IBM Tivoli Storage Manager

for Windows
Version 7.1

Administrator's Guide


IBM Tivoli Storage Manager
for Windows
Version 7.1

Administrator's Guide


 Note:
Before using this information and the product it supports, read the information in “Notices” on page 1111.

First edition (December 2013)

This edition applies to version 7, release 1, modification 0 of IBM Tivoli Storage Manager (product numbers
5608-E01, 5608-E02, 5608-E03), and to all subsequent releases and modifications until otherwise indicated in new
editions.
© Copyright IBM Corporation 1993, 2013.
US Government Users Restricted Rights – Use, duplication or disclosure restricted by GSA ADP Schedule Contract
with IBM Corp.
Contents
Preface . . . . . . . . . . . . . . xv Management of client operations . . . . . . . 21
Who should read this guide . . . . . . . . . xv Managing client nodes. . . . . . . . . . 22
Publications . . . . . . . . . . . . . . xv Security management . . . . . . . . . . 23
Conventions used in this guide. . . . . . . . xv Managing client data with policies . . . . . 24
Schedules for client operations . . . . . . . 25
New for IBM Tivoli Storage Manager Server maintenance . . . . . . . . . . . . 26
Server-operation management . . . . . . . 26
Version 7.1 . . . . . . . . . . . . xvii Server script automation . . . . . . . . . 27
New in Version 7.1 . . . . . . . . . . . xvii Database and recovery-log management . . . . 28
Operations Center updates . . . . . . . . xvii Sources of information about the server . . . . 28
Automatic client failover for restore operations Tivoli Storage Manager server networks . . . . 29
from replicated servers . . . . . . . . . xvii Exporting and importing data . . . . . . . 30
Software upgrade for the server database Protecting Tivoli Storage Manager and client data 30
manager. . . . . . . . . . . . . . xviii Protecting the server . . . . . . . . . . 31
Deprecated device types . . . . . . . . xviii
EMC Centera availability on Linux x86_64
systems for Tivoli Storage Manager servers . . xviii Part 2. Configuring and managing
File-space level collocation groups . . . . . xviii storage devices . . . . . . . . . . 33
Improved server efficiency when migrating from
disk for nodes with a large number of file spaces xix Chapter 3. Storage device concepts . . 35
Shared memory for database backup and restore
Road map for key device-related task information 35
operations . . . . . . . . . . . . . xix
Tivoli Storage Manager storage devices . . . . . 36
Immediate use of space that is added to the
Tivoli Storage Manager storage objects . . . . . 36
server database . . . . . . . . . . . . xix
Libraries . . . . . . . . . . . . . . 36
Server components that are not delivered with
Drives . . . . . . . . . . . . . . . 40
Version 7.1 . . . . . . . . . . . . . . xx
Device class . . . . . . . . . . . . . 40
New best practices . . . . . . . . . . . . xx
Library, drive, and device-class objects . . . . 43
Duplicate identification processes and memory
Storage pools and storage-pool volumes . . . . 43
requirements . . . . . . . . . . . . . xx
Data movers . . . . . . . . . . . . . 45
Server startup and setting ulimits . . . . . . xx
Paths . . . . . . . . . . . . . . . 45
Server objects. . . . . . . . . . . . . 45
Part 1. Tivoli Storage Manager Tivoli Storage Manager volumes . . . . . . . 46
basics . . . . . . . . . . . . . . . 1 Volume inventory for an automated library. . . 47
Device configurations . . . . . . . . . . . 47
Devices on local area networks . . . . . . . 47
Chapter 1. Tivoli Storage Manager Devices on storage area networks . . . . . . 47
overview . . . . . . . . . . . . . . 3 LAN-free data movement. . . . . . . . . 49
How client data is stored . . . . . . . . . . 5 Network-attached storage . . . . . . . . 50
Data-protection options . . . . . . . . . . 8 Mixed device types in libraries . . . . . . . 52
Data movement to server storage . . . . . . 14 Library sharing . . . . . . . . . . . . 54
Consolidation of backed-up client data . . . . 14 Removable media mounts and dismounts . . . . 54
How the server manages storage . . . . . . . 15 How Tivoli Storage Manager uses and reuses
Device support for server storage . . . . . . 15 removable media . . . . . . . . . . . . 55
Data migration through the storage hierarchy . . 16 Definitions for storage devices . . . . . . . . 58
Removal of expired data . . . . . . . . . 16 Example: Mapping devices to device classes . . 59
Example: Mapping storage pools to device
Chapter 2. Tivoli Storage Manager classes and devices . . . . . . . . . . . 59
concepts . . . . . . . . . . . . . . 17 Planning for server storage . . . . . . . . . 60
Interfaces to Tivoli Storage Manager . . . . . . 17 Server options that affect storage operations . . . 62
Storage configuration and management . . . . . 18
Disk devices . . . . . . . . . . . . . 18 Chapter 4. Magnetic disk devices . . . 63
Removable media devices . . . . . . . . 19 Requirements for disk systems . . . . . . . . 63
Migrating data from disk to tape . . . . . . 19 Comparison of random access and sequential access
Techniques for managing your storage pools and disk devices . . . . . . . . . . . . . . 65
volumes . . . . . . . . . . . . . . 20 Configuring random access volumes on disk devices 69

© Copyright IBM Corp. 1993, 2013 iii

Configuring FILE sequential volumes on disk Considerations for using virtual tape libraries 116
devices . . . . . . . . . . . . . . . . 70 Adding a virtual tape library to your
Varying disk volumes online or offline . . . . . 70 environment . . . . . . . . . . . . . 117
Cache copies for files stored on disk . . . . . . 71 Defining or deleting all drives and paths for a
Freeing space on disk . . . . . . . . . . . 71 single library . . . . . . . . . . . . 118
Scratch FILE volumes . . . . . . . . . . . 72 Configuring Tivoli Storage Manager servers to
Volume history file and volume reuse . . . . . 72 share SAN-connected devices . . . . . . . . 119
Setting up server communications . . . . . 119
Chapter 5. Attaching devices for the Setting up the library manager server . . . . 119
server . . . . . . . . . . . . . . . 73 Setting up the library client servers . . . . . 121
Configuring Tivoli Storage Manager for LAN-free
Attaching a manual drive. . . . . . . . . . 73
data movement. . . . . . . . . . . . . 122
Attaching an automated library device . . . . . 74
Validating your LAN-free configuration . . . 123
Device alias names . . . . . . . . . . . . 75
Troubleshooting device configuration . . . . . 123
Displaying information about devices with tsmdlst 76
Displaying device information. . . . . . . 123
Selecting a device driver . . . . . . . . . . 78
Troubleshooting problems with devices. . . . 124
Drivers for IBM devices . . . . . . . . . 78
Impact of device changes on the SAN . . . . 125
Drivers for non IBM devices. . . . . . . . 79
Increased block size for writing to tape . . . . . 126
Installing device drivers for IBM 3494 libraries . 80
Installing the Tivoli Storage Manager device
driver . . . . . . . . . . . . . . . 80 Chapter 7. Managing removable media
Uninstalling the Tivoli Storage Manager device operations. . . . . . . . . . . . . 127
driver . . . . . . . . . . . . . . . 81 Defining volumes . . . . . . . . . . . . 127
Using Windows device drivers with the Managing volumes . . . . . . . . . . . 128
GENERICTAPE format . . . . . . . . . 81 Partially-written volumes . . . . . . . . 128
Creating a file to list devices and their attributes 82 Volume inventory for automated libraries . . . 128
Controlling devices with the Tivoli Storage Changing the status of database-backup and
Manager device driver. . . . . . . . . . 82 database-export volumes . . . . . . . . 129
Installing the Centera SDK for Centera shared Preparing media for automated libraries . . . . 129
libraries . . . . . . . . . . . . . . . 83 Labeling media . . . . . . . . . . . . 129
Checking media into automated library devices 130
Chapter 6. Configuring storage devices 85 Write-once, read-many tape media . . . . . 133
Planning for device configuration . . . . . . . 86 Managing media in automated libraries . . . . 135
Manually configuring devices . . . . . . . 86 Changing the status of automated library
Configuring devices using Tivoli Storage Manager volumes . . . . . . . . . . . . . . 135
commands. . . . . . . . . . . . . . . 91 Removing volumes from automated libraries 136
Defining Tivoli Storage Manager storage objects Returning partially-written volumes to
with commands . . . . . . . . . . . . 92 automated libraries . . . . . . . . . . 136
Determining backup strategies . . . . . . . 95 Returning reclaimed volumes to a library . . . 137
Determining the media and device type for client Auditing volume inventories in libraries . . . 137
backups . . . . . . . . . . . . . . 95 Adding scratch volumes to automated library
Configuring IBM 3494 libraries . . . . . . . . 96 devices . . . . . . . . . . . . . . 138
Categories in an IBM 3494 library . . . . . . 96 Setting up volume overflow locations for
Configuring an IBM 3494 library for use by one automated libraries . . . . . . . . . . 138
server . . . . . . . . . . . . . . . 97 Modifying volume access modes . . . . . . 139
Sharing an IBM 3494 library among servers . . 102 Shared libraries. . . . . . . . . . . . 140
Migrating a shared IBM 3494 library to a library Category numbers for IBM 3494 libraries . . . 141
manager . . . . . . . . . . . . . . 104 Media reuse in automated libraries . . . . . 141
Sharing an IBM 3494 library by static Media management in manual libraries . . . . 142
partitioning of drives . . . . . . . . . . 105 Tivoli Storage Manager server requests . . . . . 143
ACSLS-managed libraries . . . . . . . . . 109 Starting the administrative client as a server
Configuring an ACSLS-managed library . . . 109 console monitor . . . . . . . . . . . 143
Configuring an ACSLS library with a single Displaying information about volumes that are
drive device type . . . . . . . . . . . 110 currently mounted. . . . . . . . . . . 144
Configuring an ACSLS library with multiple Displaying information about mount requests
drive device type . . . . . . . . . . . 111 that are pending . . . . . . . . . . . 144
Setting up an ACSLS library manager server . . 113 Replying to mount requests . . . . . . . 144
Setting up an ACSLS library client server . . . 114 Canceling mount requests . . . . . . . . 144
Checking in and labeling ACSLS library Responding to requests for volume checkin . . 145
volumes . . . . . . . . . . . . . . 115 Dismounting idle volumes . . . . . . . . 145
Managing virtual tape libraries . . . . . . . 116 Dismounting volumes from stand-alone
removable-file devices . . . . . . . . . 146

iv IBM Tivoli Storage Manager for Windows: Administrator's Guide

 Obtaining tape alert messages . . . . . . . 146 NDMP operations management . . . . . . . 206
Tape rotation . . . . . . . . . . . . . 146 Managing NAS file server nodes . . . . . . 207
Labeling volumes using commands . . . . . . 148 Managing data movers used in NDMP
Tivoli Storage Manager media-manager support 148 operations . . . . . . . . . . . . . 208
Using external media managers to control media 149 Dedicating a Tivoli Storage Manager drive to
Requirements for managing media in external NDMP operations . . . . . . . . . . . 208
libraries . . . . . . . . . . . . . . 149 Storage pool management for NDMP operations 209
Troubleshooting database errors . . . . . . . 150 Managing tables of contents . . . . . . . 209
Managing libraries . . . . . . . . . . . 150 Preventing long-running, inactive NDMP
Obtaining information about libraries . . . . 150 connections from closing . . . . . . . . 210
Updating automated libraries . . . . . . . 151 Configuring Tivoli Storage Manager for NDMP
Deleting libraries . . . . . . . . . . . 152 operations . . . . . . . . . . . . . . 211
Managing drives . . . . . . . . . . . . 152 Configuring Tivoli Storage Manager policy for
Requesting information about drives . . . . 152 NDMP operations . . . . . . . . . . . 212
Updating drives . . . . . . . . . . . 153 Tape libraries and drives for NDMP operations 215
Using drive encryption . . . . . . . . . 154 Attaching tape library robotics for NAS-attached
Validating data during read/write operations to libraries . . . . . . . . . . . . . . 219
tape . . . . . . . . . . . . . . . 155 Registering NAS nodes with the Tivoli Storage
Replacing tape drives. . . . . . . . . . 159 Manager server. . . . . . . . . . . . 223
Cleaning drives . . . . . . . . . . . 164 Defining a data mover for the NAS file server 224
Deleting drives . . . . . . . . . . . . 168 Defining paths for NDMP operations . . . . 224
Managing paths . . . . . . . . . . . . 169 Labeling and checking tapes into the library . . 228
Obtaining information about paths . . . . . 169 Scheduling NDMP operations . . . . . . . 228
Updating paths. . . . . . . . . . . . 169 Defining virtual file spaces . . . . . . . . 228
Deleting paths . . . . . . . . . . . . 170 Tape-to-tape copy to back up data . . . . . 229
Managing data movers . . . . . . . . . . 170 Tape-to-tape copy to move data . . . . . . 229
Obtaining information about data movers . . . 170 Backing up and restoring NAS file servers using
Updating data movers . . . . . . . . . 170 NDMP . . . . . . . . . . . . . . . 230
Deleting data movers. . . . . . . . . . 171 NAS file servers: backups to a single Tivoli
Managing disks . . . . . . . . . . . . 171 Storage Manager server . . . . . . . . . 231
Obtaining information about disks . . . . . 171 Backing up NDMP file server to Tivoli Storage
Updating disks . . . . . . . . . . . . 171 Manager server backups. . . . . . . . . 232
Deleting disks . . . . . . . . . . . . 172 File-level backup and restore for NDMP operations 233
Interfaces for file-level restore . . . . . . . 234
Chapter 8. Tips for defining devices 173 International characters for NetApp file servers 234
Defining devices and paths . . . . . . . . . 173 File level restore from a directory-level backup
Defining libraries . . . . . . . . . . . 173 image . . . . . . . . . . . . . . . 235
Defining drives . . . . . . . . . . . . 174 Directory-level backup and restore . . . . . . 235
Defining data movers . . . . . . . . . 175 Directory-level backup and restore for NDMP
Defining paths . . . . . . . . . . . . 176 operations . . . . . . . . . . . . . 236
Defining device classes . . . . . . . . . . 177 Backing up and restoring with snapshots . . . 236
Sequential-access device types . . . . . . . 179 Backup and restore using NetApp SnapMirror to
Defining tape device classes . . . . . . . 180 Tape feature . . . . . . . . . . . . . . 237
Defining device classes for removable media NDMP backup operations using Celerra file server
devices . . . . . . . . . . . . . . 186 integrated checkpoints . . . . . . . . . . 238
Defining sequential-access disk (FILE) device Replicating NAS nodes with NDMP backup data 238
classes. . . . . . . . . . . . . . . 187
Defining LTO device classes . . . . . . . 191 Chapter 10. Managing storage pools
Defining SERVER device classes . . . . . . 194 and volumes. . . . . . . . . . . . 239
Defining device classes for StorageTek VolSafe Storage pools . . . . . . . . . . . . . 240
devices . . . . . . . . . . . . . . 195 Primary storage pools . . . . . . . . . 240
Defining device classes for CENTERA devices 197 Copy storage pools . . . . . . . . . . 241
Obtaining information about device classes . . 198 Active-data pools . . . . . . . . . . . 241
How Tivoli Storage Manager fills volumes . . 199 Example: Setting up server storage . . . . . 243
Defining storage pools . . . . . . . . . 245
Chapter 9. Using NDMP for operations Task tips for storage pools . . . . . . . . 251
with NAS file servers . . . . . . . . 203 Storage pool volumes . . . . . . . . . . 252
NDMP requirements . . . . . . . . . . . 203 Random-access storage pool volumes . . . . 253
Interfaces for NDMP operations . . . . . . 205 Sequential-access storage pool volumes. . . . 253
Data formats for NDMP backup operations . . 206 Preparing volumes for random-access storage
pools . . . . . . . . . . . . . . . 255

Contents v
 Preparing volumes for sequential-access storage Reclaiming space in sequential-access storage pools 367
pools . . . . . . . . . . . . . . . 255 How Tivoli Storage Manager reclamation works 368
Updating storage pool volumes . . . . . . 257 Reclamation thresholds . . . . . . . . . 370
Access modes for storage pool volumes . . . 259 Reclaiming volumes with the most reclaimable
Storage pool hierarchies . . . . . . . . . . 260 space . . . . . . . . . . . . . . . 370
Setting up a storage pool hierarchy . . . . . 261 Starting reclamation manually or in a schedule 371
How the server groups files before storing . . 262 Optimizing drive usage using multiple
Where the server stores files . . . . . . . 263 concurrent reclamation processes . . . . . . 371
Example: How the server determines where to Reclaiming volumes in a storage pool with one
store files in a hierarchy . . . . . . . . . 264 drive . . . . . . . . . . . . . . . 372
Backing up the data in a storage hierarchy . . 265 Reducing the time to reclaim tape volumes with
Staging client data from disk to tape . . . . 270 high capacity . . . . . . . . . . . . 373
Migrating files in a storage pool hierarchy. . . . 271 Reclamation of write-once, read-many (WORM)
Migrating disk storage pools . . . . . . . 273 media . . . . . . . . . . . . . . . 373
Migrating sequential-access storage pools . . . 279 Controlling reclamation of virtual volumes . . 374
The effect of migration on copy storage pools Reclaiming copy storage pools and active-data
and active-data pools . . . . . . . . . . 283 pools . . . . . . . . . . . . . . . 374
Caching in disk storage pools . . . . . . . . 284 How collocation affects reclamation . . . . . 378
How the server removes cached files . . . . 284 Estimating space needs for storage pools . . . . 379
Effect of caching on storage pool statistics . . . 285 Estimating space requirments in random-access
Deduplicating data . . . . . . . . . . . 285 storage pools . . . . . . . . . . . . 380
Data deduplication overview . . . . . . . 285 Estimating space needs in sequential-access
Data deduplication limitations. . . . . . . 288 storage pools . . . . . . . . . . . . 381
Planning guidelines for data deduplication . . 291 Monitoring storage-pool and volume usage . . . 382
Detecting possible security attacks during Monitoring space available in a storage pool 382
client-side data deduplication . . . . . . . 302 Monitoring the use of storage pool volumes . . 384
Evaluating data deduplication in a test Monitoring migration processes . . . . . . 392
environment. . . . . . . . . . . . . 303 Monitoring the use of cache space on disk
Managing deduplication-enabled storage pools 306 storage . . . . . . . . . . . . . . 395
Controlling data deduplication . . . . . . 310 Obtaining information about the use of storage
Displaying statistics about server-side data space . . . . . . . . . . . . . . . 396
deduplication . . . . . . . . . . . . 318 Moving data from one volume to another volume 400
Displaying statistics about client-side data Data movement within the same storage pool 401
deduplication . . . . . . . . . . . . 320 Data movement to a different storage pool . . 402
Querying about data deduplication in file Data movement from offsite volumes in copy
spaces . . . . . . . . . . . . . . . 323 storage pools or active-data pools . . . . . 402
Scenarios for data deduplication . . . . . . 324 Moving data . . . . . . . . . . . . 403
Data deduplication and data compatibility . . 329 Moving data belonging to a client node . . . . 406
Data deduplication and disaster recovery Moving data in all file spaces belonging to one
management . . . . . . . . . . . . 330 or more nodes . . . . . . . . . . . . 406
Writing data simultaneously to primary, copy, and Moving data in selected file spaces belonging to
active-data pools . . . . . . . . . . . . 331 a single node . . . . . . . . . . . . 407
Guidelines for using the simultaneous-write Obtaining information about data-movement
function . . . . . . . . . . . . . . 332 processes . . . . . . . . . . . . . . 408
Limitations that apply to simultaneous-write Troubleshooting incomplete data-movement
operations . . . . . . . . . . . . . 333 operations . . . . . . . . . . . . . 408
Controlling the simultaneous-write function . . 335 Renaming storage pools . . . . . . . . . . 409
Simultaneous-write operations: Examples . . . 339 Defining copy storage pools and active-data pools 409
Planning simultaneous-write operations . . . 352 Example: Defining a copy storage pool . . . . 411
Simultaneous-write function as part of a backup Properties of primary, copy, and active-data
strategy: Example . . . . . . . . . . . 356 pools . . . . . . . . . . . . . . . 412
Keeping client files together using collocation . . 357 Deleting storage pools . . . . . . . . . . 413
The effects of collocation on operations . . . . 359 Deleting storage pool volumes . . . . . . . 414
How the server selects volumes with collocation Deleting empty storage pool volumes . . . . 415
enabled . . . . . . . . . . . . . . 361 Deleting storage pool volumes that contain data 415
How the server selects volumes with collocation
disabled . . . . . . . . . . . . . . 363
Collocation on or off settings . . . . . . . 363
Part 3. Managing client operations 417
Collocation of copy storage pools and
active-data pools . . . . . . . . . . . 364 Chapter 11. Adding client nodes . . . 419
Planning for and enabling collocation . . . . 365 Overview of clients and servers as nodes . . . . 419

vi IBM Tivoli Storage Manager for Windows: Administrator's Guide

Installing client node software . . . . . . . . 420 Chapter 13. Implementing policies for
Registering nodes with the server . . . . . . 420 client data . . . . . . . . . . . . . 475
Accepting default closed registration or enabling Basic policy planning . . . . . . . . . . . 476
open registration . . . . . . . . . . . 421 Reviewing the standard policy . . . . . . 477
Registering nodes with client options sets . . . 423 Getting users started . . . . . . . . . . 478
Registering a network-attached storage file Changing policy . . . . . . . . . . . 478
server as a node . . . . . . . . . . . 423 File expiration and expiration processing . . . 479
Registering a source server as a node on a target Client operations controlled by policy . . . . . 480
server . . . . . . . . . . . . . . . 424 Backup and restore . . . . . . . . . . 480
Registering an API to the server . . . . . . 424 Archive and retrieve . . . . . . . . . . 481
Connecting nodes with the server . . . . . . 425 Client migration and recall . . . . . . . . 481
Required client options . . . . . . . . . 425 The parts of a policy . . . . . . . . . . . 482
UNIX and Linux client options . . . . . . 426 Relationships among clients, storage, and policy 484
Updating the password for scheduling operations 426 More on management classes . . . . . . . . 486
Creating or updating a client options file . . . . 426 Contents of a management class . . . . . . 486
Using a text editor to create or configure a client Default management classes . . . . . . . 487
options file . . . . . . . . . . . . . 427 The include-exclude list . . . . . . . . . 488
Using the client configuration wizard to create How files and directories are associated with a
or update a client options file . . . . . . . 427 management class . . . . . . . . . . . 489
Comparing network-attached nodes to local nodes 427 How Tivoli Storage Manager selects files for policy
Adding clients through the administrative operations . . . . . . . . . . . . . . 492
command line client . . . . . . . . . . . 428 Incremental backup . . . . . . . . . . 492
Enabling open registration . . . . . . . . 428 Selective backup . . . . . . . . . . . 494
Example: registering three client nodes using Logical volume backup . . . . . . . . . 495
the administrative command line . . . . . . 428 Archive . . . . . . . . . . . . . . 495
Automatic migration from a client node . . . 496
Chapter 12. Managing client nodes 431 Creating your own policies . . . . . . . . . 496
Managing nodes . . . . . . . . . . . . 431 Example: sample policy objects . . . . . . 497
Managing client nodes across a firewall . . . 432 Defining and updating a policy domain . . . 498
Updating client node information . . . . . 434 Defining and updating a policy set . . . . . 500
Command-line backup-archive client automatic Defining and updating a management class . . 501
deployment feature: overview . . . . . . . 434 Defining and updating a backup copy group 502
Renaming client nodes . . . . . . . . . 439 Defining and updating an archive copy group 508
Locking and unlocking client nodes . . . . . 440 Assigning a default management class . . . . 510
Deleting client nodes . . . . . . . . . . 440 Validating and activating a policy set . . . . 510
Consolidating multiple clients under a single Assigning client nodes to a policy domain. . . . 512
client node name . . . . . . . . . . . 441 Running expiration processing to delete expired
Displaying information about client nodes. . . 443 files . . . . . . . . . . . . . . . . 512
Overview of remote access to web Running expiration processing automatically 513
backup-archive clients . . . . . . . . . 445 Using commands to control expiration
Managing client access authority levels . . . . 447 processing . . . . . . . . . . . . . 513
Collecting processor value unit information in a Additional expiration processing with disaster
VMware host environment . . . . . . . . 450 recovery manager . . . . . . . . . . . 514
Managing file spaces . . . . . . . . . . . 451 Protection and expiration of archive data . . . . 514
Defining client nodes and file spaces . . . . 451 Data retention protection . . . . . . . . 514
Supporting Unicode-enabled clients . . . . . 453 Deletion hold . . . . . . . . . . . . 515
Displaying information about file spaces . . . 462 Protecting data using the NetApp SnapLock
Moving data for a client node . . . . . . . 463 licensed feature. . . . . . . . . . . . . 516
Deleting file spaces . . . . . . . . . . 463 Reclamation and the SnapLock feature . . . . 517
Managing client option files . . . . . . . . 464 Set up SnapLock volumes as Tivoli Storage
Creating client option sets on the server . . . 465 Manager WORM FILE volumes . . . . . . 521
Managing client option sets . . . . . . . 467 Policy configuration scenarios . . . . . . . . 522
Managing IBM Tivoli Storage Manager sessions 467 Configuring policy for direct-to-tape backups 522
Displaying information about IBM Tivoli Configuring policy for Tivoli Storage Manager
Storage Manager sessions . . . . . . . . 468 application clients . . . . . . . . . . . 523
Canceling an IBM Tivoli Storage Manager Policy for logical volume backups . . . . . 523
session . . . . . . . . . . . . . . 469 Configuring policy for NDMP operations . . . 524
When a client session is automatically canceled 470 Configuring policy for LAN-free data
Disabling or enabling access to the server . . . 471 movement . . . . . . . . . . . . . 525
Managing client restartable restore sessions . . 471 Policy for Tivoli Storage Manager servers as
clients . . . . . . . . . . . . . . . 527

Contents vii
 Setting policy to enable point-in-time restore for Copying existing schedules . . . . . . . . 570
clients . . . . . . . . . . . . . . . 527 Modifying schedules . . . . . . . . . . 570
Distributing policy using enterprise configuration 528 Deleting schedules . . . . . . . . . . 570
Querying policy . . . . . . . . . . . . 529 Displaying information about schedules . . . 571
Querying copy groups . . . . . . . . . 529 Managing node associations with schedules . . . 571
Querying management classes. . . . . . . 530 Adding new nodes to existing schedules . . . 572
Querying policy sets . . . . . . . . . . 530 Moving nodes from one schedule to another 572
Querying policy domains . . . . . . . . 531 Displaying nodes associated with schedules . . 572
Deleting policy . . . . . . . . . . . . . 531 Removing nodes from schedules . . . . . . 573
Deleting copy groups. . . . . . . . . . 532 Managing event records . . . . . . . . . . 573
Deleting management classes . . . . . . . 532 Displaying information about scheduled events 573
Deleting policy sets . . . . . . . . . . 532 Managing event records in the server database 575
Deleting policy domains. . . . . . . . . 532 Managing the throughput of scheduled operations 576
Modifying the default scheduling mode . . . 576
Chapter 14. Managing data for client Specifying the schedule period for incremental
nodes. . . . . . . . . . . . . . . 535 backup operations . . . . . . . . . . . 579
Balancing the scheduled workload for the server 579
Validating a node's data . . . . . . . . . . 535
Controlling how often client nodes contact the
Performance considerations for data validation 536
server . . . . . . . . . . . . . . . 581
Validating a node's data during a client session 536
Specifying one-time actions for client nodes . . . 583
Encrypting data on tape . . . . . . . . . . 536
Determining how long the one-time schedule
Choosing an encryption method . . . . . . 537
remains active . . . . . . . . . . . . 583
Changing your encryption method and
hardware configuration . . . . . . . . . 538
Securing sensitive client data . . . . . . . . 539 Part 4. Maintaining the server . . . 585
Setting up shredding . . . . . . . . . . 540
Ensuring that shredding is enforced . . . . . 541 Chapter 17. Managing the storage
Creating and using client backup sets . . . . . 543
environment with the Operations
Generating client backup sets on the server . . 544
Restoring backup sets from a backup-archive Center . . . . . . . . . . . . . . 587
client . . . . . . . . . . . . . . . 548 Operations Center overview . . . . . . . . 587
Moving backup sets to other servers. . . . . 548 Opening the Operations Center . . . . . . 589
Managing client backup sets . . . . . . . 549 Alert monitoring overview . . . . . . . . . 590
Enabling clients to use subfile backup . . . . . 552 Changing the Operations Center configuration . . 591
Setting up clients to use subfile backup. . . . 553 Changing the hub server . . . . . . . . 591
Managing subfile backups . . . . . . . . 553 Removing a spoke server . . . . . . . . 591
Archiving data . . . . . . . . . . . . . 554 Restoring the configuration to the preconfigured
Archive operations overview . . . . . . . 555 state . . . . . . . . . . . . . . . 592
Managing storage usage for archives . . . . 555 Starting and stopping the web server . . . . 593
Restarting the initial configuration wizard . . . 593
Chapter 15. Scheduling operations for
client nodes . . . . . . . . . . . . 559 Chapter 18. Managing server
Prerequisites to scheduling operations . . . . . 559 operations. . . . . . . . . . . . . 595
Scheduling a client operation . . . . . . . . 560 Licensing IBM Tivoli Storage Manager . . . . . 595
Defining client schedules . . . . . . . . 561 Registering licensed features . . . . . . . 596
Associating client nodes with schedules . . . 562 Monitoring licenses . . . . . . . . . . 597
Starting the scheduler on the clients . . . . . 562 Role of processor value units in assessing
Displaying schedule information . . . . . . 563 licensing requirements . . . . . . . . . 598
Checking the status of scheduled operations . . 563 Estimating processor value units . . . . . . 601
Creating schedules for running command files . . 564 Collecting processor value unit information in a
Updating the client options file to automatically VMware host environment . . . . . . . . 603
generate a new password . . . . . . . . . 565 Starting the Tivoli Storage Manager server . . . 604
Configuring the scheduler to run under the Starting the server on Windows . . . . . . 604
site-server account. . . . . . . . . . . . 565 Running multiple server instances on a single
Overview of the Tivoli Storage Manager scheduler system . . . . . . . . . . . . . . 606
running as a Windows service . . . . . . . . 566 Halting the server . . . . . . . . . . . . 607
Moving the Tivoli Storage Manager server to
another system . . . . . . . . . . . . . 608
Chapter 16. Managing schedules for
Date and time on the server . . . . . . . . 609
client nodes . . . . . . . . . . . . 569 Managing server processes . . . . . . . . . 609
Managing IBM Tivoli Storage Manager schedules 569 Requesting information about server processes 610
Adding new schedules . . . . . . . . . 569

viii IBM Tivoli Storage Manager for Windows: Administrator's Guide

 Canceling server processes . . . . . . . . 610 Increasing the size of the active log . . . . . . 678
Preempting operations . . . . . . . . . 611 Reducing the size of the active log . . . . . . 679
Changing names in Tivoli Storage Manager . . . 613 Moving the database and recovery log on a server 679
Setting the server name . . . . . . . . . 613 Moving both the database and recovery log . . 680
Changing the host name for a Tivoli Storage Moving only the database . . . . . . . . 680
Manager server. . . . . . . . . . . . 613 Moving only the active log, archive log, or
Add or update server options . . . . . . . . 614 archive failover log . . . . . . . . . . 681
Adding or updating a server option without Specifying alternative locations for database log
restarting the server . . . . . . . . . . 615 files . . . . . . . . . . . . . . . . 682
Deleting information about volume history . . . 615 Specifying an alternative location with the
Getting help on commands and error messages . . 616 ARCHFAILOVERLOGDIRECTORY server option or
parameter . . . . . . . . . . . . . 683
Chapter 19. Automating server Specifying an alternative location with the
operations. . . . . . . . . . . . . 619 ARCHLOGDIRECTORY server option or parameter . 683
Specifying the location of RstDbLog using the
Automating a basic administrative command
RECOVERYDIR parameter . . . . . . . . . 684
schedule . . . . . . . . . . . . . . . 620
Adding optional logs after server initialization . . 684
Defining the schedule . . . . . . . . . 620
Transaction processing . . . . . . . . . . 685
Verifying the schedule . . . . . . . . . 621
Files moved as a group between client and
Tailoring schedules . . . . . . . . . . . 621
server . . . . . . . . . . . . . . . 685
Using classic and enhanced command schedules 624
Copying schedules . . . . . . . . . . . 624
Deleting schedules . . . . . . . . . . . 625 Chapter 21. Managing a network of
Managing scheduled event records . . . . . . 625 Tivoli Storage Manager servers . . . 687
Querying events . . . . . . . . . . . 625 Concepts for managing server networks . . . . 687
Removing event records from the database . . 626 Enterprise configuration . . . . . . . . . 688
Tivoli Storage Manager server scripts . . . . . 626 Command routing. . . . . . . . . . . 689
Defining a server script . . . . . . . . . 627 Central monitoring for the Tivoli Storage
Managing server scripts . . . . . . . . . 633 Manager server. . . . . . . . . . . . 689
Running a server script . . . . . . . . . 636 Data storage on another server . . . . . . 690
Using macros . . . . . . . . . . . . . 637 Examples: management of multiple Tivoli
Writing commands in a macro. . . . . . . 637 Storage Manager servers . . . . . . . . 690
Writing comments in a macro . . . . . . . 638 Enterprise-administration planning . . . . . . 692
Using continuation characters . . . . . . . 638 Setting up communications among servers . . . 692
Using substitution variables in a macro. . . . 639 Setting up communications for enterprise
Running a macro . . . . . . . . . . . 640 configuration and enterprise event logging . . 692
Command processing in a macro . . . . . . 640 Setting up communications for command
routing . . . . . . . . . . . . . . 696
Chapter 20. Managing the database Updating and deleting servers. . . . . . . 700
and recovery log . . . . . . . . . . 643 Setting up enterprise configurations . . . . . . 701
Enterprise configuration scenario . . . . . . 702
Database and recovery log overview . . . . . 643
Creating the default profile on a configuration
Database: Overview . . . . . . . . . . 644
manager . . . . . . . . . . . . . . 706
Connecting the server to the database with
Creating and changing configuration profiles 707
TCP/IP . . . . . . . . . . . . . . 646
Getting information about profiles . . . . . 714
Recovery log . . . . . . . . . . . . 646
Subscribing to a profile . . . . . . . . . 717
Setting the user data limit for the database . . . 650
Refreshing configuration information . . . . . 721
Disk space requirements for the server database
Managing problems with configuration refresh 722
and recovery log . . . . . . . . . . . . 650
Returning managed objects to local control . . . 722
Capacity planning . . . . . . . . . . . . 651
Setting up administrators for the servers . . . . 723
Estimating database space requirements . . . 652
Managing problems with synchronization of
Estimating recovery log space requirements . . 656
profiles . . . . . . . . . . . . . . . 723
Monitoring space utilization for the database
Switching a managed server to a different
and recovery logs . . . . . . . . . . . 669
configuration manager . . . . . . . . . . 724
Monitoring the database and recovery log . . . . 671
Deleting subscribers from a configuration manager 724
Increasing the size of the database . . . . . . 672
Renaming a managed server . . . . . . . . 725
Manually redistributing data and reclaiming
Completing tasks on multiple servers . . . . . 725
space after you add directories to the database . 674
Managing multiple servers through a web-based
Reducing the size of the database . . . . . . 676
interface . . . . . . . . . . . . . . 725
Scheduling table and index reorganization. . . . 676
Routing commands . . . . . . . . . . 726
Restrictions to table and index reorganization 677
Setting up server groups . . . . . . . . 728
Scheduling table or index reorganization . . . 678
Querying server availability . . . . . . . 732

Contents ix
Using virtual volumes to store data on another Using SQL to query the IBM Tivoli Storage
server . . . . . . . . . . . . . . . . 732 Manager database . . . . . . . . . . . . 794
Setting up source and target servers for virtual Using SELECT commands . . . . . . . . 794
volumes . . . . . . . . . . . . . . 734 Using SELECT commands in Tivoli Storage
Performance limitations for virtual volume Manager scripts . . . . . . . . . . . 797
operations . . . . . . . . . . . . . 735 Querying the SQL activity summary table . . . 798
Performing operations at the source server . . 736 Creating output for use by another application 799
Reconciling virtual volumes and archive files 739 Using the Tivoli Storage Manager activity log . . 799
Requesting information from the activity log 800
Chapter 22. Exporting and importing Setting a retention period for the activity log 801
data . . . . . . . . . . . . . . . 741 Setting a size limit for the activity log . . . . 801
Reviewing data that can be exported and imported 741
Exporting restrictions. . . . . . . . . . 742 Chapter 25. Monitoring Tivoli Storage
Deciding what information to export . . . . 742 Manager accounting records. . . . . 803
Deciding when to export . . . . . . . . 743
Exporting data directly to another server . . . . 744 Chapter 26. Reporting and monitoring
Options to consider before exporting . . . . 745 with Tivoli Monitoring for Tivoli
Preparing to export to another server for
immediate import . . . . . . . . . . . 748
Storage Manager . . . . . . . . . . 805
Monitoring the server-to-server export process 750
Exporting administrator information to another Chapter 27. Logging IBM Tivoli
server . . . . . . . . . . . . . . . 751 Storage Manager events to receivers . 807
Exporting client node information to another Enabling and disabling events . . . . . . . . 808
server . . . . . . . . . . . . . . . 751 Beginning and ending event logging . . . . . 809
Exporting policy information to another server 752 Logging events to the IBM Tivoli Storage Manager
Exporting server data to another server . . . 752 server console and activity log . . . . . . . 809
Exporting and importing data using sequential Logging events to a file exit and a user exit . . . 810
media volumes . . . . . . . . . . . . . 753 Adding a file exit or user exit option . . . . 810
Using preview before exporting or importing Logging events to the Tivoli Enterprise Console 811
data . . . . . . . . . . . . . . . 753 Controlling the format of events . . . . . . 812
Planning for sequential media used to export Encoding events to UTF-8 . . . . . . . . 814
data . . . . . . . . . . . . . . . 754 Setting up a Tivoli Enterprise Console as a
Exporting tasks. . . . . . . . . . . . 755 receiver . . . . . . . . . . . . . . 814
Importing data from sequential media volumes 758 Logging events to an SNMP manager . . . . . 815
Monitoring export and import processes . . . 770 Configuring IBM Tivoli Storage Manager SNMP 818
Exporting and importing data from virtual Logging events to the Windows event log . . . . 820
volumes . . . . . . . . . . . . . . 773 Enterprise event logging: logging events to another
server . . . . . . . . . . . . . . . . 820
Part 5. Monitoring operations . . . 775 Querying event logging . . . . . . . . . . 822
User exit and file exit receivers . . . . . . . 822
Sample user-exit declarations . . . . . . . 823
Chapter 23. Daily monitoring tasks 777 Sample user exit program . . . . . . . . 825
Monitoring operations using the command line 778 Readable text file exit (FILETEXTEXIT) format 826
Monitoring your server processes daily. . . . 778
Monitoring your database daily . . . . . . 779
Monitoring disk storage pools daily . . . . . 782 Part 6. Protecting the server . . . 829
Monitoring sequential access storage pools daily 783
Monitoring scheduled operations daily . . . . 786 Chapter 28. Managing Tivoli Storage
Manager security . . . . . . . . . . 831
Chapter 24. Basic monitoring methods 789 Securing communications . . . . . . . . . 831
Using IBM Tivoli Storage Manager queries to Secure Sockets Layer and Transport Layer
display information . . . . . . . . . . . 789 Security communication . . . . . . . . . 832
Requesting information about IBM Tivoli Setting up SSL . . . . . . . . . . . . 833
Storage Manager definitions . . . . . . . 789 Securing the server console . . . . . . . . . 843
Requesting information about client sessions 790 Managing access to the server and clients . . . . 844
Requesting information about server processes 791 Managing Tivoli Storage Manager administrator
Requesting information about server settings 792 IDs . . . . . . . . . . . . . . . . 844
Querying server options . . . . . . . . . 792 Managing passwords and logon procedures . . . 852
Querying the system . . . . . . . . . . 793 Configuring a directory server for password
authentication . . . . . . . . . . . . 855

x IBM Tivoli Storage Manager for Windows: Administrator's Guide

 Setting the policy for an LDAP-authenticated Client node attributes that are updated during
password . . . . . . . . . . . . . . 856 replication . . . . . . . . . . . . . 932
Configuring the Tivoli Storage Manager server Node replication restrictions . . . . . . . . 933
to authenticate passwords with an LDAP Task tips for node replication . . . . . . . . 935
directory server . . . . . . . . . . . 857 Change replication rules . . . . . . . . . 935
Registering nodes and administrator IDs to Add and remove client nodes for replication 936
authenticate passwords with an LDAP directory Manage replication servers . . . . . . . . 936
server . . . . . . . . . . . . . . . 858 Validate a configuration and preview results 937
Updating nodes and administrator IDs to Manage replication processing. . . . . . . 937
authenticate passwords with a directory server . 859 Monitor replication processing and verify results 938
Determining which nodes and administrator IDs Planning for node replication . . . . . . . . 939
are configured to authenticate with an LDAP Determining server database requirements for
server . . . . . . . . . . . . . . . 860 node replication . . . . . . . . . . . 940
Modifying the default password expiration Estimating the total amount of data to be
period for passwords that are managed by the replicated. . . . . . . . . . . . . . 941
Tivoli Storage Manager server . . . . . . . 860 Estimating network bandwidth required for
Scenarios for authenticating passwords . . . . 861 replication . . . . . . . . . . . . . 941
Setting a limit for invalid password attempts 863 Calculating the time that is required for
Setting a minimum length for a password . . . 864 replication . . . . . . . . . . . . . 942
Disabling the default password authentication 865 Selecting a method for the initial replication . . 943
Enabling unified logon with backup-archive Scheduling incremental replication after the
clients . . . . . . . . . . . . . . . 865 initial replication . . . . . . . . . . . 945
Setting up the default replication configuration . . 946
Chapter 29. Protecting and recovering Step 1: Setting up server-to-server
the server infrastructure and client communications . . . . . . . . . . . 948
Step 2: Specifying a target replication server . . 950
data . . . . . . . . . . . . . . . 869 Step 3: Configuring client nodes for replication 950
Protecting the database and infrastructure setup Customizing a node replication configuration . . 952
files . . . . . . . . . . . . . . . . 870 Changing replication rules . . . . . . . . 952
Backing up the server database . . . . . . 870 Scenario: Converting to node replication from
Protecting infrastructure setup files . . . . . 877 import and export operations . . . . . . . 961
Protecting client data . . . . . . . . . . . 882 Adding and removing client nodes for
Protecting the data that is in primary storage replication . . . . . . . . . . . . . 962
pools . . . . . . . . . . . . . . . 882 Managing source and target replication servers 965
Auditing storage pool volumes . . . . . . 887 Verifying a node replication setup before
Fixing damaged files . . . . . . . . . . 896 processing . . . . . . . . . . . . . . 969
Scenario: Protecting the database and storage pools 898 Validating a replication configuration . . . . 970
Recovering the database and client data . . . . 900 Previewing node replication results . . . . . 970
Restoring the database . . . . . . . . . 900 Managing data replication . . . . . . . . . 971
Restoring storage pools and storage pool Replicating data by command . . . . . . . 971
volumes . . . . . . . . . . . . . . 907 Controlling throughput for node replication . . 976
Restoring to a point-in-time in a shared library Disabling and enabling node replication . . . 978
environment. . . . . . . . . . . . . . 915 Purging replicated data in a file space . . . . 983
Restoring to a point-in-time a library manager Replicating client node data after a database
server . . . . . . . . . . . . . . . 915 restore. . . . . . . . . . . . . . . 983
Restoring to a point-in-time a library client Canceling replication processes . . . . . . 985
server . . . . . . . . . . . . . . . 916 Monitoring node replication processing and
Example: Recovering to a point-in-time. . . . 916 verifying results . . . . . . . . . . . . 985
Displaying information about node replication
Chapter 30. Replicating client node settings . . . . . . . . . . . . . . 985
data . . . . . . . . . . . . . . . 919 Displaying information about node replication
Automatic failover for data recovery overview . . 920 processes . . . . . . . . . . . . . . 985
Source and target node-replication servers. . . . 921 Measuring the effectiveness of a replication
Replication server configurations . . . . . . 921 configuration . . . . . . . . . . . . 986
Policy management for node replication . . . 922 Measuring the effects of data deduplication on
Node replication processing . . . . . . . . 923 node replication processing. . . . . . . . 987
Replication rules . . . . . . . . . . . 923 Retaining replication records . . . . . . . 988
Replication state . . . . . . . . . . . 927 Recovering and storing client data after a disaster 988
Replication mode . . . . . . . . . . . 930 Restoring, retrieving, and recalling data
Replication of deduplicated data . . . . . . 931 manually from a target replication server . . . 988

Contents xi
 Converting client nodes for store operations on Managing copy storage pool volumes and
a target replication server . . . . . . . . 989 active-data pool volumes at the recovery site . 1042
Removing a node replication configuration . . . 989 Disaster recovery manager checklist . . . . . 1043
The disaster recovery plan file . . . . . . . 1048
Chapter 31. Configuring clustered Breaking out a disaster recovery plan file. . . 1048
environments . . . . . . . . . . . 991 Structure of the disaster recovery plan file . . 1051
Example disaster recovery plan file. . . . . 1054
Clustered environment overview . . . . . . . 991
Cluster nodes . . . . . . . . . . . . 992
Configuring a Windows clustered environment . . 992 Chapter 33. Integrating disaster
Microsoft Failover Cluster environment recovery manager and node
overview . . . . . . . . . . . . . . 993 replication into your disaster
Tape failover for nodes in a cluster . . . . . 994 recovery strategy . . . . . . . . . 1075
Planning for a clustered environment . . . . 996 Plan for a disaster recovery strategy . . . . . 1076
Setting up Tivoli Storage Manager in a Tier 0: No disaster recovery capability. . . . . 1077
Microsoft Failover Cluster . . . . . . . . 999 Tier 1: Offsite vaulting from a single production
Maintaining the clustered environment . . . 1001 site . . . . . . . . . . . . . . . . 1077
Tier 2: Offsite vaulting with a recovery site . . . 1078
Chapter 32. Disaster recovery Tier 3: Electronic vaulting of critical data . . . . 1078
manager . . . . . . . . . . . . . 1005 Tier 4: Active data management at peer sites . . 1079
Querying defaults for the disaster recovery plan Tier 5: Synchronous replication . . . . . . . 1080
file . . . . . . . . . . . . . . . . 1006 Tier 6: Synchronous replication with coupling or
Specifying defaults for the disaster recovery clustering applications . . . . . . . . . . 1081
plan file . . . . . . . . . . . . . . 1006
Specifying defaults for offsite recovery media Part 7. Appendixes . . . . . . . 1083
management . . . . . . . . . . . . 1009
Specifying recovery instructions for your site . . 1011
Specifying information about your server and Appendix A. External media
client node machines . . . . . . . . . . 1013 management interface description . . 1085
Specifying recovery media for client machines 1017 CreateProcess call . . . . . . . . . . . 1085
Creating and storing the disaster recovery plan 1018 Processing during server initialization . . . . . 1086
Storing the disaster recovery plan locally . . . 1019 Processing for mount requests . . . . . . . 1086
Storing the disaster recovery plan on a target Processing for release requests . . . . . . . 1087
server . . . . . . . . . . . . . . 1019 Processing for batch requests . . . . . . . . 1087
Disaster recovery plan environmental Error handling . . . . . . . . . . . . 1088
considerations . . . . . . . . . . . . 1020 Begin batch request . . . . . . . . . . . 1088
Managing disaster recovery plan files stored on End batch request . . . . . . . . . . . 1088
target servers . . . . . . . . . . . . . 1022 Volume query request . . . . . . . . . . 1089
Displaying information about recovery plan Initialization requests . . . . . . . . . . 1090
files . . . . . . . . . . . . . . . 1022 Volume eject request . . . . . . . . . . 1090
Displaying the contents of a recovery plan file 1022 Volume release request . . . . . . . . . . 1091
Restoring a recovery plan file . . . . . . 1023 Volume mount request . . . . . . . . . . 1092
Expiring recovery plan files automatically . . 1023 Volume dismount request . . . . . . . . . 1095
Deleting recovery plan files manually . . . . 1024
Moving backup media . . . . . . . . . . 1024 Appendix B. User exit and file exit
Moving copy storage pool and active-data pool receivers . . . . . . . . . . . . . 1097
volumes offsite . . . . . . . . . . . 1026 Sample user-exit declarations. . . . . . . . 1097
Moving copy storage pool and active-data pool Sample user exit program . . . . . . . . . 1099
volumes on-site . . . . . . . . . . . 1028 Readable text file exit (FILETEXTEXIT) format 1100
Managing the Disaster Recovery Manager tasks 1029
Preparing for disaster recovery . . . . . . . 1032
Recovering from a disaster . . . . . . . . 1034
Appendix C. Tivoli support
Server recovery scenario . . . . . . . . 1034 information . . . . . . . . . . . . 1103
Client recovery scenario . . . . . . . . 1038 Communities and other learning resources . . . 1103
Recovering with different hardware at the Searching knowledge bases . . . . . . . . 1105
recovery site . . . . . . . . . . . . . 1040 Searching the Internet . . . . . . . . . 1105
Automated SCSI library at the original and Using IBM Support Assistant . . . . . . . 1105
recovery sites . . . . . . . . . . . . 1040 Finding product fixes . . . . . . . . . 1106
Automated SCSI library at the original site and Receiving notification of product fixes . . . . 1106
a manual scsi library at the recovery site . . . 1041 Contacting IBM Software Support . . . . . . 1106
Setting up and managing support contracts 1107

xii IBM Tivoli Storage Manager for Windows: Administrator's Guide

 Determining the business impact . . . . . 1107 F . . . . . . . . . . . . . . . . . 1122
Describing the problem and gathering G . . . . . . . . . . . . . . . . . 1122
background information . . . . . . . . 1107 H . . . . . . . . . . . . . . . . . 1123
Submitting the problem to IBM Software I . . . . . . . . . . . . . . . . . 1124
Support . . . . . . . . . . . . . . 1108 J . . . . . . . . . . . . . . . . . 1124
K . . . . . . . . . . . . . . . . . 1124
Appendix D. Accessibility features L . . . . . . . . . . . . . . . . . 1125
for the Tivoli Storage Manager M . . . . . . . . . . . . . . . . . 1126
N . . . . . . . . . . . . . . . . . 1127
product family . . . . . . . . . . 1109 O . . . . . . . . . . . . . . . . . 1128
P . . . . . . . . . . . . . . . . . 1128
Notices. . . . . . . . . . . . . . 1111 Q . . . . . . . . . . . . . . . . . 1129
Trademarks. . . . . . . . . . . . . . 1113 R . . . . . . . . . . . . . . . . . 1130
Privacy policy considerations . . . . . . . . 1113 S . . . . . . . . . . . . . . . . . 1131
T . . . . . . . . . . . . . . . . . 1134
Glossary . . . . . . . . . . . . . 1115 U . . . . . . . . . . . . . . . . . 1134
A . . . . . . . . . . . . . . . . . 1115 V . . . . . . . . . . . . . . . . . 1135
B . . . . . . . . . . . . . . . . . 1117 W . . . . . . . . . . . . . . . . . 1136
C . . . . . . . . . . . . . . . . . 1118
D . . . . . . . . . . . . . . . . . 1119 Index . . . . . . . . . . . . . . 1137
E . . . . . . . . . . . . . . . . . 1121

Contents xiii
xiv IBM Tivoli Storage Manager for Windows: Administrator's Guide
Preface
IBM® Tivoli® Storage Manager is a client/server program that provides storage
management solutions to customers in a multi-vendor computer environment. IBM
Tivoli Storage Manager provides an automated, centrally scheduled,
policy-managed backup, archive, and space-management facility for file servers
and workstations.

Who should read this guide

This guide is intended for anyone who is registered as an administrator for Tivoli
Storage Manager. A single administrator can manage Tivoli Storage Manager, or
several people can share administrative responsibilities.

You should be familiar with the operating system on which the server resides and
the communication protocols required for the client/server environment. You also
need to understand the storage management practices of your organization, such
as how you are currently backing up workstation files and how you are using
storage devices.

Publications
Publications for the Tivoli Storage Manager family of products are available online.
The Tivoli Storage Manager product family includes IBM Tivoli Storage
FlashCopy® Manager, IBM Tivoli Storage Manager for Space Management, IBM
Tivoli Storage Manager for Databases, and several other storage management
products from IBM Tivoli.

To search across all publications or to download PDF versions of individual

publications, go to the Tivoli Storage Manager information center at
[Link]

You also can find the Tivoli Storage Manager product family information centers
and other information centers that contain official product documentation for
current and previous versions of Tivoli products at Tivoli Documentation Central.
Tivoli Documentation Central is available at [Link]
developerworks/community/wikis/home/wiki/Tivoli Documentation Central.

Conventions used in this guide

v Command to be entered on the Windows command line:
> dsmadmc
v Command to be entered on the command line of an administrative client:
query devclass

In the usage and descriptions for administrative commands, the term characters
corresponds to the number of bytes available to store an item. For languages in
which it takes a single byte to represent a displayable character, the character to
byte ratio is 1 to 1. However, for DBCS and other multi-byte languages, the
reference to characters refers only to the number of bytes available for the item and
may represent fewer actual characters.

© Copyright IBM Corp. 1993, 2013 xv

xvi IBM Tivoli Storage Manager for Windows: Administrator's Guide
New for IBM Tivoli Storage Manager Version 7.1
Many features in the Tivoli Storage Manager Version 7.1 server are new for
previous Tivoli Storage Manager users.

New in Version 7.1

New features and other changes are available in the Tivoli Storage Manager
Version 7.1 server.

Operations Center updates

New features are available in Tivoli Storage Manager Operations Center Version
7.1 to help you manage your storage environment.

You can now do the following additional administrative tasks through the
Operations Center:
v Customize the settings for servers and clients by editing the server and client
properties.
v View or cancel the sessions and processes that are in progress by using the
Active Tasks view for servers.
v View the sessions and processes that succeeded or failed by using the
Completed Tasks view for servers.
v View activity log messages that are related to specific alerts, sessions, and
processes. These activity log messages are available on the Alerts page and in
the Active Tasks and Completed Tasks views.
v Manually back up clients, storage pools, and the server database.
v View the replication server configuration for a client.
v Register clients, and configure basic backup settings.
v Suppress risk warnings for specific clients.
v View the authority level of the administrator ID that is used to log in to the
Operations Center.
Related tasks:
Chapter 17, “Managing the storage environment with the Operations Center,” on
page 587

Automatic client failover for restore operations from replicated

servers
A Tivoli Storage Manager Version 7.1 client can automatically fail over to a target
replication server, if the source replication server is unavailable.

To use automatic failover when node replication is in use, both the source and
target replication servers and the client must be at V7.1. You can use only one
failover server for each replicating node at any time. The failover server is the last
server that a node successfully replicated to. The client can recover data from the
target replication server, but it cannot store data during failover processing.

© Copyright IBM Corp. 1993, 2013 xvii

 Software upgrade for the server database manager
Tivoli Storage Manager Version 7.1 is installed with the latest version of the IBM
DB2® database software, V10.5. The DB2 technology provides management
functions for the server database.

The previous Tivoli Storage Manager server release, V6.3, was packaged with a
DB2 V9.7 database.

If you are upgrading the server, you must ensure that the correct version of DB2 is
installed. If you are using a Tivoli Storage Manager V6.1 server, and you want to
upgrade to V7.1, you must upgrade the V6.1 server to V6.2 or V6.3. Then, upgrade
the server from V6.2 or V6.3 to V7.1. The Tivoli Storage Manager V6.1 server was
packaged with a DB2 V9.5 database, which cannot be upgraded to V10.5. The
Tivoli Storage Manager V6.2 and V6.3 servers were packaged with a DB2 V9.7
database, which can be upgraded to V10.5.

Deprecated device types

In Tivoli Storage Manager Version 7.1, certain device types are deprecated.

The following device types are deprecated. While you can use these device types
with Tivoli Storage Manager Version 7.1, plan to migrate data to new devices
before you upgrade to a later release of Tivoli Storage Manager.
v 3490
v 3570
v CARTRIDGE
v OPTICAL
v WORM
v QIC
v DTF

For a list of devices and valid device class formats, see the Tivoli Storage Manager
Supported Devices website for your operating system:
[Link]
IBM_TSM_Supported_Devices_for_AIXHPSUNWIN.html

EMC Centera availability on Linux x86_64 systems for Tivoli

Storage Manager servers
You can now use EMC Centera for versions of Linux x86_64 that are supported by
the Tivoli Storage Manager Version 7.1 server.

For details about Centera-related commands, refer to the Administrator's Reference.

File-space level collocation groups

File-space level collocation groups are introduced in Tivoli Storage Manager
Version 7.1.

You can group file spaces that belong to a single node, which allows data for these
file spaces to be collocated efficiently without requiring separate volumes for each
file space. When you use file-space level collocation groups, you can group data
for a limited set of file systems, for example, virtual machines. Fewer volumes are
required for the data and placement can be coordinated in the server storage
hierarchy.

xviii IBM Tivoli Storage Manager for Windows: Administrator's Guide

 Use the DEFINE COLLOCMEMBER command to define members of a file-space
collocation group.

For more information about collocation, see Keeping client files together using
collocation in the Administrator's Guide.

For detailed information about the DEFINE COLLOCMEMBER command, see the
Administrator's Reference.

Improved server efficiency when migrating from disk for

nodes with a large number of file spaces
With the new functions that are added to migration processing, you can improve
the efficiency of the server by using file space level migration. Nodes with multiple
large file spaces can take advantage of faster migration processing for
random-access storage pools.

For updated information about how collocation settings affect data processing and
how the server migrates files, see How the server selects files to migrate in the
Administrator's Guide.

Shared memory for database backup and restore operations

You can now use shared memory to reduce processor load and improve
throughput, if the database backup performance is slow.

You can manually configure a Tivoli Storage Manager server, or use the instance
configuration wizard, to use shared memory with DB2.

Immediate use of space that is added to the server database

When you add space to the database, new database directories are now available
for immediate use and parallel I/O performance is improved.

You can add directories to the database by using the EXTEND DBSPACE command. In
Version 7.1, the updates for this operation include distributing data across all
database directories and then reclaiming unused space and returning it to the
system. Because redistribution operations take considerable system resources, plan
ahead when you want to add space to the database. You must complete the
process while the server is not handling a heavy load.

For detailed requirements, see the EXTEND DBSPACE command in the Administrator’s
Reference.

As part of these enhancements, a new server utility, DSMSERV EXTEND DBSPACE, is

available to perform the same function as the EXTEND DBSPACE command, while the
server is offline.
Related tasks:
“Increasing the size of the database” on page 672

New for IBM Tivoli Storage Manager Version 7.1 xix

Server components that are not delivered with Version 7.1
Tivoli Monitoring for Tivoli Storage Manager and the Administration Center are
not delivered with Tivoli Storage Manager Version 7.1.
Tivoli Monitoring for Tivoli Storage Manager
You can use Tivoli Monitoring for Tivoli Storage Manager V6.3.4 with any
Tivoli Storage Manager V 5.5 or later server. For more information about
installing and configuring Tivoli Monitoring for Tivoli Storage Manager,
see Technote 1649883 at: [Link]
[Link]?uid=swg21649883
Administration Center
The Administration Center is not delivered in V7.1 or later releases. You
can use the V6.3.4 Administration Center with any Tivoli Storage Manager
V6.3 or later server, including to automatically update backup-archive
clients. For more information about the Tivoli Storage Manager
Administration Center, see the Tivoli Storage Manager V6.3 Information
Center at: [Link]
In V7.1, you can use the Operations Center, a web-based interface for
managing your storage environment. The Operations Center has many of
the same functions as the Administration Center. For more information
about the Operations Center, see Installing and upgrading the Operations
Center and Managing the storage environment with the Operations Center.

Restriction: The Operations Center cannot be installed on HP-UX or

Oracle Solaris systems. However, you can use the Operations Center to
manage Tivoli Storage Manager V6.3.4 or later servers that run on HP-UX
or Oracle Solaris systems.

New best practices

Learn about new best practices that you can implement for the Tivoli Storage
Manager server and client.

Duplicate identification processes and memory requirements

To get the best performance for your data deduplication processes, you can specify
an increased number of duplicate identification processes.

When you create a storage pool for data deduplication, you can now specify 0 - 50
parallel processes for server-side duplicate identification.

You can use additional memory to optimize the frequent access of deduplicate
extent information that is stored in the Tivoli Storage Manager database.

Server startup and setting ulimits

Before you start the Tivoli Storage Manager server on an IBM AIX®, HP-UX, Linux,
or Solaris operating system, verify access rights and user limits, also known as
ulimits.

If you do not verify user limits, the server might become unstable or fail to
respond. Set the ulimit value as 65536 for the maximum number of open files.
Ensure that the system-wide limit is at least the value of the ulimit.

xx IBM Tivoli Storage Manager for Windows: Administrator's Guide

Part 1. Tivoli Storage Manager basics

© Copyright IBM Corp. 1993, 2013 1

2 IBM Tivoli Storage Manager for Windows: Administrator's Guide
Chapter 1. Tivoli Storage Manager overview
IBM Tivoli Storage Manager is an enterprise-wide storage management application.
It provides automated storage management services to workstations, personal
computers, and file servers from various vendors, with various operating systems.

Tivoli Storage Manager includes the following components:

Server
Server program
The server program provides backup, archive, and space
management services to the clients.
You can set up multiple servers in your enterprise network to
balance storage, processor, and network resources.
Administrative interface
The administrative interface allows administrators to control and
monitor server activities, define management policies for clients,
and set up schedules to provide services to clients at regular
intervals.
Administrative interfaces include a command-line administrative
client and a web-based interface that is called the Operations
Center. With Tivoli Storage Manager you can manage and control
multiple servers from a single interface that runs in a web browser.
Server database and recovery log
The Tivoli Storage Manager server uses a database to track
information about server storage, clients, client data, policy, and
schedules. The server uses the recovery log as a scratch pad for the
database, recording information about client and server actions
while the actions are being performed.
Server storage
The server can write data to hard disk drives, disk arrays and
subsystems, stand-alone tape drives, tape libraries, and other forms
of random-access and sequential-access storage. The media that the
server uses are grouped into storage pools.
The storage devices can be connected directly to the server, or
connected through a local area network (LAN) or a storage area
network (SAN).
Client Nodes
A client node can be a workstation, a personal computer, a file server, or
even another Tivoli Storage Manager server. The client node has IBM Tivoli
Storage Manager client software that is installed and is registered with the
server.
Network-attached storage (NAS) file servers can also be client nodes, but
when you use NDMP, they do not have Tivoli Storage Manager client
software installed.
Backup-archive client
The backup-archive client allows users to maintain backup versions
of files, which they can restore if the original files are lost or

© Copyright IBM Corp. 1993, 2013 3

 damaged. Users can also archive files for long-term storage and
retrieve the archived files when necessary. Users themselves or
administrators can register workstations and file servers as client
nodes with a Tivoli Storage Manager server.
The storage agent is an optional component that can also be
installed on the same system as the client node. The storage agent
enables LAN-free data movement for client operations and is
supported on a number of operating systems.
Network-attached storage file server (by using NDMP)
The server can use the Network Data Management Protocol
(NDMP) to back up and restore file systems that are stored on a
network-attached storage (NAS) file server. The data on the NAS
file server is backed up to a tape library. The Tivoli Storage
Manager software does not need to be installed on the NAS file
server. A NAS file server can also be backed up over the LAN to a
Tivoli Storage Manager server. For more information, including
details about supported NAS file servers, see Chapter 9, “Using
NDMP for operations with NAS file servers,” on page 203.
Application client
Application clients allow users to initiate online backups of data
for applications such as database programs. After the application
program initiates a backup or restore, the application client acts as
the interface to Tivoli Storage Manager. The Tivoli Storage Manager
server then applies its storage management functions to the data.
The application client can initiate its functions while application
users are working, with minimal disruption.
The following products provide application clients for use with the
Tivoli Storage Manager server:
v Tivoli Storage Manager for Databases
v Tivoli Storage Manager for Enterprise Resource Planning
v Tivoli Storage Manager for Mail
Application programming interface (API)
The API can help you to enhance existing applications to use the
backup, archive, restore, and retrieve services that Tivoli Storage
Manager provides. Tivoli Storage Manager API clients can register
as client nodes with a Tivoli Storage Manager server.
Tivoli Storage Manager for Space Management
Tivoli Storage Manager for Space Management provides space
management services for workstations on some operating systems. The
space management function is essentially a more automated version of
archive. Tivoli Storage Manager for Space Management automatically
migrates files that are less frequently used to server storage, freeing space
on the workstation. The migrated files are also called space-managed files.
Users can recall space-managed files automatically by accessing them as
they normally would from the workstation. Tivoli Storage Manager for
Space Management is also known as the space manager client, or the
hierarchical storage management (HSM) client.
HSM for Windows
The Tivoli Storage Manager for HSM for Windows client provides
hierarchical storage management (HSM) for Windows NTFS file systems.

4 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 HSM is a data storage system that automatically moves data between
high-cost and low-cost storage media.
High-speed storage devices are more expensive per byte stored than slower
devices, such as magnetic tape drives. While it would be ideal to have all
data available on high-speed devices all the time, this is prohibitively
expensive for many organizations. HSM is available to store data on slower
devices, and then copy data to faster disk drives only when needed.

For information about supported operating systems for clients, see the IBM Tivoli
Storage Manager website at [Link]
Overview/Software/Tivoli/Tivoli_Storage_Manager.

Client programs, such as the backup-archive client and the HSM client (space
manager), are installed on systems that are connected through a LAN and that are
registered as client nodes. From these client nodes, users can back up, archive, or
migrate files to the server.

The following sections present key concepts and information about IBM Tivoli
Storage Manager. The sections describe how Tivoli Storage Manager manages client
files that are based on information that is provided in administrator-defined
policies, and manages devices and media that are based on information that is
provided in administrator-defined Tivoli Storage Manager storage objects.

The final section gives an overview of tasks for the administrator of the server,
including options for configuring the server and how to maintain the server.

How client data is stored

Tivoli Storage Manager policies are rules that determine how the client data is
stored and managed. The rules include where the data is initially stored, how
many backup versions are kept, how long archive copies are kept, and more.

You can have multiple policies and assign the different policies as needed to
specific clients, or files. Policy assigns a location in server storage where data is
initially stored. Server storage is divided into storage pools that are groups of
storage volumes and can include disk and tape volumes.

When you install Tivoli Storage Manager, you have a default policy that you can
use. For details about this default policy, see “Reviewing the standard policy” on
page 477. You can modify this policy and define policies.

Clients use Tivoli Storage Manager to store data for any of the following purposes:
Back up and restore
The backup process copies data from client workstations to server storage
to protect against data-loss. The server retains versions of a file according
to policy, and replaces older versions of the file with newer versions. The
policy includes the number of versions of files to keep and the retention
time for the versions.
A client can restore any version of a file that is kept on the Tivoli Storage
Manager server.
Archive and retrieve
The archive process copies data from client workstations to the Tivoli
Storage Manager server for long-term storage. The process can optionally
delete the archived files from the client workstations. The server retains

Chapter 1. Tivoli Storage Manager overview 5

 archive copies according to the policy for archive retention time. A client
can retrieve an archived copy of a file.
Instant archive and rapid recovery
Instant archive is the creation of a complete set of backed-up files for a
client. The set of files is called a backup set, which is created on the server
from the most recently backed-up files.
You can copy a backup set onto compatible portable media, which you can
take directly to the client for rapid recovery without the use of a network
and communicating with the Tivoli Storage Manager server.
Migration and recall
Migration is a function of the Tivoli Storage Manager for Space
Management program. Migration can free up client storage space by
copying files from workstations to the Tivoli Storage Manager server. On
the client, the Tivoli Storage Manager for Space Management program
replaces the original file with a stub file that points to the original file on
the Tivoli Storage Manager server. Files can be recalled to the workstations
when needed.
The migration process is also called hierarchical storage management
(HSM). After migration is configured, the process is transparent to the
users. Files are migrated and recalled automatically.
A policy determines when files are considered for automatic migration. On
UNIX or Linux systems that support the Tivoli Storage Manager for Space
Management program, the policies determine whether files must be backed
up to the server before it is migrated. Space management is also integrated
with backup. If the file to be backed up is already migrated to server
storage, the file is backed up from there.

Figure 1 on page 7 shows how a policy is part of the Tivoli Storage Manager
process for storing client data.

6 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 Client Server Client Storage Pools
Data
4

Migration
Backup
or
Archive

Database

Policy Domain

Policy Set

Management Class

Copy Group

Figure 1. How IBM Tivoli Storage Manager Controls Backup, Archive, and Migration
Processes

The steps in the process are as follows:

v
1 A client initiates a backup, archive, or migration operation. The file is bound
to a management class. The management class is either the default or one
specified for the file in the client options file or client options set.
v
2 Based on information in the management class, the client sends the file and
file information to the server. The information that is sent depends on the client
operation that was performed.
v
3 The server checks the management class that is bound to the file to
determine the destination storage pool where the file is stored. For backed-up
and archived files, the destination is stored in the backup and archive copy
groups within management classes. For space-managed files, the destination is
stored directly in the management class.
The storage pool can be a group of disk volumes, or tape volumes.
v
4 The server stores the file in the storage pool that is identified as the storage
destination.
The Tivoli Storage Manager server saves information in its database about each
file that it backs up, archives, or migrates.
If you set up server storage in a hierarchy, Tivoli Storage Manager can later
migrate the file to a storage pool different from the one where the file was

Chapter 1. Tivoli Storage Manager overview 7

 initially stored. For example, you might want to set up server storage so that
Tivoli Storage Manager migrates files from a disk storage pool to tape volumes
in a tape storage pool.

Files remain in the storage pools until they expire and expiration processing
occurs, or until they are deleted on the server. A file expires because of criteria that
are set in the policy. For example, the criteria include the number of file versions
that are allowed and the number of days that have elapsed since a file was deleted
from a client-node file system. If data retention protection is activated, an archive
object cannot be inadvertently deleted.

For information about assigning storage destinations in copy groups and

management classes, and on binding management classes to client files, see
Chapter 13, “Implementing policies for client data,” on page 475.

For information about managing the database, see Chapter 20, “Managing the
database and recovery log,” on page 643.

For information about storage pools and storage pool volumes, see Chapter 10,
“Managing storage pools and volumes,” on page 239.

For information about event-based policy, deletion hold, and data retention
protection, see Chapter 13, “Implementing policies for client data,” on page 475.

Data-protection options
Tivoli Storage Manager provides a variety of backup and archive operations,
allowing you to select the right protection for the situation.

Table 1 shows some examples of the protection options.

Table 1. Examples of meeting your goals with Tivoli Storage Manager
For this goal... Do this...
Back up files that are on a user's workstation, Use the backup-archive client to perform
and have the ability to restore individual incremental backups or selective backups.
files.
Back up a file server, and have the ability to Use the backup-archive client to perform
restore individual files. incremental backups or selective backups.

If the file server is a network-attached

storage file server that is supported, you can
have the server use NDMP to perform image
backups. This support is available in the
Tivoli Storage Manager Extended Edition
product.
Make restore media portable, or make Use the backup-archive client to perform
restores easier to perform remotely. incremental backups, and then generate
backup sets by using the Tivoli Storage
Manager server.
Provide the ability to more easily restore the Use the backup-archive client to perform
entire contents of a single logical volume, logical volume backups (also called image
instead of restoring individual files. backups).
Set up records retention to meet legal or Use the backup-archive client to occasionally
other long-term storage needs. perform archiving. To ensure that the
archiving occurs at the required intervals, use
central scheduling.

8 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Table 1. Examples of meeting your goals with Tivoli Storage Manager (continued)
For this goal... Do this...
Create an archive for a backup-archive client, Use the backup-archive client to perform
from data that is already stored for backup. incremental backups, and then generate a
backup set by using the Tivoli Storage
Manager server. This is also called instant
archive.
Provide the ability to restore data to a point Use the backup-archive client to regularly
in time. perform incremental backups (either
manually or automatically through
schedules). Then do one of the following:
v Set up policy to ensure that data is
preserved in server storage long enough to
provide the required service level. See
“Setting policy to enable point-in-time
restore for clients” on page 527 for details.
v Create backup sets for the backup-archive
client on a regular basis. Set the retention
time to provide the required service level.
See “Creating and using client backup
sets” on page 543 for details.
Save a set of files and directories before Use the backup-archive client to archive the
making significant changes to them. set of files and directories.

If this kind of protection is needed regularly,

consider creating backup sets from backup
data already stored for the client. Using
backup sets instead of frequent archive
operations can reduce the amount of
metadata that must be stored in the server's
database.
Manage a set of related files, which are not Use the backup group command on the
in the same file system, with the same backup-archive client to create a logical
backup, restore, and server policies. grouping of a set of files, which can be from
one or more physical file systems. The group
backup process creates a virtual file space in
server storage to manage the files, because
the files might not be from one file system
on the client. Actions such as policy binding,
migration, expiration, and export are applied
to the group as a whole. See the
Backup-Archive Clients Installation and User's
Guide for details.
Back up data for an application that runs Use the appropriate application client. For
continuously, such as a database application example, use Tivoli Storage Manager for Mail
(for example, DB2 or Oracle) or a mail to protect the Lotus Domino application.
application (Lotus® Domino®).
Exploit disk hardware capable of data Use the appropriate component in the IBM
snapshots. Tivoli Storage FlashCopy Manager product,
such as System Storage® Archive Manager for
IBM Enterprise Storage Server® for DB2.
Make backups transparent to end users. Use the backup-archive client with centrally
scheduled backups that run during off-shift
hours. Monitor the schedule results.

Chapter 1. Tivoli Storage Manager overview 9

 Table 1. Examples of meeting your goals with Tivoli Storage Manager (continued)
For this goal... Do this...
Reduce the load on the LAN by moving Use LAN-free data movement or, for
backup data over your SAN. supported network-attached storage (NAS)
file servers, use NDMP operations.

Schedule the backups of client data to help enforce the data management policy
that you establish. If you schedule the backups, rather than rely on the clients to
perform the backups, the policy that you establish is followed more consistently.
See Chapter 15, “Scheduling operations for client nodes,” on page 559.

The standard backup method that Tivoli Storage Manager uses is called progressive
incremental backup. It is a unique and efficient method for backup. See “Progressive
incremental backups” on page 13.

Table 2 summarizes the client operations that are available. In all cases, the server
tracks the location of the backup data in its database. Policy that you set
determines how the backup data is managed.
Table 2. Summary of client operations
Type of Description Usage Restore options For more
operation information
Progressive The standard method of Helps ensure complete, The user can restore just See “Incremental
incremental backup used by Tivoli effective, policy-based the version of the file that backup” on page 492
backup Storage Manager. After backup of data. Eliminates is needed. and the
the first, full backup of a the need to retransmit Backup-Archive
client system, incremental backup data that has not Tivoli Storage Manager Clients Installation
backups are done. been changed during does not need to restore a and User's Guide.
Incremental backup by successive backup base file followed by
date is also available. operations. incremental backups. This
means reduced time and
No additional full fewer tape mounts, as
backups of a client are well as less data
required after the first transmitted over the
backup. network.
Selective Backup of files that are Allows users to protect a The user can restore just See “Selective
backup selected by the user, subset of their data the version of the file that backup” on page 494
regardless of whether the independent of the is needed. and the
files have changed since normal incremental Backup-Archive
the last backup. backup process. Tivoli Storage Manager Clients Installation
does not need to restore a and User's Guide.
base file followed by
incremental backups. This
means reduced time and
fewer tape mounts, as
well as less data
transmitted over the
network.

10 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Table 2. Summary of client operations (continued)
Type of Description Usage Restore options For more
operation information
Adaptive A backup method that Maintains backups of data The base file plus a See “Enabling clients
subfile backs up only the parts of while minimizing connect maximum of one subfile to use subfile
backup a file that have changed time and data is restored to the client. backup” on page 552
since the last backup. The transmission for the and the
server stores the base file backup of mobile and Backup-Archive
(the complete initial remote users. Clients Installation
backup of the file) and and User's Guide.
subsequent subfiles (the
changed parts) that
depend on the base file.

The process works with

either the standard
progressive incremental
backup or with selective
backup.

Applicable to clients on
Windows systems.
Journal- Aids all types of backups Reduces the amount of Journal-based backup has See the
based (progressive incremental time required for backup. no effect on how files are Backup-Archive
backup backup, selective backup, The files eligible for restored; this depends on Clients Installation
adaptive subfile backup) backup are known before the type of backup and User's Guide.
by basing the backups on the backup operation performed.
a list of changed files. begins.
The list is maintained on
the client by the journal Applicable to clients on
engine service of IBM AIX and Windows
Tivoli Storage Manager. systems, except Windows
2003 64-bit IA64.
Image Full volume backup. Allows backup of an The entire image is See “Policy for
backup entire file system or raw restored. logical volume
Nondisruptive, on-line volume as a single object. backups” on page
backup is possible for Can be selected by 523 and the
Windows clients by using backup-archive clients on Backup-Archive
the Tivoli Storage UNIX, Linux, and Clients Installation
Manager snapshot Windows systems. and User's Guide.
function.
Image Full volume backup, Used only for the image The full image backup See Chapter 9,
backup which can be followed by backups of NAS file plus a maximum of one “Using NDMP for
with subsequent differential servers, performed by the differential backup are operations with NAS
differential backups. server using NDMP restored. file servers,” on page
backups operations. 203.
Backup A method of backup that Implements Details depend on the See the
using exploits the capabilities high-efficiency backup hardware. documentation for
hardware of IBM Enterprise Storage and recovery of Tivoli Storage
snapshot Server FlashCopy and business-critical FlashCopy Manager.
capabilities EMC TimeFinder to make applications while
copies of volumes used virtually eliminating
by database servers. The backup-related downtime
Tivoli Storage FlashCopy or user disruption on the
Manager product then database server.
uses the volume copies to
back up the database
volumes.

Chapter 1. Tivoli Storage Manager overview 11

Table 2. Summary of client operations (continued)
Type of Description Usage Restore options For more
operation information
Group A method that backs up Creates a consistent The user can select to See the
backup point-in-time backup of a
files that you specify as a restore the entire group or Backup-Archive
named group. The files group of related files. The just selected members of Clients Installation
can be from one or more files can reside in different the group. The user can and User's Guide.
file spaces. The backup file spaces on the client. restore just the version of
can be a full or a All objects in the group the file that is needed.
differential backup. are assigned to the same
management class. The
Applicable to clients on server manages the group
UNIX and Linux systems. as a single logical entity,
and stores the files in a
virtual file space in server
storage.

A group can be included

in a backup set.
Archive The process creates a Use for maintaining The selected version of See “Archive” on
copy of files and stores copies of vital records for the file is retrieved on page 495 and the
them for a specific time. legal or historical request. Backup-Archive
purposes. Clients Installation
Note: If you need to and User's Guide.
frequently create archives
for the same data,
consider using instant
archive (backup sets)
instead. Frequent archive
operations can create a
large amount of metadata
in the server database
resulting in increased
database growth and
decreased performance for
server operations such as
expiration. Frequently,
you can achieve the same
objectives with
incremental backup or
backup sets. Although the
archive function is a
powerful way to store
inactive data with fixed
retention, it should not be
used on a frequent and
large scale basis as the
primary backup method.

12 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Table 2. Summary of client operations (continued)
Type of Description Usage Restore options For more
operation information
Instant The process creates a Use when portability of The files are restored See “Creating and
archive backup set of the most the recovery media or directly from the backup using client backup
recent versions of the rapid recovery of a set. The backup set sets” on page 543.
files for the client, using backup-archive client is resides on media that can
files already in server important. Also use for be mounted on the client
storage from earlier efficient archiving. system, such as a CD, a
backup operations. tape drive, or a file
system. The Tivoli Storage
Manager server does not
have to be contacted for
the restore process, so the
process does not use the
network or the server.

Progressive incremental backups

The terms differential and incremental are often used to describe backups. The
standard method of backup used by Tivoli Storage Manager is progressive
incremental.

The terms differential and incremental have the following meanings:

v A differential backup backs up files that have changed since the last full backup.
– If a file changes after the full backup, the changed file is backed up again by
every subsequent differential backup.
– All files are backed up at the next full backup.
v An incremental backup backs up only files that have changed since the last
backup, whether that backup was a full backup or another incremental backup.
– If a file changes after the full backup, the changed file is backed up only by
the next incremental backup, not by all subsequent incremental backups.
– If a file has not changed since the last backup, the file is not backed up.

Tivoli Storage Manager takes incremental backup one step further. After the initial
full backup of a client, no additional full backups are necessary because the server,
using its database, keeps track of whether files need to be backed up. Only files
that change are backed up, and then entire files are backed up, so that the server
does not need to reference base versions of the files. This means savings in
resources, including the network and storage.

If you choose, you can force full backup by using the selective backup function of
a client in addition to the incremental backup function. You can also choose to use
adaptive subfile backup, in which the server stores the base file (the complete
initial backup of the file) and subsequent subfiles (the changed parts) that depend
on the base file.

Backup methods are summarized in Table 2 on page 10.

Chapter 1. Tivoli Storage Manager overview 13

 Storage-pool and server-database backups
Tivoli Storage Manager protects client data by letting you back up storage pools
and the database.

You can back up client backup, archive, and space-managed data in primary
storage pools to copy storage pools. You can also copy active versions of client
backup data from primary storage pools to active-data pools. The server can
automatically access copy storage pools and active-data pools to retrieve data. See
“Protecting client data” on page 882.

You can also back up the server's database. The database is key to the server's
ability to track client data in server storage. See “Protecting the database and
infrastructure setup files” on page 870.

These backups can become part of a disaster recovery plan, created automatically
by the disaster recovery manager. See Chapter 32, “Disaster recovery manager,” on
page 1005.

Data movement to server storage

Tivoli Storage Manager provides several methods for sending client data to server
storage.

In many configurations, the Tivoli Storage Manager client sends its data to the
server over the LAN. The server then transfers the data to a device that is attached
to the server. You can also use storage agents that are installed on client nodes to
send data over a SAN. This minimizes use of the LAN and the use of the
computing resources of both the client and the server. For details, see “LAN-free
data movement” on page 49.

For network-attached storage, use NDMP operations to avoid data movement over
the LAN. For details, see “NDMP backup operations” on page 51.

Consolidation of backed-up client data

By grouping the backed-up data for a client, you can minimize the number of
media mounts required for client recovery.

The server offers you methods for doing this:

Collocation
The server can keep each client's files on a minimal number of volumes
within a storage pool. Because client files are consolidated, restoring
collocated files requires fewer media mounts. However, backing up files
from different clients requires more mounts.
You can have the server collocate client data when the data is initially
stored in server storage. If you have a storage hierarchy, you can also have
the data collocated when the server migrates the data from the initial
storage pool to the next storage pool in the storage hierarchy.
Another choice you have is the level of collocation. You can collocate by
client, by file space per client, or by group. Your selection depends on the
size of the file spaces being stored and the restore requirements.
See “Keeping client files together using collocation” on page 357.
Active-data pools
Active-data pools are storage pools that contain only the active versions of

14 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 client backup data. Archive data and data migrated by Hierarchical Space
Management (HSM) clients are not allowed in active-data pools.
Active-data pools can be associated with three types of devices:
sequential-access disk (FILE), removable media (tape), or sequential-access
volumes on another Tivoli Storage Manager server. There are three types of
active-data pool, each of which has distinct advantages. For example, an
active-data pool associated with sequential-access disk is particularly
well-suited for fast restores of client data because tapes do not have to be
mounted and because the server does not have to position past inactive
files.
For more information, see “Backing up primary storage pools” on page
883.
Backup set creation
You can generate a backup set for each backup-archive client. A backup set
contains all active backed-up files that currently exist for that client in
server storage. The process is also called instant archive.
The backup set is portable and is retained for the time that you specify.
Creation of the backup set consumes more media because it is a copy in
addition to the backups that are already stored.
See “Creating and using client backup sets” on page 543.
Moving data for a client node
You can consolidate data for a client node by moving the data within
server storage. You can move it to a different storage pool, or to other
volumes in the same storage pool.
See “Moving data belonging to a client node” on page 406.

How the server manages storage

Through the server, you manage the devices and media used to store client data.
The server integrates the management of storage with the policies that you define
for managing client data.

Device support for server storage

With Tivoli Storage Manager, you can use direct-attached storage devices and
network-attached devices for server storage.

Tivoli Storage Manager represents physical storage devices and media with the
following administrator-defined objects:
Library
A library is one or more drives (and possibly robotic devices) with similar
media-mounting requirements.
Drive Each drive represents a drive mechanism in a tape device.
Data mover
A data mover represents a device that accepts requests from Tivoli Storage
Manager to transfer data on behalf of the server. Data movers transfer data
between storage devices.
Path A path represents how a source accesses a destination. For example, the
source can be a server, and the destination can be a tape drive. A path
defines the one-to-one relationship between a source and a destination.
Data can flow from the source to the destination, and back.

Chapter 1. Tivoli Storage Manager overview 15

 Device class
Each device is associated with a device class that specifies the device type
and how the device manages its media.
Storage pools and volumes
A storage pool is a named collection of volumes that have the same media
type. A storage pool is associated with a device class. A storage pool
volume is associated with a specific storage pool.
For example, an LTO tape storage pool contains only LTO tape volumes.

For details about device concepts, see Chapter 3, “Storage device concepts,” on
page 35.

Data migration through the storage hierarchy

You can organize the server's storage pools into one or more hierarchical
structures. This storage hierarchy allows flexibility in a number of ways. For
example, you can set policy to have clients send their backup data to disks for
faster backup operations, then later have the server automatically migrate the data
to tape.

See “Storage pool hierarchies” on page 260.

Removal of expired data

A policy that you define controls when client data automatically expires from the
Tivoli Storage Manager server. The expiration process is how the server
implements the policy.

For example, you have a backup policy that specifies that three versions of a file be
kept. File A is created on the client, and backed up. Over time, the user changes
file A, and three versions of the file are backed up to the server. Then the user
changes file A again. When the next incremental backup occurs, a fourth version of
file A is stored, and the oldest of the four versions is eligible for expiration.

To remove data that is eligible for expiration, a server expiration process marks
data as expired and deletes metadata for the expired data from the database. The
space occupied by the expired data is then available for new data.

You control the frequency of the expiration process by using a server option, or
you can start the expiration processing by command or scheduled command.

See “Running expiration processing to delete expired files” on page 512.

Media reuse by reclamation

As server policies automatically expire data, the media where the data is stored
accumulates unused space. The Tivoli Storage Manager server implements a
process, called reclamation, which allows you to reuse media without traditional
tape rotation.

Reclamation is a server process that automatically defragments media by

consolidating unexpired data onto other media when the free space on media
reaches a defined level. The reclaimed media can then be used again by the server.
Reclaiming media allows the automated circulation of media through the storage
management process. Use of reclamation can help minimize the number of media
that you need to have available.

16 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Chapter 2. Tivoli Storage Manager concepts
The server comes with many defaults so that you can begin using its services
immediately. The amount and importance of the data protected by Tivoli Storage
Manager, your business process requirements, and other factors make it likely that
you need to adjust and customize the server's behavior.

Your changing storage needs and client requirements can mean on-going
configuration changes and monitoring. The server's capabilities are described in the
following topics.

Interfaces to Tivoli Storage Manager

You can work with many different applications by using Tivoli Storage Manager
interfaces.

The following interfaces are provided:

v Graphical user interfaces
For the clients, there are graphical user interfaces for the backup-archive client
and the space manager client (if installed, on supported operating systems). For
information about using the interfaces, see the online information or the
Installation Guide.
v The Windows server includes a special interface, the IBM Tivoli Storage
Manager for Windows program folder.
v Web interfaces for server administration and for the backup-archive client
The Operations Center provides web and mobile access to status information
about the Tivoli Storage Manager environment. You can use this interface to
monitor multiple servers and complete some administrative tasks. The interface
also provides web access to the command-line. See Chapter 17, “Managing the
storage environment with the Operations Center,” on page 587 for more
information.
The web backup-archive client (web client) allows an authorized user to
remotely access a client to run backup, archive, restore, and retrieve processes.
The web browser must have the appropriate support for Oracle Java™. See the
Backup-Archive Clients Installation and User's Guide for requirements.
v The command-line interface
For information about using the command-line interface of the administrative
client, see the Administrator's Reference. For information about using the
command-line interface of the backup-archive client or other clients, see the
user's guide for that client.
v The application programming interface
For more information, see the IBM Tivoli Storage Manager Using the Application
Program Interface.
v Access to information in the server database using standard SQL SELECT
statements. Tivoli Storage Manager Versions 6.1 and later use the DB2 open
database connectivity (ODBC) driver to query the database and display the
results. For more information, see “Using SQL to query the IBM Tivoli Storage
Manager database” on page 794.

© Copyright IBM Corp. 1993, 2013 17

Storage configuration and management
Configuring and managing storage for efficiency, and capacity are important tasks
for an administrator.

The Tivoli Storage Manager server uses its storage for the data it manages for
clients. The storage can be a combination of devices.
v Disk
v Tape drives that are either manually operated or automated
v Other drives that use removable media

Devices can be locally attached, or accessible through a storage area network

(SAN). When you configure and manage the storage, consider the following tasks:
v Selecting the devices and media that will form the server storage, and whether
library drives want to be shared among Tivoli Storage Manager servers.
v Designing the storage hierarchy for efficient backups and optimal storage usage.
v Using product features that allow the server to provide services to clients while
minimizing traffic on the communications network:
– LAN-free data movement
– Data movement by using NDMP to protect data on network-attached storage
(NAS) file servers when backups occur to libraries directly attached to the
NAS file servers
v Using the Tivoli Storage Manager product to help you to manage the drives and
media, or by using an external media manager to do the management outside of
the Tivoli Storage Manager product.

For an introduction to storage concepts, see Chapter 3, “Storage device concepts,”

on page 35.

Disk devices
Disk devices can be used with Tivoli Storage Manager for storing the database and
recovery log or client data that is backed up, archived, or migrated from client
nodes.

The server can store data on disk by using random-access volumes (device type of
DISK) or sequential-access volumes (device type of FILE).

The Tivoli Storage Manager product allows you to exploit disk storage in ways
that other products do not. You can have multiple client nodes back up to the
same disk storage pool at the same time, and still keep the data for the different
client nodes separate. Other products also allow you to back up different systems
at the same time, but only by interleaving the data for the systems, leading to
slower restore processes.

If you have enough disk storage space, data can remain on disk permanently or
temporarily, depending on the amount of disk storage space that you have. Restore
process performance from disk can be very fast compared to tape.

You can have the server later move the data from disk to tape; this is called
migration through the storage hierarchy. Other advantages to this later move to
tape include:
v Ability to collocate data for clients as the data is moved to tape
v Streaming operation of tape drives, leading to better tape drive performance

18 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 v More efficient use of tape drives by spreading out the times when the drives are
in use

For information about storage hierarchy and setting up storage pools on disk
devices, see:
Chapter 4, “Magnetic disk devices,” on page 63 and “Storage pool hierarchies”
on page 260

Removable media devices

Removable media devices can be used with Tivoli Storage Manager for storage of
client data that is backed up, archived, or migrated from client nodes; storage of
database backups; and the exporting, that is, moving, of data to another server.

The following topics provide an overview of how to use removable media devices
with Tivoli Storage Manager.

For guidance and scenarios on configuring your tape devices, see:

Chapter 6, “Configuring storage devices,” on page 85

Device classes
A device class represents a set of storage devices with similar availability,
performance, and storage characteristics.

You must define device classes for the drives available to the Tivoli Storage
Manager server. You specify a device class when you define a storage pool so that
the storage pool is associated with drives.

For more information about defining device classes, see “Defining device classes”
on page 177.

Removable media operations

Routine removable media operations include preparing and controlling media for
reuse, ensuring that sufficient media are available, and mounting volumes in
response to server requests, for manually operated drives. Removable media
operations also include managing libraries and drives.

For information about removable media operations, see:

Chapter 7, “Managing removable media operations,” on page 127

Migrating data from disk to tape

After you set up disk and tape storage pools, you can configure the server so that
client data can be migrated to tape. By migrating data to tape from a disk storage
pool, you can verify that tape devices are properly set up.

About this task

The server can perform migration as needed, based on migration thresholds that
you set for the storage pools. Because migration from a disk to a tape storage pool
uses resources such as drives and operators, you may want to ensure that
migration occurs at a time that is best for your situation. You can control when
migration occurs by setting migration thresholds with the DEFINE STGPOOL
command.

Chapter 2. Tivoli Storage Manager concepts 19

 You might not want to empty the disk storage pool every time migration occurs by
setting the low migration threshold to 0. You can typically keep the low threshold
at 40%, and vary the high threshold from as high as 90% to as low as 50%.

To migrate from a disk storage pool to a tape storage pool, devices must be
allocated and tapes must be mounted. Mount messages are directed to the console
message queue and to any administrative client that is started with either the
mount mode or console mode option.

See “Migrating disk storage pools” on page 273 and the Administrator's Reference
for more information.

Techniques for managing your storage pools and volumes

Backed-up, archived, and space-managed files are stored in groups of volumes that
are called storage pools.

About this task

You can use the following techniques to manage your storage pool data:
Collocation
The server can keep data that belongs to a single client node, group of
client nodes, or client file space assigned to as few volumes as possible.
When you restore collocated files, fewer media mounts is needed, however,
when you back up files from different clients, more mounts are needed.
Reclamation
Reclamation makes the fragmented space on volumes usable again by
moving any remaining active files from one volume to another volume,
thus making the original volume available for reuse.
Storage pool backup
Client backup, archive, and space-managed data in primary storage pools
can be backed up to copy storage pools for disaster recovery purposes. You
can configure the storage pool to simultaneously write to copy storage
pools when client data is written to the primary storage pools.
Copy active data
The active versions of client backup data can be copied to active-data
pools. Active-data pools provide a number of benefits. For example, if the
device type associated with an active-data pool is sequential-access disk
(FILE), you can eliminate the need for disk staging pools. Restoring client
data is faster because FILE volumes are not physically mounted, and the
server does not have to position past inactive files that do not have to be
restored.
An active-data pool that uses removable media, such as tape, reduces the
number of volumes for onsite and offsite storage. (Like volumes in copy
storage pools, volumes in active-data pools can be moved offsite for
protection in case of disaster.) If you vault data electronically to a remote
location, a SERVER-type active-data pool saves bandwidth by copying and
restoring only active data.
As backup client data is written to primary storage pools, the active
versions can be simultaneously written to active-data pools.
Cache When the server migrates files from disk storage pools, duplicate copies of
the files can remain in cache (disk storage) for faster retrieval. Cached files

20 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 are deleted only when space is needed. However, client backup operations
that use the disk storage pool can have poorer performance.

You can establish a hierarchy of storage pools. The hierarchy can be based on the
speed or the cost of the devices that are associated with the pools. Tivoli Storage
Manager can migrate client files through this hierarchy to ensure the most efficient
use of storage devices.

You manage storage volumes by defining, updating, and deleting volumes, and by
monitoring the use of server storage. You can also move files within and across
storage pools to optimize the use of server storage.

For more information about storage pools and volumes and taking advantage of
storage pool features, see Chapter 10, “Managing storage pools and volumes,” on
page 239.

Increasing disk storage-pool sizes

When the server is installed, Tivoli Storage Manager creates default disk storage
pools. You must define volumes for these storage pools.

About this task

At server installation, the following storage pools are created:

v BACKUPPOOL
v ARCHIVEPOOL
v SPACEMGPOOL
Before defining storage-pool volumes, estimate how much client data you want to
store. Consider defining at least as many volumes as you need to hold one day of
client backups. To format and define storage pool volumes, use the DEFINE VOLUME
command.

Tip: To define disk volumes, you can also use the Server Initialization wizard. This
wizard is displayed during the server configuration process.

Management of client operations

Because the key task of the server is to provide services to clients, many of the
server administrator's tasks deal with client operations.

Tasks include the following:

v Registering clients and customizing client operations
v Ensuring that client operations meet security requirements
v Providing required levels of service by customizing policies
v Automating protection by using schedules

After you have created schedules, you manage and coordinate those schedules.
Your tasks include the following:
v Verify that the schedules ran successfully.
v Determine how long Tivoli Storage Manager retains information about schedule
results (event records) in the database.
v Balance the workload on the server so that all scheduled operations complete.
For more information about client operations, see the following sections:

Chapter 2. Tivoli Storage Manager concepts 21

 v For setting up an include-exclude list for clients, see “Getting users started” on
page 478.
v For automating client operations, see Chapter 15, “Scheduling operations for
client nodes,” on page 559.
v For running the scheduler on a client system, see the user's guide for the client.
v For setting up policy domains and management classes, see Chapter 13,
“Implementing policies for client data,” on page 475.

For more information about these tasks, see Chapter 16, “Managing schedules for
client nodes,” on page 569

Managing client nodes

A basic administrative task is adding client nodes and giving the systems that the
nodes represent access to the services and resources of the IBM Tivoli Storage
Manager server.

The Tivoli Storage Manager server supports a variety of client nodes. You can
register the following types of clients and servers as client nodes:
v Tivoli Storage Manager backup-archive client
v Application clients that provide data protection through one of the following
products: Tivoli Storage Manager for Application Servers, Tivoli Storage
Manager for Databases, Tivoli Storage Manager for Enterprise Resource
Planning, or Tivoli Storage Manager for Mail.
v Tivoli Storage Manager for Space Management client (called space manager
client or HSM client)
v A NAS file server for which the Tivoli Storage Manager server uses NDMP for
backup and restore operations
v Tivoli Storage Manager source server (registered as a node on a target server)

When you register clients, you have choices to make about the following:
v Whether the client should compress files before sending them to the server for
backup
v Whether the client node ID has the authority to delete its files from server
storage
v Whether an administrator ID that matches the client ID is created, for remote
client operations

Other important tasks include the following:

Controlling client options from the server
Client options on client systems allow users to customize backup, archive,
and space management operations, as well as schedules for these
operations. On most client systems, the options are in a file called [Link].
In some cases, you may need or want to provide the clients with options to
use. To help users get started, or to control what users back up, you can
define sets of client options for clients to use. Client options sets are
defined in the server database and are used by the clients that you
designate.
Among the options that can be in a client option set are the include and
exclude options. These options control which files are considered for the
client operations.
For more information, see:

22 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 v Chapter 11, “Adding client nodes,” on page 419
v Chapter 12, “Managing client nodes,” on page 431
Allowing subfile backups

Restriction: Subfile backups are only applicable to Windows clients.

For mobile and remote users, you want to minimize the data sent over the
network as well as the time that they are connected to the network. You
can set the server to allow a client node to back up changed portions of
files that have been previously backed up, rather than entire files. The
portion of the file that is backed up is called a subfile. For more
information, see Chapter 14, “Managing data for client nodes,” on page
535.
Creating backup sets for client nodes
You can perform an instant archive for a client by creating a backup set. A
backup set copies a client node's active, backed-up files from server storage
onto sequential media. If the sequential media can be read by a device
available to the client system, you can restore the backup set directly to the
client system without using the network. The server tracks backup sets that
you create and retains the backup sets for the time you specify.
For more information, see Chapter 14, “Managing data for client nodes,”
on page 535.

For more information on managing client nodes, see the Backup-Archive Clients
Installation and User's Guide.

Security management
Tivoli Storage Manager includes security features for user registration and
passwords. Also included are features that can help ensure security when clients
connect to the server across a firewall.

Registration for clients can be closed or open. With closed registration, a user with
administrator authority must register all clients. With open registration, clients can
register themselves at first contact with the server. See “Registering nodes with the
server” on page 420.

You can ensure that only authorized administrators and client nodes are
communicating with the server by requiring passwords. Passwords can
authenticate with an LDAP directory server or the Tivoli Storage Manager server.
Most password-related commands work for both kinds of servers. The PASSEXP and
RESET PASSEXP commands do not work for passwords that authenticate with an
LDAP directory server. You can use the LDAP directory server to give more
options to your passwords, independent of the Tivoli Storage Manager server.

Restriction: The Tivoli Storage Manager backup-archive client must be at version

6.4.0 or later to use passwords that authenticate with an LDAP directory server.

Whether you store your passwords on an LDAP directory server, or on the Tivoli
Storage Manager server, you can set the following requirements for passwords:
v Minimum number of characters in a password.
v Expiration time.

Chapter 2. Tivoli Storage Manager concepts 23

 v A limit on the number of consecutive, invalid password attempts. When the
client exceeds the limit, Tivoli Storage Manager stops the client node from
accessing the server. The limit can be set on the Tivoli Storage Manager server,
and on the LDAP directory server.

Important: The invalid password limit is for passwords that authenticate with the
Tivoli Storage Manager server and any LDAP directory servers. Invalid password
attempts can be configured on an LDAP directory server, outside of the Tivoli
Storage Manager server. But the consequence of setting the number of invalid
attempts on the LDAP directory server might pose some problems. For example,
when the REGISTER NODE command is issued, the default behavior is to name the
node administrator the same name as the node. The LDAP server does not
recognize the difference between the node “NODE_Q” and the administrator
“NODE_Q”. The node and the administrator can authenticate to the LDAP server
if they have the same password. If the node and administrator have different
passwords, the authentication fails for either the node or administrator. If the node
or the administrator fail to logon consistently, their IDs are locked. You can avoid
this situation by issuing the REGISTER NODE command with USERID=userid or
USERID=NONE.

Tivoli Storage Manager provides several ways to manage security. You can control
the authority of administrators. An organization can name a single administrator
or distribute the workload among a number of administrators and grant them
different levels of authority. For better security when clients connect across a
firewall, you can control whether clients can initiate contact with the server for
scheduled operations. Server options can be set to keep client and administrative
traffic on separate server ports.
Related concepts:
“Managing Tivoli Storage Manager administrator IDs” on page 844
“Managing passwords and logon procedures” on page 852
Related reference:
Chapter 28, “Managing Tivoli Storage Manager security,” on page 831

Managing client data with policies

As the administrator, you define the rules for client backup, archive, and migration
operations, based on user or business requirements.

About this task

The rules are called policies. Policies identify:

v The criteria for backup, archive, and migration of client data
v Where the client data is initially stored
v How the data is managed by the server (how many backup versions are kept,
for how long)

In Tivoli Storage Manager, you define policies by defining policy domains, policy
sets, management classes, and backup and archive copy groups. When you install
Tivoli Storage Manager, you have a default policy that consists of a single policy
domain named STANDARD.

The default policy provides basic backup protection for end-user workstations. To
provide different levels of service for different clients, you can add to the default
policy or create new policy. For example, because of business needs, file servers are

24 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 likely to require a policy different from policy for users' workstations. Protecting
data for applications such as Lotus Domino also may require a unique policy.

For more information about the default policy and establishing and managing new
policies, see Chapter 13, “Implementing policies for client data,” on page 475.

Schedules for client operations

Scheduling client operations can mean better protection for data, because
operations can occur consistently without user intervention.

Scheduling also can mean better utilization of resources such as the network.
Client backups that are scheduled at times of lower usage can minimize the impact
on user operations on a network.

You can automate operations for clients by using schedules. Tivoli Storage
Manager provides a central scheduling facility. You can also use operating system
utilities or other scheduling tools to schedule Tivoli Storage Manager operations.

With Tivoli Storage Manager schedules, you can perform the operations for a client
immediately or schedule the operations to occur at regular intervals.

The key objects that interact are:

Include-exclude options on each client
The include-exclude options determines which files are backed up,
archived, or space-managed, and determines management classes,
encryption, and type of backup for files.
The client can specify a management class for a file or group of files, or
can use the default management class for the policy domain. The client
specifies a management class by using an INCLUDE option in the client's
include-exclude list or file. You can have central control of client options
such as INCLUDE and EXCLUDE by defining client option sets on the
server. When you register a client, you can specify a client option set for
that client to use. See “Managing client option files” on page 464 for
details.
Association defined between client and schedule
Associations determine which schedules are run for a client.
Clients are assigned to a policy domain when they are registered. To
automate client operations, you define schedules for a domain. Then you
define associations between schedules and clients in the same domain.
Schedule
The schedule determines when a client operation automatically occurs.
Schedules that can automate client operations are associated with a policy
domain.
The scheduled client operations are called events. The Tivoli Storage
Manager server stores information about events in its database. For
example, you can query the server to determine which scheduled events
completed successfully and which failed.
Management class
The management class determines where client files are initially stored and
how they are managed.

Chapter 2. Tivoli Storage Manager concepts 25

 The management class contains information that determines how Tivoli
Storage Manager handles files that clients backup, archive, or migrate. For
example, the management class contains the backup copy group and the
archive copy group. Each copy group points to a destination, a storage pool
where files are first stored when they are backed up or archived.

For a schedule to work on a particular client, the client machine must be turned
on. The client either must be running the client scheduler or must allow the client
acceptor daemon to start the scheduler when needed.
Related tasks:
“Scheduling a client operation” on page 560

Server maintenance
If you manage more than one server, you can ensure that the multiple servers are
consistently managed by using the enterprise management functions of Tivoli
Storage Manager.

You can set up one server as the configuration manager and have other servers
obtain configuration information from it.

To keep the server running well, you can perform these tasks:
v Managing server operations, such as controlling client access to the server
v Automating repetitive administrative tasks
v Monitoring and adjusting space for the database and the recovery log
v Monitoring the status of the server, server storage, and clients

Server-operation management
When managing your server operations, you can choose from a variety of
associated tasks.

Some of the more common tasks that you can perform to manage your server
operations are shown in the following list:
v Start and stop the server.
v Allow and suspend client sessions with the server.
v Query, cancel, and preempt server processes such as backing up the server
database.
v Customize server options.

Other tasks that are needed less frequently include:

v Maintain compliance with the license agreement.
v Move the server.

See “Licensing IBM Tivoli Storage Manager” on page 595. For suggestions about
the day-to-day tasks required to administer the server, see Chapter 18, “Managing
server operations,” on page 595.

26 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 Server options
Server options let you customize the server and its operations.

Server options can affect the following:

v Server communications
v Storage
v Database and recovery log operations
v Client transaction performance

Server options are in the server options file. Some options can be changed and
made active immediately by using the SETOPT command. Most server options are
changed by editing the server options file and then halting and restarting the
server to make the changes active.

See the Administrator's Reference for details about the server options file and
reference information for all server options.

Server script automation

Repetitive, manual tasks associated with managing the server can be automated
through Tivoli Storage Manager schedules and scripts. Using schedules and scripts
can minimize the daily tasks for administrators.

About this task

You can define schedules for the automatic processing of most administrative
commands. For example, a schedule can run the command to back up the server's
database every day.

Tivoli Storage Manager server scripts allow you to combine administrative

commands with return code checking and processing. The server comes with
scripts that you can use to do routine tasks, or you can define your own. The
scripts typically combine several administrative commands with return code
checking, or run a complex SQL SELECT command.

Results

For more information about automating Tivoli Storage Manager operations, see
Chapter 19, “Automating server operations,” on page 619.

Modifying a maintenance script

You can modify your maintenance script to add, subtract, or reposition commands.

Before you begin

If you have a predefined maintenance script, you can add or subtract commands
using the maintenance script wizard. You can add, subtract, or reposition
commands if you have a custom maintenance script. Both methods can be accessed
through the same process. If you want to convert your predefined maintenance
script to a custom maintenance script, select a server with the predefined script,
click Select Action > Convert to Custom Maintenance Script.

About this task

Perform the following tasks to modify a maintenance script:

Chapter 2. Tivoli Storage Manager concepts 27

 Procedure
1. Click Server Maintenance in the navigation tree.
2. Select a server that has either Predefined or Custom designated in the
Maintenance Script column.
3. Click Select Action > Modify Maintenance Script. If you are modifying a
predefined maintenance script, the maintenance script wizard opens your script
for you to modify. If you are modifying a custom maintenance script, the
maintenance script editor opens your script so that you can modify it.

Database and recovery-log management

The Tivoli Storage Manager database contains information about registered client
nodes, policies, schedules, and the client data in storage pools. The database is key
to the operation of the server.

The information about the client data, also called metadata, includes the file name,
file size, file owner, management class, copy group, and location of the file in
server storage. The server records changes made to the database (database
transactions) in its recovery log. The recovery log is used to maintain the database
in a transactionally consistent state, and to maintain consistency across server
startup operations.

For more information about the Tivoli Storage Manager database and recovery log
and about the tasks associated with them, see Chapter 20, “Managing the database
and recovery log,” on page 643.

Sources of information about the server

Tivoli Storage Manager provides you with many sources of information about
server and client status and activity, the state of the server database and storage,
and resource usage. By monitoring selected information, you can provide reliable
services to users while you are making the best use of available resources.

The Operations Center provides visibility to the key aspects of your Tivoli Storage
Manager environment, including clients, servers, policy, storage pools, and storage
devices. See Chapter 17, “Managing the storage environment with the Operations
Center,” on page 587 for more information.

You can use Tivoli Storage Manager queries and SQL queries to get information
about the server. You can also set up automatic logging of information about Tivoli
Storage Manager clients and server events. Daily checks of some indicators are
suggested.

See the following sections for more information about these tasks:
v Part 5, “Monitoring operations,” on page 775
v Chapter 23, “Daily monitoring tasks,” on page 777
v “Using SQL to query the IBM Tivoli Storage Manager database” on page 794
v Chapter 27, “Logging IBM Tivoli Storage Manager events to receivers,” on page
807

28 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Tivoli Storage Manager server networks
You might have a number of Tivoli Storage Manager servers in your network, at
the same or different locations.

About this task

Some examples of different configurations are:

v Your users are scattered across many locations, so you have located Tivoli
Storage Manager servers close to the users to manage network bandwidth
limitations.
v You have set up multiple servers to provide services to different organizations at
one location.
v You have multiple servers on your network to make disaster recovery easier.

Servers connected to a network can be centrally managed. Tivoli Storage Manager

provides functions to help you configure, manage, and monitor the servers. An
administrator working at one Tivoli Storage Manager server can work with servers
at other locations around the world.

When you have a network of Tivoli Storage Manager servers, you can simplify
configuration and management of the servers by using enterprise administration
functions. You can do the following:
v Designate one server as a configuration manager that distributes configuration
information such as policy to other servers. See “Setting up enterprise
configurations” on page 701.
v Route commands to multiple servers while logged on to one server. See
“Routing commands” on page 726.
v Log events such as error messages to one server. This allows you to monitor
many servers and clients from a single server. See “Enterprise event logging:
logging events to another server” on page 820.
v Store data for one Tivoli Storage Manager server in the storage of another Tivoli
Storage Manager server. The storage is called server-to-server virtual volumes.
See “Using virtual volumes to store data on another server” on page 732 for
details.
v Share an automated library among Tivoli Storage Manager servers. See “Devices
on storage area networks” on page 47.
v Store a recovery plan file for one server on another server, when using disaster
recovery manager. You can also back up the server database and storage pools to
another server. See Chapter 32, “Disaster recovery manager,” on page 1005 for
details.
v Back up the server database and storage pools to another server. See “Using
virtual volumes to store data on another server” on page 732 for details.
v To simplify password management, have client nodes and administrators
authenticate their passwords on multiple servers using an LDAP directory
server. See “Managing passwords and logon procedures” on page 852.

Chapter 2. Tivoli Storage Manager concepts 29

 Exporting and importing data
As conditions change, you can move data from one server to another by using
export and import processes.

About this task

For example, you may need to balance workload among servers by moving client
nodes from one server to another. The following methods are available:
v You can export part or all of a server's data to sequential media, such as tape or
a file on hard disk. You can then take the media to another server and import
the data to that server
v You can export part or all of a server's data and import the data directly to
another server, if server-to-server communications are set up.

For more information about moving data between servers, see Chapter 22,
“Exporting and importing data,” on page 741.

Protecting Tivoli Storage Manager and client data

The database, recovery log, and storage pools are critical to the operation of the
server and must be properly protected.

About this task

Attention: If the database is unusable, the entire Tivoli Storage Manager server is
unavailable. If a database is lost and cannot be recovered, it might be difficult or
impossible to recover data that is managed by that server. Therefore, It is critically
important to back up the database. However, even without the database, fragments
of data or complete files might easily be read from storage pool volumes that are
not encrypted. Even if data is not completely recovered, security can be
compromised. For this reason, always encrypt sensitive data by using the Tivoli
Storage Manager client or the storage device, unless the storage media is physically
secured. See Part 6, “Protecting the server,” on page 829 for steps that you can take
to protect your database.

IBM Tivoli Storage Manager provides a number of ways to protect your data,
including backing up your storage pools and database. For example, you can
define schedules so that the following operations occur:
v After the initial full backup of your storage pools, incremental storage pool
backups are done nightly.
v Full database backups are done weekly.
v Incremental database backups are done nightly.

In addition, disaster recovery manager (DRM), an optional feature of Tivoli Storage

Manager, can assist you in many of the tasks that are associated with protecting
and recovering your data. For details, see Chapter 32, “Disaster recovery manager,”
on page 1005.

30 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Protecting the server
Tivoli Storage Manager provides a number of ways to protect and recover your
server from media failure or from the loss of the Tivoli Storage Manager database
or storage pools.

About this task

Recovery is based on the following preventive measures:

v Mirroring, by which the server maintains a copy of the active log
v Periodic backup of the database
v Periodic backup of the storage pools
v Audit of storage pools for damaged files, and recovery of damaged files when
necessary
v Backup of the device configuration and volume history files
v Validation of the data in storage pools, using cyclic redundancy checking
v Keeping the [Link] file in a safe place to ensure that SSL (Secure Sockets
Layer) is secure for authenticating passwords
For information about protecting the server with these measures, see Chapter 29,
“Protecting and recovering the server infrastructure and client data,” on page 869.

In addition to taking these actions, you can prepare a disaster recovery plan to
guide you through the recovery process by using the disaster recovery manager,
which is available with Tivoli Storage Manager Extended Edition. The disaster
recovery manager (DRM) assists you in the automatic preparation of a disaster
recovery plan. You can use the disaster recovery plan as a guide for disaster
recovery as well as for audit purposes to certify the recoverability of the Tivoli
Storage Manager server.

The disaster recovery methods of DRM are based on taking the following
measures:
v Sending server backup volumes offsite or to another Tivoli Storage Manager
server
v Creating the disaster recovery plan file for the Tivoli Storage Manager server
v Storing client machine information
v Defining and tracking client recovery media

For more information about protecting your server and for details about recovering
from a disaster, see Chapter 29, “Protecting and recovering the server infrastructure
and client data,” on page 869.

Chapter 2. Tivoli Storage Manager concepts 31

32 IBM Tivoli Storage Manager for Windows: Administrator's Guide
Part 2. Configuring and managing storage devices
To make storage devices available to a server, you must attach the devices,
configure device drivers, and create Tivoli Storage Manager objects that represent
those devices. To use the devices to store client data, you also create objects
representing storage pools and storage-pool volumes.

© Copyright IBM Corp. 1993, 2013 33

34 IBM Tivoli Storage Manager for Windows: Administrator's Guide
Chapter 3. Storage device concepts
To work with storage devices, you must be familiar with Tivoli Storage Manager
storage objects and other basic concepts.

“Tivoli Storage Manager storage devices” on page 36

“Tivoli Storage Manager storage objects” on page 36
“Tivoli Storage Manager volumes” on page 46
“Planning for server storage” on page 60
“Device configurations” on page 47
“Removable media mounts and dismounts” on page 54
“How Tivoli Storage Manager uses and reuses removable media” on page 55
“Definitions for storage devices” on page 58

The examples in topics show how to perform tasks using the Tivoli Storage
Manager command-line interface. For information about the commands, see the
Administrator's Reference, or issue the HELP command from the command line of a
Tivoli Storage Manager administrative client.

Road map for key device-related task information

Key tasks include configuring and managing disk devices, physically attaching
storage devices to your system, and so on. In this document, information about
tasks is organized into linked topics.

Use the following table to identify key tasks and the topics that describe how to
perform those tasks.

Task Topic
Configure and manage magnetic disk Chapter 4, “Magnetic disk devices,” on page
devices, which Tivoli Storage Manager uses 63
to store client data, the database, database
backups, recovery log, and export data.
Physically attach storage devices to your Chapter 5, “Attaching devices for the
system. Install and configure the required server,” on page 73
device drivers.
Configure devices to use with Tivoli Storage Chapter 6, “Configuring storage devices,” on
Manager, using detailed scenarios of page 85
representative device configurations.
Plan, configure, and manage an environment Chapter 9, “Using NDMP for operations
for NDMP operations with NAS file servers,” on page 203
Perform routine operations such as labeling Chapter 7, “Managing removable media
volumes, checking volumes into automated operations,” on page 127
libraries, and maintaining storage volumes
and devices.
Define and manage device classes. “Defining device classes” on page 177

© Copyright IBM Corp. 1993, 2013 35

Tivoli Storage Manager storage devices
With Tivoli Storage Manager, you can use a range of manual and automated
devices for server storage. Both direct and network-attached storage provide
options for storing data. Tivoli Storage Manager devices can be physical, such as
disk drives and tape drives, or logical, such as files on disk or storage on another
server.

Tivoli Storage Manager supports the following types of devices:

v Tape devices
v Removable file devices
v Disk devices
v Storage area network (SAN) devices
Devices in a SAN environment must be supported by the Tivoli Storage
Manager server.

For a summary of supported devices, see Table 3 on page 58. For details and
updates, see the Tivoli Storage Manager device support website:

[Link]
IBM_TSM_Supported_Devices_for_AIXHPSUNWIN.html

Tivoli Storage Manager storage objects

Devices and media are represented by objects that you define. Information about
these objects is stored in the Tivoli Storage Manager database.

You can query, update, and delete the following objects:

v Library
v Drive
v Device class
v Disk Devices
v Storage pool
v Storage pool volume
v Data mover
v Path
v Server

Libraries
A physical library is a collection of one or more drives that share similar
media-mounting requirements. That is, the drive can be mounted by an operator or
by an automated mounting mechanism.

A library object definition specifies the library type, for example, SCSI or 349X, and
other characteristics associated with the library type, for example, the category
numbers used by an IBM TotalStorage 3494 Tape Library for private, scratch
volumes, and scratch, write-once, read-many (WORM) volumes.

Tivoli Storage Manager supports a variety of library types.

36 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Shared libraries
Shared libraries are logical libraries that are represented physically by SCSI, 349X,
ACSLS, or VTL libraries. The physical library is controlled by the Tivoli Storage
Manager server that is configured as a library manager. Tivoli Storage Manager
servers that use the SHARED library type are library clients to the library manager
server. Shared libraries reference a library manager.

Automated cartridge system library software libraries

An automated cartridge system library software (ACSLS) library is a type of
external library that is controlled by Oracle StorageTek ACSLS media-management
software. The server can act as a client application to the ACSLS software to use
the drives.

The StorageTek software performs the following functions:

v Mounts volumes, both private and scratch
v Dismounts volumes
v Returns library volumes to scratch status
The ACSLS software selects an appropriate drive for media-access operations. You
do not define the drives, check in media, or label the volumes in an external
library.

Restriction: To use ACSLS functions, the StorageTek Library Attach software must
be installed. For more information, see “ACSLS-managed libraries” on page 109.

For additional information regarding ACSLS libraries, refer to your StorageTek

documentation. If you are using an ACSLS library and have a firewall enabled,
refer to the StorageTek Automated Cartridge System Library Software Administrator’s
Guide, Version 8.1 (April 2012 edition).

Manual libraries
In manual libraries, operators mount the volumes in response to mount-request
messages issued by the server.

The server sends these messages to the server console and to administrative clients
that were started by using the special MOUNTMODE or CONSOLEMODE parameter.

You can also use manual libraries as logical entities for sharing sequential-access
disk (FILE) volumes with other servers.

You cannot combine drives of different types or formats, such as Digital Linear
Tape (DLT) and 8MM, in a single manual library. Instead, you must create a
separate manual library for each device type.

For information about configuring a manual library, see:

Chapter 6, “Configuring storage devices,” on page 85
For information about monitoring mount messages for a manual library, see:
“Tivoli Storage Manager server requests” on page 143

Chapter 3. Storage device concepts 37

 SCSI libraries
A SCSI library is controlled through a SCSI interface, attached either directly to the
server's host using SCSI cabling or by a storage area network. A robot or other
mechanism automatically handles volume mounts and dismounts.

The drives in a SCSI library can be of different types. A SCSI library can contain
drives of mixed technologies, for example LTO Ultrium and DLT drives. Some
examples of this library type are:
v The Oracle StorageTek L700 library
v The IBM 3590 tape device, with its Automatic Cartridge Facility (ACF)

Remember: Although it has a SCSI interface, the IBM 3494 Tape Library
Dataserver is defined as a 349X library type.

For information about configuring a SCSI library, see:

Chapter 6, “Configuring storage devices,” on page 85

Virtual tape libraries

A virtual tape library (VTL) is a hardware component that can emulate a tape
library while using a disk as the underlying storage hardware.

Using a VTL, you can create variable numbers of drives and volumes because they
are only logical entities within the VTL. The ability to create more drives and
volumes increases the capability for parallelism, giving you more simultaneous
mounts and tape I/O.

VTLs use SCSI and Fibre Channel interfaces to interact with applications. Because
VTLs emulate tape drives, libraries, and volumes, an application such as Tivoli
Storage Manager cannot distinguish a VTL from real tape hardware unless the
library is identified as a VTL.

For information about configuring a VTL library, see “Managing virtual tape
libraries” on page 116.

349X libraries
A 349X library is a collection of drives in an IBM 3494. Volume mounts and
dismounts are handled automatically by the library. A 349X library has one or
more library management control points (LMCP) that the server uses to mount and
dismount volumes in a drive. Each LMCP provides an independent interface to the
robot mechanism in the library.

Categories in an IBM 3494 library

The library manager that is built into the IBM 3494 library tracks the category
number of each volume in the library. A single category number identifies all
volumes that are used for the same purpose or application. Category numbers are
useful when multiple systems share the resources of a single library.

Attention: If other systems or other Tivoli Storage Manager servers connect to the
same 3494 library, each must use a unique set of category numbers. Otherwise, two
or more systems might try to use the same volume, and cause corruption or loss of
data.

Typically, a software application that uses a 3494 library uses volumes in one or
more categories that are reserved for that application. To avoid loss of data, each

38 IBM Tivoli Storage Manager for Windows: Administrator's Guide

application that shares the library must have unique categories. When you define a
3494 library to the server, you can use the PRIVATECATEGORY and SCRATCHCATEGORY
parameters to specify the category numbers for private and scratch Tivoli Storage
Manager volumes in that library. If the volumes are IBM 3592 WORM (write once,
read many) volumes, you can use the WORMSCRATCHCATEGORY parameter to specify
category numbers for scratch WORM volumes in the library. For more information
about private, scratch, and scratch WORM volumes, see “Tivoli Storage Manager
volumes” on page 46.

When a volume is first inserted into the library, either manually or automatically at
the convenience I/O station, the volume is assigned to the insert category
(X'FF00'). A software application such as Tivoli Storage Manager can contact the
library manager to change the category number for a volume. For Tivoli Storage
Manager, you use the CHECKIN LIBVOLUME command.

349X library objects contain only one device type (3590 or 3592) of drives. Thus, if
you have 3590 and 3592 drives in your 349X library, you must define two library
objects: one for your 3590 drives and one for your 3592 drives. Each of these
library objects must have the same device parameter when their paths are defined.

For information about configuring a 349X library, see:

Chapter 6, “Configuring storage devices,” on page 85

External libraries
An external library is a collection of drives that are managed by an external
media-management system that is not part of Tivoli Storage Manager. The server
provides an interface that allows external media management systems to operate
with the server.

The external media-management system performs the following functions:

v Volume mounts (specific and scratch)
v Volume dismounts
v Freeing of library volumes (return to scratch)

The external media manager selects the appropriate drive for media-access
operations. You do not define the drives, check in media, or label the volumes in
an external library.

An external library allows flexibility in grouping drives into libraries and storage
pools. The library can have one drive, a collection of drives, or even a part of an
automated library.

An ACSLS or LibraryStation-controlled Oracle StorageTek library used in

conjunction with an external library manager (ELM), like Gresham's
EDT-DistribuTAPE, is a type of external library. IBM Tape Storage Library Manager
(TSLM) is another example.

For a definition of the interface that Tivoli Storage Manager provides to the
external media management system, see Appendix A, “External media
management interface description,” on page 1085.

Chapter 3. Storage device concepts 39

 Drives
A drive object represents a drive mechanism within a library that uses removable
media. For devices with multiple drives, including automated libraries, you must
define each drive separately and associate it with a library.

Drive definitions can include such information as the element address for drives in
SCSI or virtual tape libraries (VTLs), how often a tape drive is cleaned, and
whether the drive is online.

Tivoli Storage Manager drives include tape drives that can stand alone or that can
be part of an automated library. Supported removable media drives also include
removable file devices such as rewritable CDs.

Device class
Each device that is defined to Tivoli Storage Manager is associated with one device
class. The device class specifies the device type and media management
information, such as recording format, estimated capacity, and labeling prefixes.

A device type identifies a device as a member of a group of devices that share

similar media characteristics. For example, the 8MM device type applies to 8 mm
tape drives.

Device types include various removable media types along with FILE, CENTERA,
and SERVER.

A device class for a tape drive must also specify a library.

Disk devices
Using Tivoli Storage Manager, you can define random-access disk (DISK device
type) volumes using a single command. You can also use space triggers to
automatically create preassigned private volumes when predetermined
space-utilization thresholds are exceeded.

For important disk-related information, see “Requirements for disk systems” on

page 63.

Removable media
Tivoli Storage Manager provides a set of specified removable-media device types,
such as 8MM for 8 mm tape devices, or REMOVABLEFILE for Jaz or DVD-RAM
drives.

The GENERICTAPE device type is provided to support certain devices that are not
supported by the Tivoli Storage Manager server.

For more information about supported removable media device types, see
“Defining device classes” on page 177 and the Administrator's Reference.

40 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Files on disk as sequential volumes (FILE)
The FILE device type lets you create sequential volumes by creating files on disk
storage. To the server, these files have the characteristics of a tape volume. FILE
volumes can also be useful when transferring data for purposes such as electronic
vaulting or for taking advantage of relatively inexpensive disk storage devices.

FILE volumes are a convenient way to use sequential-access disk storage for the
following reasons:
v You do not need to explicitly define scratch volumes. The server can
automatically acquire and define scratch FILE volumes as needed.
v You can create and format FILE volumes using a single command. The
advantage of private FILE volumes is that they can reduce disk fragmentation
and maintenance overhead.
v Using a single device class definition that specifies two or more directories, you
can create large, FILE-type storage pools. Volumes are created in the directories
you specify in the device class definition. For optimal performance, volumes
should be associated with file systems.
v When predetermined space-utilization thresholds have been exceeded, space
trigger functionality can automatically allocate space for private volumes in
FILE-type storage pools.
v The Tivoli Storage Manager server allows concurrent read-access and
write-access to a volume in a storage pool associated with the FILE device type.
Concurrent access improves restore performance by allowing two or more clients
to access the same volume at the same time. Multiple client sessions (archive,
retrieve, backup, and restore) or server processes (for example, storage pool
backup) can read the volume concurrently. In addition, one client session or one
server process can write to the volume while it is being read.
The following server processes are allowed shared read access to FILE volumes:
– BACKUP DB
– BACKUP STGPOOL
– COPY ACTIVEDATA
– EXPORT/IMPORT NODE
– EXPORT/IMPORT SERVER
– GENERATE BACKUPSET
– RESTORE STGPOOL
– RESTORE VOLUME
The following server processes are not allowed shared read access to FILE
volumes:
– AUDIT VOLUME
– DELETE VOLUME
– MIGRATION
– MOVE DATA
– MOVE NODEDATA
– RECLAMATION

Unless sharing with storage agents is specified, the FILE device type does not
require you to define library or drive objects. The only required object is a device
class.

For important disk-related information, see “Requirements for disk systems” on

page 63.
Chapter 3. Storage device concepts 41
 Files on sequential volumes (CENTERA)
The CENTERA device type defines the EMC Centera storage device. It can be used
like any standard storage device from which files can be backed up and archived
as needed.

The Centera storage device can also be configured with the Tivoli Storage Manager
server to form a specialized storage system that protects you from inadvertent
deletion of mission-critical data such as emails, trade settlements, legal documents,
and so on.

The CENTERA device class creates logical sequential volumes for use with Centera
storage pools. These volumes share many of the same characteristics as FILE type
volumes. With the CENTERA device type, you are not required to define library or
drive objects. CENTERA volumes are created as needed and end in the suffix
"CNT."

Multiple client retrieve sessions, restore sessions, or server processes can read a
volume concurrently in a storage pool that is associated with the CENTERA device
type. In addition, one client session or one server process can write to the volume
while it is being read. Concurrent access improves restore and retrieve performance
because two or more clients can have access to the same volume at the same time.

The following server processes can share read access to Centera volumes:
v EXPORT NODE
v EXPORT SERVER
v GENERATE BACKUPSET

The following server processes cannot share read access to Centera volumes:
v AUDIT VOLUME
v DELETE VOLUME

For more information about the Centera device class, see “Defining device classes
for CENTERA devices” on page 197. For details about Centera-related commands,
refer to the Administrator's Reference.

Sequential volumes on another Tivoli Storage Manager server

(SERVER)
The SERVER device type lets you create volumes for one Tivoli Storage Manager
server that exist as archived files in the storage hierarchy of another server. These
virtual volumes have the characteristics of sequential-access volumes such as tape.
No library or drive definition is required.

You can use virtual volumes for the following:

v Device-sharing between servers. One server is attached to a large tape library
device. Other servers can use that library device indirectly through a SERVER
device class.
v Data-sharing between servers. By using a SERVER device class to export and
import data, physical media remains at the original location instead having to be
transported.
v Immediate offsite storage. Storage pools and databases can be backed up
without physically moving media to other locations.
v Offsite storage of the disaster recovery manager (DRM) recovery plan file.
v Electronic vaulting.

42 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 See “Using virtual volumes to store data on another server” on page 732.

Library, drive, and device-class objects

Library objects, drive objects, and device-class objects taken together represent
physical storage entities.

These three objects are shown in Figure 2.

Server Environment
Physical Device
Environment
Device Class

Library Represents

Drives
Device
Drive Drive

Figure 2. Removable media devices are represented by a library, drive, and device class

v For more information about the drive object, see:

“Managing drives” on page 152
“Defining drives” on page 174
v For more information about the library object, see:
“Managing libraries” on page 150
“Defining libraries” on page 173
v For more information about the device class object, see “Defining device classes”
on page 177.

Storage pools and storage-pool volumes

A storage pool is a collection of volumes that are associated with one device class
and one media type. For example, a storage pool that is associated with a device
class for 8-mm tape volumes contains only 8 mm tape volumes.

You can control the characteristics of storage pools, such as whether scratch
volumes are used.

Tivoli Storage Manager supplies default disk storage pools. .

Figure 3 on page 44 shows storage pool volumes grouped into a storage pool. Each
storage pool represents only one type of media. For example, a storage pool for
8-mm devices represents collections of only 8-mm tapes.

Chapter 3. Storage device concepts 43

 Volume Volume

Represents

Storage Media
Pool

Figure 3. Relationships of storage pool volumes, storage pools, and media

For DISK device classes, you must define volumes. For other device classes, such
as tape and FILE, you can allow the server to dynamically acquire scratch volumes
and define those volumes as needed. For details, see:
“Preparing volumes for random-access storage pools” on page 255
“Preparing volumes for sequential-access storage pools” on page 255

One or more device classes are associated with one library, which can contain
multiple drives. When you define a storage pool, you associate the pool with a
device class. Volumes are associated with pools. Figure 4 shows these relationships.

Storage Pool
Vol. Vol. Volumes

Storage Storage Storage

Pool Pool Pool

Device Class Device Class

Library

Drive Drive Drive Drive

Figure 4. Relationships between storage and device objects

For information about defining storage pool and volume objects, see Chapter 10,
“Managing storage pools and volumes,” on page 239.

For information about configuring volumes for random access see“Configuring

random access volumes on disk devices” on page 69.

44 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Data movers
Data movers are devices that accept requests from Tivoli Storage Manager to
transfer data on behalf of the server. Data movers transfer data between storage
devices without using significant server, client, or network resources.

For NDMP operations, data movers are NAS file servers. The definition for a NAS
data mover contains the network address, authorization, and data formats required
for NDMP operations. A data mover enables communication and ensures authority
for NDMP operations between the Tivoli Storage Manager server and the NAS file
server.

Tivoli Storage Manager supports two types of data movers:

v For NDMP operations, data movers are NAS file servers. The definition for a
NAS data mover contains the network address, authorization, and data formats
required for NDMP operations. A data mover enables communication and
ensures authority for NDMP operations between the Tivoli Storage Manager
server and the NAS file server.
v For server-free data movement, data movers are devices such as the IBM SAN
Data Gateway, that move data between disk devices and tape devices on the
SAN.

Paths
Paths allow access to drives, disks, and libraries. A path definition specifies a
source and a destination. The source accesses the destination, but data can flow in
either direction between the source and destination.

Here are a few examples of paths:

v Between a server and a drive or a library
v Between a storage agent and a drive
v Between a data mover and a drive, a disk, or a library

For more information about the path object, see:

“Defining paths” on page 176
“Managing paths” on page 169

Server objects
Server objects are defined to use a library that is on a SAN and that is managed by
another Tivoli Storage Manager server, to use LAN-free data movement, or to store
data in virtual volumes on a remote server.

Among other characteristics, you must specify the server TCP/IP address.

For more information, see:

v “Setting up the library client servers” on page 121
v “Using virtual volumes to store data on another server” on page 732
v Storage Agent User's Guide

Chapter 3. Storage device concepts 45

Tivoli Storage Manager volumes
A volume is the basic unit of storage for Tivoli Storage Manager storage pools.
Tivoli Storage Manager volumes are classified according to status: private, scratch,
and scratch write-once, read-many (WORM). Scratch WORM status applies to 349X
libraries only when the volumes are IBM 3592 WORM volumes.

The following definitions apply:

v A private volume is a labeled volume that is in use or owned by an application,
and may contain valid data. You must define each private volume. Alternatively,
for storage pools associated with sequential access disk (FILE) device classes,
you can use space triggers to create private, preassigned volumes when
predetermined space-utilization thresholds have been exceeded. Private FILE
volumes are allocated as a whole. The result is less risk of severe fragmentation
than with space dynamically acquired for scratch FILE volumes.
A request to mount a private volume must include the name of that volume.
Defined private volumes do not return to scratch when they become empty. For
information about defining private volumes, see “Defining storage pool
volumes” on page 256. For information about changing the status of a volume
(for example, from private to scratch) in an automated library, see the following:
– “Changing the status of automated library volumes” on page 135
v A scratch volume is a labeled volume that is empty or contains no valid data
and that can be used to satisfy any request to mount a scratch volume. When
data is written to a scratch volume, its status is changed to private, and it is
defined as part of the storage pool for which the mount request was made.
When valid data is moved from the volume and the volume is reclaimed, the
volume returns to scratch status and can be reused by any storage pool
associated with the library.
v A WORM scratch volume is similar to a conventional scratch volume. However,
WORM volumes cannot be reclaimed by Tivoli Storage Manager reclamation
processing. WORM volumes can be returned to scratch status only if they have
empty space in which data can be written. Empty space is space that does not
contain valid, expired or deleted data. (Deleted and expired data on WORM
volumes cannot be overwritten.) If a WORM volume does not have any empty
space in which data can be written (for example, if the volume is entirely full of
deleted or expired data), the volume remains private.

For each storage pool, you must decide whether to use scratch volumes. If you do
not use scratch volumes, you must define private volumes, or you can use
space-triggers if the volume is assigned to a storage pool with a FILE device type.
Tivoli Storage Manager keeps an inventory of volumes in each automated library it
manages and tracks whether the volumes are in scratch or private status. When a
volume mount is requested, Tivoli Storage Manager selects a scratch volume only
if scratch volumes are allowed in the storage pool. The server can choose any
scratch volume that has been checked into the library.

You do not need to allocate volumes to different storage pools associated with the
same automated library. Each storage pool associated with the library can
dynamically acquire volumes from the library's inventory of scratch volumes. Even
if only one storage pool is associated with a library, you do not need to explicitly
define all the volumes for the storage pool. The server automatically adds volumes
to and deletes volumes from the storage pool.

Tip: A disadvantage of using scratch volumes is that volume usage information,

which you can use to determine when the media has reached its end of life, is

46 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 deleted when a private volume is returned to the scratch volume pool.

Volume inventory for an automated library

A library's volume inventory includes only those volumes that have been checked
into that library.

This inventory is not necessarily identical to the list of volumes in the storage
pools associated with the library. For example:
v A volume can be checked into the library but not be in a storage pool (a scratch
volume, a database backup volume, or a backup set volume).
v A volume can be defined to a storage pool associated with the library (a private
volume), but not checked into the library.

For more information on how to check in volumes, see the following:

v “Checking media into automated library devices” on page 130

Device configurations
You can configure devices on a local area network, on a storage area network, for
LAN-free data movement, and as network-attached storage. Tivoli Storage
Manager provides methods for configuring storage devices.

For information about supported devices and Fibre Channel hardware and
configurations, see [Link]
Software/Tivoli/Tivoli_Storage_Manager

Devices on local area networks

In the conventional local area network (LAN) configuration, one or more tape
libraries are associated with a single Tivoli Storage Manager server.

In a LAN configuration, client data, electronic mail, terminal connection,

application program, and device control information must all be handled by the
same network. Device control information and client backup and restore data flow
across the LAN.

For information about the categories of libraries that are supported by Tivoli
Storage Manager, see “Libraries” on page 36.

Devices on storage area networks

A SAN is a dedicated storage network that can improve system performance. On a
SAN, you can consolidate storage and relieve the distance, scalability, and
bandwidth limitations of LANs and wide area networks (WANs).

By using Tivoli Storage Manager in a SAN, you can take advantage of the
following functions:
v Sharing storage devices among multiple Tivoli Storage Manager servers.
Devices that use the GENERICTAPE device type are not included.
v Moving Tivoli Storage Manager client data directly to storage devices (LAN-free
data movement) by configuring a storage agent on the client system.

In a SAN, you can share tape drives and libraries that are supported by the Tivoli
Storage Manager server, including most SCSI devices.

Chapter 3. Storage device concepts 47

 Figure 5 shows a SAN configuration in which two Tivoli Storage Manager servers
share a library.

Library Manager
Library Client Server Server
Library
Control
LAN

Library Control
Data Flow
SAN Data Flow

Tape Library

Figure 5. Library sharing in a storage area network (SAN) configuration. The servers
communicate over the LAN. The library manager controls the library over the SAN. The
library client stores data to the library devices over the SAN.

When Tivoli Storage Manager servers share a library, the server that is defined as
the library manager controls device operations. These operations include mount,
dismount, volume ownership, and managing library inventory. Tivoli Storage
Manager servers that are defined as library clients use server-to-server
communications to contact the library manager and request device service. Data
moves over the SAN between each server and the storage device.

Tivoli Storage Manager servers use the following features when sharing an
automated library:
Partitioning of the Volume Inventory
The inventory of media volumes in the shared library is partitioned among
servers. Either one server owns a particular volume, or the volume is in
the global scratch pool. No server owns the scratch pool at any time.
Serialized Drive Access
Only one server accesses each tape drive at a time. Drive access is
serialized and controlled so that servers do not dismount other servers'
volumes or write to drives where other servers mount their volumes.
Serialized Mount Access
The library autochanger completes a single mount or dismount operation
at a time. The library manager completes all mount operations to provide
this serialization.
Related tasks:

48 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 “Configuring Tivoli Storage Manager servers to share SAN-connected devices” on
page 119

LAN-free data movement

Tivoli Storage Manager allows a client, through a storage agent, to directly back up
and restore data to a tape library on a SAN.

Figure 6 shows a SAN configuration in which a client directly accesses a tape or

FILE library to read or write data.

Library control Tivoli Storage

client metadata Manager
server
LAN
Client data
Library control
Tivoli Storage Manager
backup-archive client
with storage agent
installed SAN

Tape library
File library

Figure 6. LAN-Free data movement. Client and server communicate over the LAN. The
server controls the device on the SAN. Client data moves over the SAN to the device.

LAN-free data movement requires the installation of a storage agent on the client
system. The server maintains the database and recovery log, and acts as the library
manager to control device operations. The storage agent on the client handles the
data transfer to the device on the SAN. This implementation frees up bandwidth
on the LAN that would otherwise be used for client data movement.

The following outlines a typical backup scenario for a client that uses LAN-free
data movement:
1. The client begins a backup operation. The client and the server exchange policy
information over the LAN to determine the destination of the backed up data.
For a client that uses LAN-free data movement, the destination is a storage
pool that uses a device on the SAN.
2. Because the destination is on the SAN, the client contacts the storage agent,
which handles the data transfer. The storage agent sends a request for a volume
mount to the server.
3. The server contacts the storage device and, in the case of a tape library, mounts
the appropriate media.
4. The server notifies the client of the location of the mounted media.
5. The client, through the storage agent, writes the backup data directly to the
device over the SAN.
6. The storage agent sends file attribute information to the server, and the server
stores the information in its database.

Chapter 3. Storage device concepts 49

 If a failure occurs on the SAN path, failover occurs. The client uses its LAN
connection to the Tivoli Storage Manager server and moves the client data over the
LAN.

Remember:
v Centera storage devices cannot be targets for LAN-free operations.
v For the latest information about clients that support the feature, see the IBM
Tivoli Storage Manager support page at [Link]
portal/Overview/Software/Tivoli/Tivoli_Storage_Manager.

Network-attached storage
Network-attached storage (NAS) file servers are dedicated storage machines whose
operating systems are optimized for file-serving functions. NAS file servers
typically do not run software that is acquired from another vendor. Instead, they
interact with programs like Tivoli Storage Manager through industry-standard
network protocols, such as network data management protocol (NDMP).

Tivoli Storage Manager provides two basic types of configurations that use NDMP
for backing up and managing NAS file servers. In one type of configuration, Tivoli
Storage Manager uses NDMP to back up a NAS file server to a library device
directly attached to the NAS file server. (See Figure 7.) The NAS file server, which
can be distant from the Tivoli Storage Manager server, transfers backup data
directly to a drive in a SCSI-attached tape library. Data is stored in special,
NDMP-formatted storage pools, which can be backed up to storage media that can
be moved offsite for protection in case of an onsite disaster.

Server

Offsite storage
Tape Library

NAS File
Server
Legend:

SCSI or Fibre
Channel Connection
TCP/IP NAS File Server
Connection File System
Data Flow Disks

Figure 7. Library device directly attached to a NAS file server

In the other type of NDMP-based configuration, Tivoli Storage Manager uses

NDMP to back up a NAS file server to a Tivoli Storage Manager storage-pool
hierarchy. (See Figure 8 on page 51.) With this type of configuration, you can store
NAS data directly to disk (either random access or sequential access) and then
migrate the data to tape. Data can also be backed up to storage media that can
then be moved offsite. The advantage of this type of configuration is that it gives
you all the backend-data management features associated with a conventional
50 IBM Tivoli Storage Manager for Windows: Administrator's Guide
Tivoli Storage Manager storage-pool hierarchy, including migration and
reclamation.

Disk Storage
Server Offsite storage
Tape Library

NAS File
Server
Legend:

SCSI or Fibre
Channel Connection
TCP/IP NAS File Server
Connection File System
Data Flow Disks

Figure 8. NAS file server toTivoli Storage Manager storage-pool hierarchy

In both types of configurations, Tivoli Storage Manager tracks file system image
backups and has the capability to perform NDMP file-level restores. For more
information about NDMP file-level restores, see “NDMP file-level restoration” on
page 52.

Note:
v A Centera storage device cannot be a target for NDMP operations.
v Support for filer-to-server data transfer is only available for NAS devices that
support NDMP version 4.
v For a comparison of NAS backup methods, including using a backup-archive
client to back up a NAS file server, see “Determining the location of NAS
backup” on page 213.

NDMP backup operations

In backup images produced by network data management protocol (NDMP)
operations for a NAS file server, Tivoli Storage Manager creates NAS
file-system-level or directory-level image backups.

The image backups are different from traditional Tivoli Storage Manager backups
because the NAS file server transfers the data to the drives in the library or
directly to the Tivoli Storage Manager server. NAS file system image backups can
be either full or differential image backups. The first backup of a file system on a
NAS file server is always a full image backup. By default, subsequent backups are
differential image backups containing only data that has changed in the file system
since the last full image backup. If a full image backup does not already exist, a
full image backup is performed.

Chapter 3. Storage device concepts 51

 If you restore a differential image, Tivoli Storage Manager automatically restores
the full backup image first, followed by the differential image.

NDMP file-level restoration

Tivoli Storage Manager provides a way to restore data from backup images
produced by NDMP operations. To assist users in restoring selected files, you can
create a table of contents (TOC) of file-level information for each backup image.

Using the Web backup-archive client, users can then browse the TOC and select
the files that they want to restore. If you do not create a TOC, users must be able
to specify the name of the backup image that contains the file to be restored and
the fully qualified name of the file.

You can create a TOC using one of the following commands:

v BACKUP NODE server command. For details, see the Administrator's Reference.
v BACKUP NAS client command, with [Link] specified in the client
options file or specified in the client options set. For details, see the
Backup-Archive Clients Installation and User's Guide.

Directory-level backup and restore

If you have a large NAS file system, initiating a backup on a directory level
reduces backup and restore times, and provides more flexibility in configuring
your NAS backups.

By defining virtual file spaces, a file system backup can be partitioned among
several NDMP backup operations and multiple tape drives. You can also use
different backup schedules to back up sub-trees of a file system.

The virtual file space name cannot be identical to any file system on the NAS
node. If a file system is created on the NAS device with the same name as a virtual
file system, a name conflict will occur on the Tivoli Storage Manager server when
the new file space is backed up. See the Administrator's Reference for more
information about virtual file space mapping commands.

Remember: Virtual file space mappings are only supported for NAS nodes.

Mixed device types in libraries

Tivoli Storage Manager supports mixing different device types within a single
automated library, as long as the library itself can distinguish among the different
media for the different device types.

Libraries with this capability are those models supplied from the manufacturer
already containing mixed drives, or capable of supporting the addition of mixed
drives. Check with the manufacturer, and also check the Tivoli Storage Manager
Web site for specific libraries that have been tested on Tivoli Storage Manager with
mixed device types.

For example, you can have Quantum SuperDLT drives, LTO Ultrium drives, and
StorageTek 9940 drives in a single library defined to the Tivoli Storage Manager
server. For examples of how to set this up, see:
“Defining Tivoli Storage Manager storage objects with commands” on page 92
“Configuring a 3494 library with multiple drive device types” on page 99

52 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Different media generations in a library
While the Tivoli Storage Manager server now allows mixed device types in an
automated library, the mixing of different generations of the same type of drive is
still not supported. New drives cannot write the older media formats, and old
drives cannot read new formats.

If the new drive technology cannot write to media formatted by older generation
drives, the older media must be marked read-only to avoid problems for server
operations. Also, the older drives must be removed from the library. Some
examples of combinations that the Tivoli Storage Manager server does not support
in a single library are:
v SDLT 220 drives with SDLT 320 drives
v DLT 7000 drives with DLT 8000 drives
v StorageTek 9940A drives with 9940B drives
v UDO1 drives with UDO2 drives

There are exceptions to the rule against mixing generations of LTO Ultrium drives
and media. The Tivoli Storage Manager server does support mixtures of the
following types:
v LTO Ultrium Generation 1 (LTO1) and LTO Ultrium Generation 2 (LTO2)
v LTO Ultrium Generation 2 (LTO2) with LTO Ultrium Generation 3 (LTO3)
v LTO Ultrium Generation 3 (LTO3) with LTO Ultrium Generation 4 (LTO4)
v LTO Ultrium Generation 4 (LTO4) with LTO Ultrium Generation 5 (LTO5)
v LTO Ultrium Generation 5 (LTO5) with LTO Ultrium Generation 6 (LTO6)

The server supports these mixtures because the different drives can read and write
to the different media. If you plan to upgrade all drives to Generation 2 (or
Generation 3, Generation 4, or Generation 5), first delete all existing Ultrium drive
definitions and the paths associated with them. Then you can define the new
Generation 2 (or Generation 3, Generation 4, or Generation 5) drives and paths.

Note:
1. LTO Ultrium Generation 3 drives can only read Generation 1 media. If you are
mixing Ultrium Generation 1 with Ultrium Generation 3 drives and media in a
single library, you must mark the Generation 1 media as read-only, and all
Generation 1 scratch volumes must be checked out.
2. LTO Ultrium Generation 4 drives can only read Generation 2 media. If you are
mixing Ultrium Generation 2 with Ultrium Generation 4 drives and media in a
single library, you must mark the Generation 2 media as read-only, and all
Generation 2 scratch volumes must be checked out.
3. LTO Ultrium Generation 5 drives can only read Generation 3 media. If you are
mixing Ultrium Generation 3 with Ultrium Generation 5 drives and media in a
single library, you must mark the Generation 3 media as read-only, and all
Generation 3 scratch volumes must be checked out.
4. LTO Ultrium Generation 6 drives can only read Generation 4 media. If you are
mixing Ultrium Generation 4 with Ultrium Generation 6 drives and media in a
single library, you must mark the Generation 4 media as read-only, and all
Generation 4 scratch volumes must be checked out.

To learn more about additional considerations when mixing LTO Ultrium

generations, see “Defining LTO device classes” on page 191.

Chapter 3. Storage device concepts 53

 When using Tivoli Storage Manager, you cannot mix drives that are 3592 drive
generations. Use one of three special configurations. For details, see “Defining 3592
device classes” on page 183.

If you plan to encrypt volumes in a library, do not mix media generations in the
library.

Mixed media and storage pools

You cannot mix media formats in a storage pool. Each unique media format must
be mapped to a separate storage pool through its own device class.

This includes LTO formats. Multiple storage pools and their device classes of
different types can point to the same library that can support them as explained in
“Different media generations in a library” on page 53.

You can migrate to a new generation of a media type within the same storage pool
by following these steps:
1. ALL older drives are replaced with the newer generation drives within the
library (they cannot be mixed).
2. The existing volumes with the older formats are marked R/O if the new drive
cannot append those tapes in the old format. If the new drive can write to the
existing media in their old format, this is not necessary, but Step 1 is still
required. If it is necessary to keep different drive generations that are read but
not write compatible within the same library, separate storage pools for each
must be used.

Library sharing
Library sharing or tape resource sharing allows multiple Tivoli Storage Manager
servers to use the same tape library and drives on a storage area network (SAN)
and to improve backup and recovery performance and tape hardware asset
utilization.

When Tivoli Storage Manager servers share a library, one server is set up as the
library manager and controls library operations such as mount and dismount. The
library manager also controls volume ownership and the library inventory. Other
servers are set up as library clients and use server-to-server communications to
contact the library manager and request resources.

Library clients must be at the same or a lower version than the library manager
server. A library manager cannot support library clients that are at a higher
version. For example, a version 6.2 library manager can support a version 6.1
library client but cannot support a version 6.3 library client.

Removable media mounts and dismounts

When data is to be stored in or retrieved from a storage pool, the server selects the
storage-pool volume and determines the name of the library that contains the
drives to be used for the operation. When it has finished accessing the volume and
the mount retention period has elapsed, the server dismounts the volume.

When data is to be stored in or retrieved from a storage pool, the server does the
following:
1. The server selects a volume from the storage pool. The selection is based on the
type of operation:

54 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 Retrieval
The name of the volume that contains the data to be retrieved is stored
in the database.
Store If a defined volume in the storage pool can be used, the server selects
that volume.
If no defined volumes in the storage pool can be used, and if the
storage pool allows it, the server selects a scratch volume.
2. The server checks the device class associated with the storage pool to
determine the name of the library that contains the drives to be used for the
operation.
v The server searches the library for an available drive or until all drives have
been checked. A drive status can be:
– Offline.
– Busy and not available for the mount.
– In an error state and not available for the mount.
– Online and available for the mount.
3. The server mounts the volume:
v For a manual library, the server displays a mount message for a private or a
scratch volume to be mounted in the selected drive.
v For an automated library, the server directs the library to move the volume
from a storage slot into the selected drive. No manual intervention is
required.
If a scratch mount is requested, the server checks the library's volume
inventory for a scratch volume. If one is found, its status is changed to
private, it is mounted in the drive, and it is automatically defined as part of
the original storage pool. However, if the library's volume inventory does not
contain any scratch volumes, the mount request fails.
4. The server dismounts the volume when it has finished accessing the volume
and the mount retention period has elapsed.
v For a manual library, the server ejects the volume from the drive so that an
operator can place it in its storage location.
v For an automated library, the server directs the library to move the volume
from the drive back to its original storage slot in the library.

How Tivoli Storage Manager uses and reuses removable media

Using Tivoli Storage Manager, you can control how removable media are used and
reused. After Tivoli Storage Manager selects an available medium, that medium is
used and eventually reclaimed according to its associated policy.

Tivoli Storage Manager manages the data on the media, but you manage the media
itself, or you can use a removable media manager. Managing media involves
creating a policy to expire data after a certain time or under certain conditions,
moving valid data onto new media, and reusing the empty media.

In addition to information about storage pool volumes, volume history contains

information about tapes that are used for database backups and exports (for
disaster recovery purposes). The process for reusing these tapes is slightly different
from the process for reusing tapes that contain client data backups.

Chapter 3. Storage device concepts 55

 Figure 9 shows a typical lifecycle for removable media. The numbers (such as 1)
refer to numbers in the figure.

New tape preparation

1 2

Label tape Add tape

to inventory

Tape inventory
Ongoing tape processing
3

Select tape

Data
expires or
moves

Reclaim

Figure 9. Simplified view of the lifecycle of a tape

1. You label 1 and check in 2 the media. Checking media into a manual library
means storing them (for example, on shelves). Checking media into an
automated library involves adding them to the library volume inventory.
2. If you plan to define volumes to a storage pool associated with a device, check
in the volume with its status specified as private. Use of scratch volumes is
more convenient in most cases.
3. A client sends data to the server for backup, archive, or space management.
The server stores the client data on the volume. Which volume the server
selects (3) depends on:
v The policy domain to which the client is assigned.
v The management class for the data. It can be either the default management
class for the policy set, or the class that is specified by the client in the
client's include/exclude list or file.
v The storage pool that is specified as the destination in either the management
class (for space-managed data) or copy group (for backup or archive data).
The storage pool is associated with a device class, which determines which
device and which type of media is used.
v Whether the maximum number of scratch volumes that a server can request
from the storage pool was reached, when the scratch volumes are selected.

56 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 v Whether collocation is enabled for that storage pool. When collocation is
enabled, the server tries to place data on separate volumes. The data can
come from different client nodes, groups of client nodes, file spaces, or
groups of file spaces. For more information, see “Keeping client files together
using collocation” on page 357.
Figure 10 shows more detail about the policies and storage pool specifications
that govern the volume selection that is described in this step.

Figure 10. How Tivoli Storage Manager affects media use

4. The data on a volume changes over time as a result of:

v Expiration of files 4 (affected by management class and copy group
attributes, and the frequency of expiration processing). See “Basic policy
planning” on page 476.
v Movement and deletion of file spaces by an administrator.
v Automatic reclamation of media 5
The amount of data on the volume and the reclamation threshold that is set
for the storage pool affects when the volume is reclaimed. When the volume
is reclaimed, any valid, unexpired data is moved to other volumes or
possibly to another storage pool (for storage pools with single-drive
libraries).
v Collocation, by which the server tries to keep data on a minimal number of
removable media in a storage pool. The data can belong to a single client
node, a group of client nodes, a file space, or a group of file spaces.
If the volume becomes empty because all valid data either expires or is moved
to another volume, the volume is available for reuse. It is not available for

Chapter 3. Storage device concepts 57

 reuse if a time delay was specified for the storage pool. The empty volume
becomes a scratch volume if it was initially a scratch volume. The volume starts
again at step 3 on page 56.
5. You determine when the media is at its end-of-life.
For volumes that you defined (private volumes), check the statistics on the
volumes by querying the database. The statistics include these two entries:
v The number of write passes on a volume (compared with the number of
write passes that are recommended by the manufacturer)
v The number of errors on the volume
You must move any valid data off a volume that reached end-of-life. Then, if
the volume is in an automated library, check out the volume from the library. If
the volume is not a scratch volume, delete the volume from the database.

Definitions for storage devices

Before the Tivoli Storage Manager server can use a device, the device must be
configured to the operating system and to the server.

About this task

You can use the PERFORM LIBACTION command to simplify the process when you
add devices to SCSI and VTL library types.

Table 3 summarizes the definitions that are required for different device types.
Table 3. Required definitions for storage devices
Required Definitions
Library Drive Path Device
Device Device Types Class
Magnetic disk DISK — — — Yes See note
FILE See note — — — Yes
CENTERA — — — Yes
Tape 3590 Yes Yes Yes Yes
3592
4MM
8MM
DLT
LTO
NAS
VOLSAFE

GENERICTAPE

ECARTRIDGE See note

Removable REMOVABLEFILE Yes Yes Yes Yes
media (file
system)
Virtual SERVER — — — Yes
volumes

Notes:

58 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 v The DISK device class exists at installation and cannot be changed.
v FILE libraries, drives, and paths are required for sharing with storage agents.
v The ECARTRIDGE device type is for StorageTek cartridge tape drives such as
SD-3, 9480, 9890, and 9940 drives.

Example: Mapping devices to device classes

You have internal disk drives, an automated tape library with 8 mm drives, and a
manual DLT tape drive. You create a device class for each type of storage.

About this task

To map storage devices to device classes, use the information shown in Table 4.
Table 4. Mapping storage devices to device classes
Device Class Description
DISK Storage volumes that reside on the internal disk drive

Tivoli Storage Manager provides one DISK device class that

is already defined. You do not need and cannot define
another device class for disk storage.
8MM_CLASS Storage volumes that are 8 mm tapes, used with the drives
in the automated library
DLT_CLASS Storage volumes that are DLT tapes, used on the DLT drive

You must define any device classes that you need for your removable media
devices such as tape drives. See “Defining device classes” on page 177 for
information on defining device classes to support your physical storage
environment.

Example: Mapping storage pools to device classes and

devices
After you categorize your storage devices, you can identify availability, space, and
performance requirements for client data that is stored in server storage. These
requirements help you determine where to store data for different groups of clients
and different types of data. You can then create storage pools that are storage
destinations for backed-up, archived, or space-managed files to match
requirements.

About this task

For example, you determine that users in the business department have three
requirements:
v Immediate access to certain backed-up files, such as accounts receivable and
payroll accounts.
These files should be stored on disk. However, you need to ensure that data is
moved from the disk to prevent it from becoming full. You can set up a storage
hierarchy so that files can migrate automatically from disk to the automated tape
library.
v Periodic access to some archived files, such as monthly sales and inventory
reports.
These files can be stored on 8-mm tapes, using the automated library.

Chapter 3. Storage device concepts 59

 v Occasional access to backed-up or archived files that are rarely modified, such as
yearly revenue reports.
These files can be stored using the DLT drive.

To match user requirements to storage devices, you define storage pools, device
classes, and, for device types that require them, libraries and drives. For example,
to set up the storage hierarchy so that data migrates from the BACKUPPOOL to 8
mm tapes, you specify BACKTAPE1 as the next storage pool for BACKUPPOOL.
See Table 5.
Table 5. Mapping storage pools to device classes, libraries, and drives
Library
Storage Pool Device Class (Hardware) Drives Volume Type Storage Destination
BACKUPPOOL DISK — — Storage volumes For a backup copy
on the internal group for files
disk drive requiring immediate
access
BACKTAPE1 8MM_CLASS AUTO_8MM DRIVE01, 8-mm tapes For overflow from the
(Exabyte DRIVE02 BACKUPPOOL and for
EXB-210) archived data that is
periodically accessed
BACKTAPE2 DLT_CLASS MANUAL_LIB DRIVE03 DLT tapes For backup copy
(Manually groups for files that are
mounted) occasionally accessed
Note: Tivoli Storage Manager has the following default disk storage pools:
v BACKUPPOOL
v ARCHIVEPOOL
v SPACEMGPOOL
v DISKPOOL
For more information, see
“Configuring random access volumes on disk devices” on page 69

Planning for server storage

To determine the device classes and storage pools that you need for your server
storage, you must evaluate the devices in your storage environment.

About this task

Procedure
1. Determine which drives and libraries are supported by the server. For the
most up-to-date list of supported devices and operating-system levels, see the
Supported Devices website:
[Link]
IBM_TSM_Supported_Devices_for_AIXHPSUNWIN.html
2. Determine which storage devices can be selected for use by the server. For
example, determine how many tape drives you have that you can allow the
server to use. For more information about selecting a device configuration, see
“Device configurations” on page 47
The servers can share devices in libraries that are attached through a SAN. If
the devices are not on a SAN, the server expects to have exclusive use of the
drives that are defined to it. If another application (including another Tivoli

60 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 Storage Manager server) tries to use a drive while the server to which the
drive is defined is running, some server functions might fail. For more
information about specific drives and libraries, see [Link]
support/entry/portal/Overview/Software/Tivoli/Tivoli_Storage_Manager.
3. Determine the device driver that supports the devices. For more information
about device driver support, see:
“Selecting a device driver” on page 78
4. Determine how to attach the devices to the server. For more information about
attaching devices, see:
“Attaching an automated library device” on page 74
5. Determine whether to back up client data directly to tape or to a storage
hierarchy.
6. Determine which client data is backed up to which device, if you have
multiple device types.
7. Determine the device type and device class for each of the available devices.
Group similar devices together and identify their device classes. For example,
create separate categories for 4 mm and 8 mm devices.

Tip: For sequential access devices, you can categorize the type of removable
media based on their capacity.
For example, standard length cartridge tapes and longer length cartridge tapes
require different device classes.
8. Determine how the mounting of volumes is accomplished for the devices:
v Devices that require operators to load volumes must be part of a defined
MANUAL library.
v Devices that are automatically loaded must be part of a defined SCSI, 349X,
or VTL library. Each automated library device is a separate library.
v Devices that are controlled by Oracle StorageTek Automated Cartridge
System Library Software (ACSLS) must be part of a defined ACSLS library.
v Devices that are managed by an external media management system must
be part of a defined EXTERNAL library.
9. If you are considering storing data for one Tivoli Storage Manager server by
using the storage of another Tivoli Storage Manager server, consider network
bandwidth and network traffic. If your network resources constrain your
environment, you might have problems with using the SERVER device type
efficiently.
Also, consider the storage resources available on the target server. Ensure that
the target server has enough storage space and drives to handle the load from
the source server.
10. Determine the storage pools to set up, based on the devices you have and on
user requirements. Gather users' requirements for data availability. Determine
which data needs quick access and, which does not.
11. Be prepared to label removable media. You might want to create a new
labeling convention for media so that you can distinguish them from media
that are used for other purposes.

Chapter 3. Storage device concepts 61

Server options that affect storage operations
Tivoli Storage Manager provides a number of options that you can specify in the
server options file, [Link], to configure certain server storage operations.

The Table 6 table provides brief descriptions of these options.

Table 6. Server storage options
Option Description
3494SHARED Enables sharing of an IBM TotalStorage 3494 Tape Library
between a Tivoli Storage Manager server and server
applications other than a Tivoli Storage Manager server. This
configuration is not recommended because this
configuration can cause drive contention.
ACSACCESSID Specifies the ID for the Automatic Cartridge System (ACS)
access control.
ACSLOCKDRIVE Allows the drives within ACSLS libraries to be locked.
ACSQUICKINIT Allows a quick or full initialization of the ACSLS library.
ACSTIMEOUTX Specifies the multiple for the built-in timeout value for
ACSLS API.
ASSISTVCRRECOVERY
Specifies whether the server assists an IBM 3590 drive in
recovering from a lost or corrupted Vital Cartridge Records
(VCR) condition.
CHECKTAPEPOS Specifies whether the server validates data position on tape.
Tip: This option might conflict with storage area network
(SAN) tape acceleration.
DRIVEACQUIRERETRY Specifies how many times the server retries the acquisition
of a drive in a library when there are no drives available
after it acquires a mount point.
NOPREEMPT Specifies whether the server allows certain operations to
preempt other operations for access to volumes and devices.
See “Preempting operations” on page 611 for details.
RESOURCETIMEOUT Specifies how long the server waits for a resource before it
cancels the pending acquisition of a resource.
Tip: For proper management of shared library resources,
consider setting the RESOURCETIMEOUT option at the same
time limit for all servers in a shared configuration. For error
recovery, Tivoli Storage Manager always defers to the
longest time limit.
SEARCHMPQUEUE Specifies the order in which the server satisfies requests in
the mount queue.

62 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Chapter 4. Magnetic disk devices
Using magnetic disk devices, Tivoli Storage Manager can store essential data for
server and client environments.

Tivoli Storage Manager stores data on magnetic disks in random access volumes,
as data is normally stored on disk, and in files on the disk that are treated as
sequential access volumes.

You can store the following types of data on magnetic disk devices:
v The database and recovery log
v Backups of the database
v Export and import data
v Client data that is backed up, archived, or migrated from client nodes. The client
data is stored in storage pools.

See the following sections:

Tasks:
“Configuring random access volumes on disk devices” on page 69
“Configuring FILE sequential volumes on disk devices” on page 70
“Varying disk volumes online or offline” on page 70
“Cache copies for files stored on disk” on page 71
“Freeing space on disk” on page 71
“Scratch FILE volumes” on page 72
“Volume history file and volume reuse” on page 72

Note: Some of the tasks described in this chapter require an understanding of

storage objects. For an introduction to these storage objects, see “Tivoli Storage
Manager storage objects” on page 36.

Requirements for disk systems

Tivoli Storage Manager requires certain behaviors of disk storage systems for the
database, the active and archive logs, and storage pool volumes of the DISK device
class and of FILE device types.

Review the following Tivoli Storage Manager requirements for disk devices and
compare them with information from your disk system vendor. A list of supported
disk storage devices is not available. Contact the vendor for your disk system if
you have questions or concerns about whether Tivoli Storage Manager
requirements are supported. The vendor should be able to provide the
configuration settings to meet these requirements.

I/O operation results must be reported synchronously and accurately. For the
database and the active and archive logs, unreported or asynchronously reported
write errors that result in data not being permanently committed to the storage

© Copyright IBM Corp. 1993, 2013 63

 system can cause failures that range from internal processing errors to the inability
to restart the server. Depending upon the error, the result could be the loss of some
or all stored data.

Data in Tivoli Storage Manager storage pools, database volumes, and log volumes
must be interdependent. Tivoli Storage Manager requires that the data written to
these entities can be retrieved exactly as it was written. Also data in these entities
must be consistent with one another. There cannot be timing windows in which
data that is being retrieved varies depending on the way that an I/O system
manages the writing of data. Generally, this means that replicated Tivoli Storage
Manager environments must use features such as maintenance of write-order
between the source and replication targets. It also requires that the database, log,
and disk storage pool volumes be part of a consistency group in which any I/O to
the members of the target consistency group are written in the same order as the
source and maintain the same volatility characteristics. Requirements for I/O to
disk storage systems at the remote site must also be met.

Database write operations must be nonvolatile for active and archive logs and
DISK device class storage pool volumes. Data must be permanently committed to
storage that is known toTivoli Storage Manager Tivoli Storage Manager has many
of the attributes of a database system, and data relationships that are maintained
require that data written as a group be permanently resident as a group or not
resident as a group. Intermediate states produce data integrity issues. Data must be
permanently resident after each operating-system write API invocation.

For FILE device type storage pool volumes, data must be permanently resident
following an operating system flush API invocation. This API is used at key
processing points in the Tivoli Storage Manager application. The API is used when
data is to be permanently committed to storage and synchronized with database
and log records that have already been permanently committed to disk storage.

For systems that use caches of various types, the data must be permanently
committed by the write APIs for the database, the active and archive logs, and
DISK device class storage pool volumes and by the flush API (for FILE device class
storage pool volumes). Tivoli Storage Manager uses write-through flags internally
when using storage for the database, the active and archive logs, and DISK device
class storage pool volumes. Data for the I/O operation can be lost if nonvolatile
cache is used to safeguard I/O writes to a device and the nonvolatile cache is
battery protected. If there is a power loss and power is not restored before the
battery is exhausted, then data can be lost. This would be the same as having
uncommitted storage resulting in data integrity issues.

To write properly to the Tivoli Storage Manager database, to active and archive
logs, and to DISK device class storage pool volumes, the operating system API
write invocation must synchronously and accurately report the operation results.
Similarly, the operating system API flush invocation for FILE device type storage
pool volumes must also synchronously and accurately report the operation results.
A successful result from the API for either write or flush must guarantee that the
data is permanently committed to the storage system.

These requirements extend to replicated environments such that the remote site
must maintain consistency with the source site in terms of the order of writes; I/O
must be committed to storage at the remote site in the same order that it was
written at the source site. The ordering applies to the set of files that Tivoli Storage
Manager is writing, whether the files belong to the database, recovery log, or

64 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 storage pool volumes. Tivoli Storage Manager can recover from incomplete I/O
scenarios if the ordering of writes is consistent between the source and target site.

To avoid having the Tivoli Storage Manager server at the local and remote site
losing synchronization, the server at the remote site should not be started except in
a fail-over situation. If there is a possibility that data at the source and target
locations can lose synchronization, there must be a mechanism to recognize this
situation. If synchronization is lost, the Tivoli Storage Manager server at the remote
location must be restored by conventional means by using Tivoli Storage Manager
database and storage pool restores.

Tivoli Storage Manager supports the use of remote file systems or drives for
reading and writing storage pool data, database backups, and other data
operations. Remote file systems in particular might report successful writes, even
after being configured for synchronous operations. This mode of operation causes
data integrity issues if the file system can fail after reporting a successful write.
Check with the vendor of your file system to ensure that flushes are performed to
nonvolatile storage in a synchronous manner.

Comparison of random access and sequential access disk devices

Before configuring your disk device, you should consider the differences between
the two methods of storing data on disks and the advantages and disadvantages of
each. The particular advantages provided by either device type depends on the
operating system on which your Tivoli Storage Manager server is running.

Table 7 provides some general information about the characteristics of DISK

devices (random access) and FILE devices (sequential access) and the benefits of
each.
Table 7. Comparing random access and sequential access disk devices
Function Random Access (DISK) Sequential Access (FILE) Notes®
Storage space allocation Disk blocks Volumes Space allocation and
and tracking tracking by blocks uses
more database storage
space, and requires more
processing power than
space allocation and
tracking by volume.
Concurrent volume access A volume can be accessed A volume can be accessed Concurrent volume access
concurrently by different concurrently by different means that two or more
operations operations different operations can
access the same volume at
the same time.

Chapter 4. Magnetic disk devices 65

Table 7. Comparing random access and sequential access disk devices (continued)
Function Random Access (DISK) Sequential Access (FILE) Notes®
Client restore operations One session per restore Multiple concurrent Multi-session restore
sessions access different enables backup-archive
volumes simultaneously on clients to perform multiple
both the server and the restore sessions for
storage agent. Active no-query restore operations,
versions of client backup increasing the speed of
data is collocated in restores. Active-data pools
active-data pools. defined using
sequential-access disk
(FILE) enable fast client
restore because the server
does not physically mount
tapes and does not position
past inactive files.

For more information, see

“Backing up primary
storage pools” on page 883,
and the information about
client restore operations in
the Optimizing Performance
guide.
Available for use in Not available Available for LAN-free Using LAN-free backup,
LAN-free backup backup using IBM General data moves over a
Parallel File System. dedicated storage area
network (SAN) to the
sequential-access storage
device, freeing up
bandwidth on the LAN.

For more information, see

“LAN-free data movement”
on page 49.
Volume configuration Operators need to define The Tivoli Storage Manager For more information about
volumes and specify their server acquires and defines volumes on random-access
sizes, or define space scratch volumes as needed media, see “Configuring
triggers to automatically if storage administrators set random access volumes on
allocate space when a the MAXSCRATCH parameter disk devices” on page 69.
threshold is reached. to a value greater than
zero. For more information about
volumes on FILE devices,
Operators can also define see “Configuring FILE
space triggers to sequential volumes on disk
automatically allocate space devices” on page 70.
when a threshold is
reached.
Tivoli Storage Manager Server caching is available, Server caching is not Caching can improve how
server caching (after files but overhead is incurred in necessary because access quickly the Tivoli Storage
are migrated to the next freeing the cached space. times are comparable to Manager server retrieves
storage pool in the storage For example, as part of a random access (DISK) files during client restore or
pool hierarchy) backup operation, the access times. retrieve operations.
server must erase cached
files to make room for For more information, see
storing new files. “Caching in disk storage
pools” on page 284.

66 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Table 7. Comparing random access and sequential access disk devices (continued)
Function Random Access (DISK) Sequential Access (FILE) Notes®
Recovery of disk space When caching is enabled, The server recovers disk For more information about
the space occupied by space in a process called reclamation, see
cached files is reclaimed on reclamation, which involves “Reclaiming space in
demand by the server. copying physical files to sequential-access storage
another volume, making pools” on page 367.
When caching is disabled, the reclaimed volume
the server recovers disk available for reuse. This
space immediately after all minimizes the amount of
physical files are migrated overhead because there is
or deleted from within an no mount time required.
aggregate.
Aggregate reconstruction Not available; the result is Aggregate reconstruction An aggregate is two or more
wasted space occurs as part of the files grouped together for
reclamation process. It is storage purposes. Most
also available by using the data from backup-archive
RECONSTRUCT parameter on clients is stored in
the MOVE DATA and MOVE aggregates. Aggregates
NODEDATA commands. accumulate empty space as
files are deleted, expire, or
as they are deactivated in
active-data pools.

For more information, see

“How Tivoli Storage
Manager reclamation
works” on page 368.
Available for use as copy Not available Available Copy storage pools and
storage pools or active-data active-data pools provide
pools additional levels of
protection for client data.

For more information, see

“Backing up primary
storage pools” on page 883.
File location Volume location is limited FILE volumes use
by the trigger prefix or by directories. A list of
manual specification directories can be specified.
If directories correspond
with file systems,
performance is optimized.

Chapter 4. Magnetic disk devices 67

Table 7. Comparing random access and sequential access disk devices (continued)
Function Random Access (DISK) Sequential Access (FILE) Notes®
Restoring the database to See Notes Use the REUSEDELAY Use the AUDIT VOLUME
an earlier level parameter to retain command to identify
volumes in a pending state. inconsistencies between
Volumes are not rewritten information about a volume
until the specified number in the database and the
of days have elapsed. actual content of the
During database volume. You can specify
restoration, if the data is whether the Tivoli Storage
physically present, it can be Manager server resolves the
accessed after DSMSERV database inconsistencies it
RESTORE DB. finds.

For more information about

auditing volumes, see
“Auditing storage pool
volumes” on page 887. For
more information about
reuse delay, see “Delaying
reuse of volumes for
recovery purposes” on page
886.

For command syntax, refer

to the Administrator's
Reference.
Migration Performed by node. Performed by volume. Files For more information, see
Migration from are not migrated from a “Migrating disk storage
random-access pools can volume until all files on the pools” on page 273.
use multiple processes. volume have met the
threshold for migration
delay as specified for the
storage pool. Migration
from sequential-access
pools can use multiple
processes.
Storage pool backup Performed by node and file Performed by volume. For For more information, see
space. Every storage pool a primary pool, there is no “Storage pools” on page
backup operation must need to scan every object in 240.
check every file in the the primary pool every
primary pool to determine time the pool is backed up
whether the file must be to a copy storage pool.
backed up.
Copying active data Performed by node and file Performed by volume. For For more information, see
space. Every storage pool a primary pool, there is no “Storage pools” on page
copy operation must check need to scan every object in 240.
every file in the primary the primary pool every
pool to determine whether time the active data in the
the file must be copied. pool is copied to an
active-data pool.

68 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Table 7. Comparing random access and sequential access disk devices (continued)
Function Random Access (DISK) Sequential Access (FILE) Notes®
Transferring data from Major benefits by moving Some benefit by moving For more information, see
non-collocated to collocated data from non-collocated data from non-collocated “Keeping client files
storage storage to DISK storage, storage to FILE storage, together using collocation”
and then allowing data to and then moving data to on page 357.
migrate to collocated collocated storage.
storage.

See “Restoring files to a

storage pool with
collocation enabled” on
page 908 for more
information.
Shredding data If shredding is enabled, Shredding is not supported For more information, see
sensitive data is destroyed on sequential access disk “Securing sensitive client
after it is deleted from a devices. data” on page 539.
storage pool. Write caching
on a random access devices
should be disabled if
shredding is enforced.
Data deduplication Not available Duplicate data in primary, For more information, see
copy, and active-data pools “Deduplicating data” on
can be identified and page 285.
removed, reducing the
overall amount of time that
is required to retrieve data
from disk.

Configuring random access volumes on disk devices

Tivoli Storage Manager provides a predefined DISK device class that is used with
all disk devices.

Procedure

To set up a random access volume on disk to store client backup, archive, or

space-managed data, complete the following steps:
1. Define a storage pool that is associated with the DISK device class. For
example, enter the following command to define a storage pool named
ENGBACK1:
define stgpool engback1 disk maxsize=5G highmig=85 lowmig=40
2. Prepare a volume for use in a random access storage pool by defining the
volume. If you do not specify a full path name, the command uses the current
path.
Define a 21 MB volume for the ENGBACK1 storage pool that is in the path
c:\program files\tivoli\tsm\server and named stgvol.002. Enter the
following command:
define volume engback1
c:\program files\tivoli\tsm\server\stgvol.002’ formatsize=21
3. Do one of the following:

Chapter 4. Magnetic disk devices 69

 v Specify the new storage pool as the destination for client files that are backed
up, archived, or migrated, by modifying existing policy or creating new
policy. See Chapter 13, “Implementing policies for client data,” on page 475
for details.
v Place the new storage pool in the storage pool migration hierarchy by
updating an already defined storage pool. See “Example: Updating storage
pools” on page 250.

Note: Define storage pool volumes on disk drives that reside on the server
system, not on remotely mounted file systems. Network attached drives can
compromise the integrity of the data that you are writing.
Related concepts:
“Disk devices” on page 40
Related tasks:
“Defining storage pool volumes” on page 256

Configuring FILE sequential volumes on disk devices

Magnetic disk storage uses files as volumes that store data sequentially, as on tape
volumes. The space for FILE volumes is managed by the operating system rather
than by Tivoli Storage Manager.
Related tasks:
“Defining sequential-access disk (FILE) device classes” on page 187
“Step 1: Defining device classes for database backups” on page 871
“Planning for sequential media used to export data” on page 754
“Defining storage pool volumes” on page 256
“Preparing volumes for sequential-access storage pools” on page 255

Varying disk volumes online or offline

To perform maintenance on a disk volume or to upgrade disk hardware, you can
vary a disk volume offline. If Tivoli Storage Manager encounters a problem with a
disk volume, the server automatically varies the volume offline.

About this task

Task Required Privilege Class
Vary a disk volume online or offline System or operator

For example, to vary the disk volume named STGVOL.POOL001 offline, enter:
vary offline stgvol.pool001

You can make the disk volume available to the server again by varying the volume
online. For example:
vary online stgvol.pool001

70 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Cache copies for files stored on disk
When you define a storage pool that uses disk random access volumes, you can
choose to enable or disable cache. When you use cache, a copy of the file remains
on disk storage even after the file is migrated to the next pool in the storage
hierarchy, for example, to tape. The file remains in cache until the space it occupies
is needed to store new files.

Using cache can improve how fast a frequently accessed file is retrieved. Faster
retrieval can be important for clients that are storing space-managed files. If the file
needs to be accessed, the copy in cache can be used rather than the copy on tape.
However, using cache can degrade the performance of client backup operations
and increase the space needed for the database.
Related tasks:
“Caching in disk storage pools” on page 284

Freeing space on disk

As client files expire, the space they occupy is not freed for other uses until you
run expiration processing on the server.

Expiration processing deletes information from the database about any client files
that are no longer valid according to the policies you have set. For example,
suppose that four backup versions of a file exist in server storage, and only three
versions are allowed in the backup policy (the management class) for the file.
Expiration processing deletes information about the oldest of the four versions of
the file. The space that the file occupied in the storage pool becomes available for
reuse.

You can run expiration processing by using one or both of the following methods:
v Use the EXPIRE INVENTORY command.
v Set the EXPINTERVAL server option and specify the interval so that expiration
processing runs periodically.

Shredding occurs only after a data deletion commits, but it is not necessarily
completed immediately after the deletion. The space occupied by the data to be
shredded remains occupied while the shredding takes place, and is not available as
free space for new data until the shredding is complete. When sensitive data is
written to server storage and the write operation fails, the data that was already
written is shredded.
Related concepts:
“Securing sensitive client data” on page 539
Related reference:
“Running expiration processing to delete expired files” on page 512

Chapter 4. Magnetic disk devices 71

Scratch FILE volumes
When the server needs a new volume, the server automatically creates a file that is
a scratch volume, up to the number you specify.

You can specify a maximum number of scratch volumes for a storage pool that has
a FILE device type.

When scratch volumes used in storage pools become empty, the files are deleted.
Scratch volumes can be located in multiple directories on multiple file systems.

Volume history file and volume reuse

When you back up the database or export server information, Tivoli Storage
Manager records information about the volumes that are used for these operations
in the volume history. Tivoli Storage Manager does not allow you to reuse these
volumes until you delete the volume information from the volume history.

To reuse volumes that were previously used for database backup or export, use the
DELETE VOLHISTORY command.

Note: With Tivoli Storage Manager Extended Edition, the disaster recovery
manager (DRM) function automatically deletes volume information during
processing of the MOVE DRMEDIA command.
Related tasks:
“Protecting the volume history file” on page 878
Chapter 32, “Disaster recovery manager,” on page 1005

72 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Chapter 5. Attaching devices for the server
For IBM Tivoli Storage Manager to use a device, you must attach the device to
your server system and install the appropriate device driver.

About this task

Attached devices must be on their own host bus adapter (HBA) and must not
share with other devices types (disk, CDROM, and so on). IBM tape drives have
some special requirements for HBAs and associated drivers.

Tasks:
“Attaching a manual drive”
“Attaching an automated library device” on page 74
“Device alias names” on page 75
“Selecting a device driver” on page 78
“Installing the Centera SDK for Centera shared libraries” on page 83

Attaching a manual drive

Attaching manual drives to your system allows you to utilize storage.

Procedure

Complete the following steps to attach a manual drive:

v For Fibre Channel (FC), Fibre Channel over Ethernet (FCoE), Serial Attached
SCSI (SAS), or iSCSI adapters, complete the following steps:
1. Install an adapter card and the associated drivers.
2. Attach the device to the server system hardware by following the
manufacture's instructions.
3. Install the device drivers for attached tape devices.
v For SCSI adapters, complete the following steps:
1. Install the SCSI adapter card in your system, if not already installed.
2. Determine the SCSI IDs available on the SCSI adapter card to which you are
attaching the device. Find one unused SCSI ID for each drive.
3. Follow the manufacturer's instructions to set the SCSI ID for the drive to the
unused SCSI IDs that you found. Usually this means setting switches on the
back of the device or through the device operator's panel

Note: Each device that is connected in a chain to a single SCSI bus must be
set to a unique SCSI ID. If each device does not have a unique SCSI ID, you
might have serious system problems.
4. Follow the manufacturer's instructions to attach the device to your server
system hardware.

Attention:
a. Power off your system before attaching a device to prevent damage to the
hardware.

© Copyright IBM Corp. 1993, 2013 73

 b. Attach a terminator to the last device in the chain of devices that are
connected on one SCSI adapter card.
5. Install the appropriate device drivers. See “Selecting a device driver” on page
78.
6. Determine the name for the device and record the name. This information
can help you when you must perform operations such as adding volumes.
Keep the records for future reference.

Attaching an automated library device

You can attach an automated library device to your system to store your data on
tapes.

Procedure

Perform the following steps to attach an automated library device:

1. Install the SCSI adapter card in your system, if not already installed.
2. Determine the SCSI IDs available on the SCSI adapter card to which you are
attaching the device. Find one unused SCSI ID for each drive, and one unused
SCSI ID for the library or autochanger controller.

Note: In some automated libraries, the drives and the autochanger share a
single SCSI ID, but have different LUNs. For these libraries, only a single SCSI
ID is required. Check the documentation for your device.
3. Follow the manufacturer's instructions to set the SCSI ID for the drives and
library controller to the unused SCSI IDs that you found. Usually this means
setting switches on the back of the device.

Note: Each device that is connected in a chain to a single SCSI bus must be set
to a unique SCSI ID. If each device does not have a unique SCSI ID, you might
have serious system problems.
4. Follow the manufacturer's instructions to attach the device to your server
system hardware.

Attention:
a. Power off your system before attaching a device to prevent damage to the
hardware.
b. Attach a terminator to the last device in the chain of devices that are
connected on one SCSI adapter card. Detailed instructions should be in the
documentation that came with your hardware.
5. Install the appropriate device drivers. See “Selecting a device driver” on page
78.
6. Determine the name for each drive and for the library, and record the names.
This information can help you when you need to perform operations such as
adding volumes to an autochanger. Keep the records for future reference.
7. For the IBM Tivoli Storage Manager server to access a SCSI library, set the
device for the appropriate mode. This is usually called random mode; however,
terminology can vary from one device to another. Refer to the documentation
for your device to determine how to set it to the appropriate mode.

Note:

74 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 a. Some libraries have front panel menus and displays that can be used for
explicit operator requests. However, if you set the device to respond to such
requests, it typically will not respond to IBM Tivoli Storage Manager
requests.
b. Some libraries can be placed in sequential mode, in which volumes are
automatically mounted in drives by using a sequential approach. This mode
conflicts with how IBM Tivoli Storage Manager accesses the device.

Device alias names

The server uses alias names to identify tape devices to the IBM Tivoli Storage
Manager device driver.

Device names for the Tivoli Storage Manager device driver differ from device
names for the Windows device driver. For example, an automated library device
might be known as lb0.0.0.1 to the Tivoli Storage Manager device driver and as
changerx (where x is a number 0–9), to the Windows device driver.

If you configure devices by using Tivoli Storage Manager commands, you must
provide the device names as parameters to the DEFINE PATH command. If you
modify device driver control, you must provide alias name information in the
Device Exclude List. The names can be either:
v Drive letters, for devices that are attached as local, removable file systems
v Alias names, for devices that are controlled by either the Tivoli Storage Manager
device driver or the Windows device drivers

Alias names replace the real device names in Tivoli Storage Manager commands
and screens. The Tivoli Storage Manager device driver communicates with devices
by using the alias names.

Alias names are displayed in the form mtx.y.z.n or lbx.y.z.n, where:

mt Indicates that the device is a tape device. For example:
mt3 (Tape drive at SCSI ID 3, LUN 0, bus 0, port 0).
mt5.0.0.1 (Tape drive at SCSI ID 5, LUN 0, bus 0, port 1).
lb Indicates that the device is the controller for an automated library device.
For example:
lb4.1 (Library at SCSI ID 4, LUN 1, bus 0, port 0).
x Indicates the SCSI ID for the targeted device.
y Indicates the logical unit number (LUN) for the targeted device.
z Indicates the bus number that is supported by the adapter device driver.
n Indicates the port number for the SCSI adapter device driver.
Note: Alias names can be abbreviated when the trailing numbers are zeros.

Chapter 5. Attaching devices 75

Displaying information about devices with tsmdlst
Use the tsmdlst utility to view device names and other information about
medium-changer, and tape devices on the system.

Options

After devices are configured, you can run the tsmdlst utility to display device
information. The utility is in the devices bin directory, which is \Program
Files\Tivoli\TSM\console by default.
/computer=computer_name
Specifies the name of the computer for which devices are listed. The default is
the local system.
/detail
Displays details on devices in the list. By default, a summary is shown.
/all
Displays information about all types of devices. By default, only tape drives
and tape libraries are included in the results.
/nogenerictapecheck
Skips the step for opening detected drives to see if they are supported for the
Tivoli Storage Manager GENERICTAPE device type.
/nohbacheck
Skips the step for HBA API detection, which might speed up processing. This
option can be useful when debugging is needed.
/trace
Used for diagnostic purposes. Stores trace output in the tsmdlst_trace.txt file.
/? Displays usage information about tsmdlst and its parameters.
/xinquiry
Provides an alternate way to obtain serial number and worldwide name
information. This option is used only for devices that are supported by the
IBM tape device driver. The following parameters are specific to the /xinquiry
option:
/processAll
Indicates that the process loops until all devices are processed.
/maxRetries=#
Indicates the maximum number of attempts to open each drive. This
option requires the /processAll option.
/genpathfile
Use this option to generate a list of devices and serial numbers. The
tsmdlst_pathfile.txt file is written with information for /genmacropathsync
and /genmacropathoffline.
/includelib
If specified with /genpathfile, the list of devices includes libraries in
addition to drives.
/genmacropathsync
Generates a macro to synchronize Tivoli Storage Manager paths for the storage
agent based on serial number. A drive must have a serial number defined to
Tivoli Storage Manager for this option to work.

76 IBM Tivoli Storage Manager for Windows: Administrator's Guide

/genmacropathoffline
Generates a macro to update Tivoli Storage Manager paths for the storage
agent to online or offline status based on drive accessibility. A drive is
accessible if an operating system open call results in: ERROR_SUCCESS,
ERROR_BUSY or ERROR_ACCESS_DENIED. This option only works for
devices that are using the IBM device driver. A symbolic name, for example
\\.\tape0, is required to open a device.
The following options are specific to /genmacropathsync and
/genmacropathoffline
/server=servername
Specifies the name of the server that the storage agent is using.
/stagent=stagentname
Specifies the name of the storage agent.
/tcps=address
Specifies the Tivoli Storage Manager server address.
/tcpp=port
Specifies the Tivoli Storage Manager server port. The default is 1500.
/id=id
Specifies the Tivoli Storage Manager administrative ID.
/pass=password
Specifies the Tivoli Storage Manager administrative password.
/devicetype=drivetype
Specifies that a on the device type of the drive, for example, LTO. This
option is case-sensitive and optional.
/libraryname=libname
Filters on the library name of the drive, for example LTO3584. This option
is case-sensitive and optional.
/execmacropathsync
Issues the path synchronize macro to the Tivoli Storage Manager server.
/execmacropathoffline
Issues the path offline macro to the Tivoli Storage Manager server.
/addpaths
Adds define and update path statements. This option is used with
/genmacropathsync.
/verbose
Lists both drive and path information returned from the Tivoli Storage
Manager server and contents of the pathfile.
/encodednames
If a path is set to online=no, the device name encodes time stamp, error,
and device as the updated device name.

Example: Display information about devices

Display information about tape devices and tape libraries for a local system,
ATLAS, by issuing the tsmdlst utility:
tsmdlst

Chapter 5. Attaching devices 77

 Note: The device name displayed is the alias name that can be used in the DEFINE
PATH command and the UPDATE PATH command. The alias name is not the actual
device name.
Computer Name: ATLAS
TSM Device Driver: TSMScsi - Running

2 HBAs were detected.

Manufacturer Model Driver Version Firmware Description

-----------------------------------------------------------------------------------------
QLogic Corp QLA2462 [Link] [Link] 5.04.02 QLogic QLA2462 Fibre Channel Adapter
QLogic Corp QLA2462 [Link] [Link] 5.04.02 QLogic QLA2462 Fibre Channel Adapter

TSM Name ID LUN Bus Port SSN WWN TSM Type Driver Device Identifier
-----------------------------------------------------------------------------------------
mt0.0.0.3 0 0 0 3 HU1206LY0B ..638D LTO NATIVE HP Ultrium 5-SCSI Y50S
mt1.0.0.3 1 0 0 3 HU1206LY9N ..6390 LTO NATIVE HP Ultrium 5-SCSI Y5AS
mt2.4.0.3 2 4 0 3 8395261003 ..C358 LTO IBM IBM ULT3580-TD3 5AT0
lb3.0.0.3 3 0 0 3 1333508999 ..7A14 LIBRARY TSM ATL P3000 0100
mt3.1.0.3 3 1 0 3 1333508000 ..7A14 DLT TSM QUANTUM DLT7000 0100

Selecting a device driver

To use a tape device, you must install the appropriate device driver.

About this task

Windows device drivers are recommended for all devices. IBM device drivers are
available for most IBM labeled devices. If a Windows device driver is not available
for your device, and the device is supported by the Tivoli Storage Manager device
driver, you can use the Tivoli Storage Manager device driver for your device.

Starting with Tivoli Storage Manager Version 6.2, support for native drivers
through SCSI Passthru is available. You can choose to use a Windows Hardware
Qualification Lab certified native device driver instead of the Tivoli Storage
Manager device driver to control devices. Devices that are already controlled by
the Tivoli Storage Manager device driver can be switched to a native driver
without updating drive or device class definitions.

Drivers for IBM devices

Tivoli Storage Manager supports IBM device drivers for some IBM devices.

Install the IBM device driver for the following IBM devices:
IBM 3494 library
IBM Ultrium 3580, TS2230, TS2340 tape drives
IBM 3581, 3582, 3583, 3584 tape libraries
IBM 3590, 3590E, and 3590H tape drives
IBM 3592 and TS1120 tape drives
IBM TS3100, TS3200, TS3310, TS3400, and TS3500 tape libraries

IBM device drivers are available at the Fix Central support website:
1. Go to the Fix Central website: [Link]
2. Select System Storage for the Product Group menu.
3. Select Tape systems for the System Storage menu.
4. Select Tape drivers and software for the Tape systems menu.
5. Select Tape device drivers for the Tape drivers and software menu.

78 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 6. Select your operating system for the Platform menu.
It is recommended that you install the most current driver available.

For IBM device driver documentation, see the following links:

v IBM Tape Device Drivers Installation and User's Guide: [Link]
support/[Link]?uid=ssg1S7002972
v IBM Tape Device Drivers Programming Reference: [Link]
[Link]?uid=ssg1S7003032

Tivoli Storage Manager supports all devices that are supported by IBM device
drivers. However, Tivoli Storage Manager does not support all the
operating-system levels that are supported by IBM device drivers. For the most
up-to-date list of devices and operating-system levels that are supported by IBM
device drivers, see the Tivoli Storage Manager Supported Devices website at
[Link]
IBM_TSM_Supported_Devices_for_AIXHPSUNWIN.html.

Tivoli Storage Manager Support for Multipath I/O with IBM Tape
Devices
Multipath I/O is the use of different paths to get to the same physical device (for
example, through multiple host bus adapters, switches, and so on). Multipathing
helps ensure that there is no single point of failure.

The IBM tape device driver provides multipathing support so that if a path fails,
the Tivoli Storage Manager server can use a different path to access data on a
storage device. The failure and transition to a different path are undetected by the
server. The IBM tape device driver also uses multipath I/O to provide dynamic
load balancing for enhanced I/O performance.

A computer has a unique SCSI address and Tivoli Storage Manager device name
for each path to a changer or tape device, even though some paths may be
redundant. For each set of redundant paths, you must define only one path to
Tivoli Storage Manager using one of the corresponding Tivoli Storage Manager
device names.

You can determine which Tivoli Storage Manager device names are redundant by
using a tool such as tsmdlst to review the device serial numbers. If multiple Tivoli
Storage Manager changer or tape device names have the same serial number, then
they are redundant and you must define only one to Tivoli Storage Manager.

For an overview of path failover and load balancing, as well as information about
how to enable, disable, or query the status of path failover for each device, see the
IBM Tape Device Drivers Installation and User's Guide.

Drivers for non IBM devices

Use the Windows device driver for tape drives and libraries if one is supplied with
Windows. If the Windows device driver is not available and the device is
supported, you can use the IBM Tivoli Storage Manager device driver to control
the device.

The Tivoli Storage Manager device driver is installed with the server. The Tivoli
Storage Manager device driver uses persistent reservation for some tape drives. See
technote 1470319 at [Link]
for details.

Chapter 5. Attaching devices 79

 For devices that are not currently supported by the Tivoli Storage Manager device
driver, the Windows driver might be suitable.
v Removable media devices (attached as local file systems) require the Windows
device driver.
v Unsupported tape drives require the Windows device driver and must be used
with the GENERICTAPE device class. For more information, see the DEFINE
DEVCLASS - GENERICTAPE command in the Administrator's Reference.

For details on device driver installation directories, see Installation directories in

the Installation Guide.

Installing device drivers for IBM 3494 libraries

You can install device drivers to use IBM 3494 tape libraries. The IBM tape library
driver consists of the ibmatl (a service) and other components.

About this task

To install the device driver for an IBM 3494 Tape Library Dataserver, refer to the
IBM TotalStorage Tape Device Drivers Installation and User's Guide.

To define a path for the library, you can determine the symbolic name of the
library by verifying the value entered in the C:\winnt\[Link] file. For example,
if the symbolic name for the library in the C:\winnt\[Link] file is 3494a, then
this is the name of your device. Drives in the library are set up separately.

Installing the Tivoli Storage Manager device driver

The Tivoli Storage Manager device driver is installed into the driver store through
the Device Driver Installation Wizard. The wizard is displayed during the Tivoli
Storage Manager device driver package installation.

About this task

Before installing a new version of the Tivoli Storage Manager device driver,
uninstall the previous version. Then complete this procedure during installation of
the device driver package.

Procedure
1. When the Device Driver Installation Wizard welcome panel displays, select
Next and proceed through the panels to install the device drivers.

Note: During installation, the system might display a Windows Security dialog
box asking if you would like to install the device software. Place a check mark
on Always trust software from "IBM Corporation" and select Install
2. After your device drivers have been installed, the final panel in the wizard is
displayed. Select Finish to complete the installation.

Results

After a successful installation, use the Device Manager to configure devices with
the Tivoli Storage Manager device driver.

80 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Uninstalling the Tivoli Storage Manager device driver
Uninstall the Tivoli Storage Manager device driver any time you are planning on
upgrading to a more current version.

Procedure

Complete this procedure to uninstall the Tivoli Storage Manager device driver.
1. From your Windows Control Panel, navigate to Programs and Features.
2. Remove or uninstall the IBM Tivoli Storage Manager Device Driver entry.
3. Do not manually remove the Windows Driver Package entries for tsmscsi.
These packages are automatically removed after the IBM Tivoli Storage Device
Driver program is removed in Step 2. These entries, however, might still appear
in the Add or Remove Programs or Programs and Features windows until the
window is refreshed.

Using Windows device drivers with the GENERICTAPE format

When a tape drive is controlled by the Windows device driver, in addition to the
Tivoli Storage Manager tape formats, you might be able to use the drive with the
GENERICTAPE format.

Because some tape drives do not have all of the functions that the Tivoli Storage
Manager server requires, not all tape drives can be used with the GENERICTAPE
device class. To determine if you can use the Windows device driver with a
specific tape drive, see “Creating a file to list devices and their attributes” on page
82. You can find the setup procedure for these devices at “Configuring devices not
supported by the Tivoli Storage Manager server” on page 88.

When using Windows device drivers with the GENERICTAPE device class, be
aware of the following:
v Tivoli Storage Manager does not recognize the device type.
If you add devices and intend to use the GENERICTAPE device class, you
should understand that the server does not know device types and recording
formats. For example, if you use a Windows device driver for a 4MM drive
using the DDS2 recording format, Tivoli Storage Manager knows only that the
device is a tape drive. The default recording format is used.
The server cannot prevent errors when it does not know the device type. For
example, if one GENERICTAPE device class points to a manual library device
containing a 4MM drive and an 8MM drive, the server might make an
impossible request: mount a 4MM cartridge into an 8MM drive.
v Device problems might be more difficult to solve.
The server cannot report I/O errors with as much detail. The server can obtain
only minimal information for display in the server console log.

It is recommended that you use the GENERICTAPE device class only with
unsupported tape devices.

Chapter 5. Attaching devices 81

 Creating a file to list devices and their attributes
A device can be used with the GENERICTAPE device class if it is supported by the
Windows device driver and has specific capabilities.

About this task

The device should be able to complete the following tasks:

v Write in variable mode
v Write filemarks
v Can forward/reverse space blocks
v Can forward/reverse space filemarks

Procedure

A file listing devices and their attributes can be created by completing the
following procedure.
1. Click Start > Programs > Command Prompt on the Windows Start button. The
Command Prompt dialog appears.
2. Change directory to the directory in which the Tivoli Storage Manager Server
was installed. For default installations, the path resembles the following:
c:\program files\tivoli\tsm\server
3. To create the file, type in the following command:
tsmdlst > [Link]
4. To view the file, type in the following command:
notepad [Link]

Controlling devices with the Tivoli Storage Manager device

driver
On Windows systems, devices are automatically controlled by the default Windows
device driver, even if you install the Tivoli Storage Manager driver (tsmscsi).

About this task

Tape drives can be automatically controlled by the Tivoli Storage Manager device
driver if the Windows device drivers are not available. If the devices are not
automatically configured and controlled by the Tivoli Storage Manager device
driver, you must manually update the controlling driver for each device that you
want controlled by the tsmscsi device driver.

Procedure

Perform this procedure from the Device Manager Console:

1. Right click on the device and select Properties. Select the Driver tab and Driver
File Details. This allows you to see the driver that is controlling your device.
2. You will need to configure the device to be used by [Link] by
right-clicking on the device and selecting Update Driver or by selecting Action
and then Update Driver. The Hardware Update Wizard appears.
3. Select Browse my computer for driver software.
4. Select Let me pick from a list of device drivers on my computer.
5. Select the IBM Tivoli Storage Manager device driver to control the device.
6. Click Next.

82 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 7. Verify that the device is configured correctly for tsmscsi:
a. Right-click the device and select Properties.
b. Select the driver tab and driver details.

Installing the Centera SDK for Centera shared libraries

Beginning with Tivoli Storage Manager Version 5.5, Centera shared libraries are not
installed with the server. To use Centera with Tivoli Storage Manager, the Centera
SDK must be installed.

Procedure

Perform the following steps when setting up the Tivoli Storage Manager server to
access Centera.
1. Install the Tivoli Storage Manager server.
2. If you are upgrading from a previous level of Tivoli Storage Manager, delete
the following Centera SDK library files from the directory where the server was
installed:
[Link]
[Link]
[Link]
[Link]
3. Contact your EMC representative to obtain the installation packages and
instructions to install the Centera SDK Version 3.2 or later.
4. Install the Centera SDK. During the installation, take note of the directory
where the Centera SDK is installed.
a. Unzip and untar the package in a working directory.
b. Copy the files in the lib32 directory to the directory with the server
executable ([Link]).
5. Start the Tivoli Storage Manager server and set up the policy, device class, and
storage pools for Centera.

Chapter 5. Attaching devices 83

84 IBM Tivoli Storage Manager for Windows: Administrator's Guide
Chapter 6. Configuring storage devices
Configure storage devices by defining the objects that represent the devices to the
Tivoli Storage Manager server.

About this task

For the most up-to-date list of supported devices and operating-system levels, see
the Supported Devices website at [Link]
products/support/IBM_TSM_Supported_Devices_for_AIXHPSUNWIN.html.

Concepts:
“Planning for device configuration” on page 86
“Mixed device types in libraries” on page 52
“Server options that affect storage operations” on page 62
“Impact of device changes on the SAN” on page 125
“Defining devices and paths” on page 173

Use the following table to locate instructions for specific tasks:

Tasks:
“Configuring devices not supported by the Tivoli Storage Manager server” on page 88
“Configuring removable media devices” on page 89
“Configuring devices using Tivoli Storage Manager commands” on page 91
“Configuring Tivoli Storage Manager servers to share SAN-connected devices” on page 119
“Configuring Tivoli Storage Manager for LAN-free data movement” on page 122
“Validating your LAN-free configuration” on page 123
Chapter 9, “Using NDMP for operations with NAS file servers,” on page 203
“Configuring IBM 3494 libraries” on page 96
“ACSLS-managed libraries” on page 109
“Troubleshooting device configuration” on page 123

You can perform configuration tasks by using the command-line interface. For
information about Tivoli Storage Manager commands, see the Administrator's
Reference or issue the HELP command from the command line of a Tivoli Storage
Manager administrative client.

The tasks require an understanding of Tivoli Storage Manager storage objects. For
an introduction to these storage objects, see “Tivoli Storage Manager storage
objects” on page 36.

© Copyright IBM Corp. 1993, 2013 85

Planning for device configuration
Before Tivoli Storage Manager can use a removable media device, you must plan,
attach, and configure the device.

Procedure

The following steps give an overview of the device configuration process:

1. Plan for the device.
2. Attach the device to the server. See the manufacturer documentation of the
device for information about attaching the device.
3. Start the appropriate device driver. Both the Tivoli Storage Manager device
driver and the native Windows device driver can be used. It might be
necessary to specify which device driver acquires which devices.
4. Configure the device.

Important: In most cases, the server expects to have exclusive use of devices
that are defined to the server. Attempting to use a Tivoli Storage Manager
device with another application might cause the server operations to fail. This
restriction does not apply to 3494 library devices, or when you use a storage
area network (SAN) to share library devices.
5. Determine the media type and device type for client data.
You can link clients to devices by directing client data to a type of media. For
example, accounting department data might be directed to LTO Ultrium tapes,
and as a result the server would select LTO Ultrium devices.
You can direct data to a specific media type through Tivoli Storage Manager
policy. When you register client nodes, you specify the associated policy.
For configuring devices by using Tivoli Storage Manager commands, you must
also define or update the Tivoli Storage Manager policy objects that links
clients to the pool of storage volumes and to the device.
6. Register clients to the policy domain defined or updated in Step 5. This step
links clients and their data with storage volumes and devices.
7. Prepare media for the device.
Label tapes before they can be used. For automated library devices, you must
add the media to the volume inventory by checking media into the library
device.

Results

Manually configuring devices

If the tape drives are not automatically configured with the Tivoli Storage Manager
driver, you must manually configure them.

Before you begin

When the Tivoli Storage Manager device driver is installed, some tape drives, and
medium changers might automatically be configured if the Windows device
drivers for the devices are not available.

Complete the following steps to determine whether a device was automatically

configured with the Tivoli Storage Manager device driver:
1. Start the Device Manager.

86 IBM Tivoli Storage Manager for Windows: Administrator's Guide

2. Right-click the device and select Properties.
3. Select the Driver tab and Driver File Details. You can see the device driver
that is controlling your device.

Alternatively, you can run [Link] in the devices bin directory, which is
\Program Files\Tivoli\TSM\console by default, to determine whether the devices
were configured with the Tivoli Storage Manager device driver. If the devices were
not configured with the Tivoli Storage Manager device driver, the TSM Type
shows GENERICTAPE.

Procedure

Complete the following steps to manually configure devices for the Tivoli Storage
Manager device driver:
1. Locate the device in the Device Manager console ([Link]) and select it.
Tape drives are listed under Tape drives, and medium changers are under
Medium Changers.
2. Configure the device for use by [Link].
a. Click ActionUpdate Driver and select Update Driver. You can also
right-click on the device and select Update Driver Software.
b. Select Browse my computer for driver software.
3. Select Let me pick from a list of device drivers on my computer.
4. Click Next.
5. Select one of the following options, depending on the device that you are
configuring:
v For a tape drive, select IBM Tivoli Storage Manager for Tape Drives.
v For a medium changer, select IBM Tivoli Storage Manager for Medium
Changers.
6. Click Next.
7. Click Close.
8. Verify that the device was configured correctly for tsmsci.
a. Right-click on the device and select Properties.
b. Select the Driver tab and Driver Details.
c. The Driver Details panel shows the device driver that is controlling the
device. The name of the device driver is [Link] for 32-bit Windows
Server 2008, and [Link] for 64-bit Windows Server 2008.

What to do next

For Windows Server 2008 Server Core, devices cannot be configured through
Device Manager. If the devices are not automatically configured, you must use the
Tivoli Storage Manager CHANGE DEVDRIVER command to configure the devices. For
more information, see technote 1320150 at [Link]
[Link]?uid=swg21320150.

Chapter 6. Configuring storage devices 87

 Configuring devices not supported by the Tivoli Storage
Manager server
You can configure devices that run with either their own device driver or with a
Windows device driver when the devices meet Tivoli Storage Manager
requirements.

About this task

Devices that are not supported by the Tivoli Storage Manager server can be added
by using the Tivoli Storage Manager command line.

Procedure
1. Attach the device to the system.
Follow the manufacturer's instructions to attach the device to the system.
2. Set up the appropriate Windows device driver for the device.
3. Configure the device. The following guidelines must be followed:
v Define the drive path with GENERICTAPE=Yes.
v The device class must have a device type of GENERICTAPE.
v Define a different device class and a different manual library device for every
unique device type that is controlled by the Windows device driver. For
example, to use a 4 mm drive and an 8 mm drive, define two manual
libraries, and two device classes (both with device type GENERICTAPE).
4. Determine your backup strategy.
Determine which device the server backs up client data to, and whether client
data is backed up to disk, and then migrated to tape, or if it is backed up
directly to tape.
5. Update the Tivoli Storage Manager policy.
Define the Tivoli Storage Manager policy that links client data with media for
the device.
6. Label volumes.

Results

For more information, see the following topics:

“Configuring devices using Tivoli Storage Manager commands” on page 91
“Creating a file to list devices and their attributes” on page 82
“Defining Tivoli Storage Manager storage objects with commands” on page 92
“Defining and updating a policy domain” on page 498
“Planning for server storage” on page 60
“Selecting a device driver” on page 78

88 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Configuring removable media devices
You can add removable media devices by using Tivoli Storage Manager
commands.

About this task

The following guidelines must be followed:

If a removable media device can be formatted with a file system, Tivoli Storage
Manager might be able to use the device. The server recognizes the device as a
device with type REMOVABLEFILE. To use device type REMOVABLEFILE for a device, the
device must adhere to the following principles:
v It must not be supported by a device type that is available for a Tivoli Storage
Manager device class.
v It must be a device with removable media, for example, Iomega Zip or Jaz
drives, CD drive, or DVD drive.
v It must be viewed by the operating system as a removable media drive, and not
as a fixed, hard disk drive. The server cannot use the device if the storage
adapter card makes the removable media drive show up as a fixed disk drive to
the operating system.

Tip: If a data cartridge that is associated with a REMOVABLEFILE device class has
two sides, the server treats each side as a separate Tivoli Storage Manager volume.
Tivoli Storage Manager REMOVABLEFILE device class supports only single-sided
media.

You can use the CD or DVD media as input media on a target Tivoli Storage
Manager server by using the REMOVABLEFILE device class for input. Using the
REMOVABLEFILE device class allows the server to distinguish media volumes by a
volume label to prompt for the next media, and to dismount media.

With CD support for Windows, you can also use CD media as an output device
class. Using CD media as output requires other software that uses a file system on
top of the CD media. This media allows other software to write to a CD by using a
drive letter and file names. The media can be either CD-R (read) or CD-RW
(read/write).

With DVD support for Windows, you can also use DVD media as an output device
class. Using DVD media as output requires other software that uses a file system
on top of the DVD media. DVDFORM software is a tool that comes with some
DVD-RAM device drivers. The DVDFORM software, for example, allows you to
label the media, which must be DVD-RAM, by using uppercase letters and
numbers. When the media is formatted, you can use the LABEL system command
to change the label.

Procedure

To set up a device, complete the following steps:

1. Attach the device to the system.
Follow the manufacturer documentation to attach the device to the system.
2. Set up the appropriate device driver for the device.
3. Configure the device.
The following parameters must be specified:
v The device class must have device type of REMOVABLEFILE.
Chapter 6. Configuring storage devices 89
 v The library type can be either MANUAL or SCSI.
v The device name that is used in defining the drives is the drive letter by
which the system knows the drive.
4. Determine your backup strategy.
Determine which device the server backs up client data to, and whether client
data is backed up to disk, and then migrated to tape, or if it is backed up
directly to tape.
5. Label removable file media.
Utilities are not supplied by the server to format or label CDs or DVDs. You
must label CDs or DVDs with the device utilities from the manufacturer or
Windows utilities because Tivoli Storage Manager does not provide utilities to
format or label these media. The operating system utilities include the Disk
Administrator program (a graphical user interface) and the label command.

Results

For more information, see the following topics:

“Configuring devices using Tivoli Storage Manager commands” on page 91
“Defining Tivoli Storage Manager storage objects with commands” on page 92
“Defining and updating a policy domain” on page 498
“Labeling media” on page 129
“Planning for server storage” on page 60
“Selecting a device driver” on page 78

Example of removable file support (CD):

The steps are included as an example of Tivoli Storage Manager REMOVABLEFILE

support. This example takes an export object and moves it from one server to
another.

About this task

v Server A:
– Define a device class named expfile with a device type of FILE.
define devclass expfile devtype=file directory=c:\data\move maxcap=650M
– Export the node. This command creates a file named CDR03 in the
c:\data\move directory. CDR03 contains the export data for node USER1.
export node user1 filedata=all devclass=expfile vol=CDR03

You can use software for writing CDs to create a CD with volume label
CDR03 that contains the file named CDR03.
v Server B:
– Insert the CD in a drive on the Windows system, for example, E:
– Issue the following Tivoli Storage Manager commands to import the node
data on the CD volume CDR03:
define library manual
define devclass cdrom devtype=removablefile library=manual
define drive manual cddrive
define path server01 cddrive srctype=server desttype=drive
library=manual directory=e:\ device=e:
import node user1 filedata=all devclass=cdrom vol=CDR03

90 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 Example of removable file support (DVD-RAM):

The steps (similar to CD support) are used to move data from one server to
another.

About this task

The following example shows how DVD-RAM drives work inside a SCSI library:
v Server A:
– Configure the device.
– For the library, follow the normal tape library configuration method.
– To configure the DVD-RAM drives, use the following procedure:
1. From your desktop, right click My Computer.
2. Select Device Manager.
3. Select the correct SCSI CD-ROM Device and right click for Properties.
4. Select Drivers.
5. Select Update Driver and choose the [Link] file for the driver.
v Issue the following Tivoli Storage Manager commands to manage the library
functions on the DVD-RAM volume DVD01 (use the library element map in the
IBM Tivoli Storage Manager device support pages for your library to determine
the correct element of each drive):
define library dvdlib libtype-scsi
define drive dvdlib drv1 element 6001
define path sever1 dvdlib srctype=server desttype=library device=lb6.0.0.3
define path server1 drv1 srctype=server desttype=drive
library=dvdlib directory=i:\ device=i:
checkin libv dvdlib search=yes status=scratch
checkout libv dvdlib DVD01 rem=no
define devclass a_class devtype=removablefile library=dvdlib

Configuring devices using Tivoli Storage Manager commands

You can add devices by issuing Tivoli Storage Manager commands.

About this task

The scenario that is documented adds a manual tape device, automated library
devices, and a removable file system device such as an Iomega Jaz drive.

Automated library devices can have more than one type of device. The scenario
shows the case of a library with one type of device (a DLT 8000 drive) and a
library with two types of devices (a DLT 8000 drive and an LTO Ultrium drive).

Procedure

Perform the following steps to add a device:

1. Attach the device to the system. Follow the manufacturer's instructions to
attach the device to the system.
2. Set up the appropriate device driver for the device.
3. Configure the device.
4. Determine which device the server backs up client data to, and whether client
data is backed up to disk and then migrated to tape, or if it is backed up
directly to tape.

Chapter 6. Configuring storage devices 91

 5. Label the media.
6. Add new volumes to the library.

Results

Some of the tasks that are described require an understanding of Tivoli Storage
Manager storage objects. For more information about Tivoli Storage Manager
commands, see the Administrator's Reference.

For more information, see:

“Checking media into automated library devices” on page 130
“Defining and updating a policy domain” on page 498
“Planning for server storage” on page 60
“Selecting a device driver” on page 78
“Tivoli Storage Manager storage objects” on page 36

Defining Tivoli Storage Manager storage objects with

commands
You can use commands to define storage objects. These objects are used to
represent each library device and its drives, as well as their respective paths and
the policy used to manage the media associated with each library device.

About this task

For additional information, see:

“Defining libraries”
“Defining drives in the library” on page 93

Defining libraries
All devices must be defined as libraries. Manual devices require a manual type
library, and most automated devices require the SCSI type library. Automated
libraries also require a path defined to them using the DEFINE PATH command.

About this task

You define libraries with the DEFINE LIBRARY command. See the following
examples of the different ways to define a library:
Manual device
define library manual8mm libtype=manual
Automated library device with one device type
define library autodltlib libtype=scsi

Note: If you have a SCSI library with a barcode reader and you would like
to automatically label tapes before they are checked in, you can specify the
following:
define library autodltlib libtype=scsi autolabel=yes
define path server01 autodltlib srctype=server desttype=library
device=lb3.0.0.0
Automated library device with two device types
define library automixlib libtype=scsi

92 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 define path server01 automixlib srctype=server desttype=library
device=lb3.0.0.0
Removable file system device (Iomega Jaz drive)
define library manualjaz libtype=manual

For more information about defining Tivoli Storage Manager libraries, see
“Defining devices and paths” on page 173.

Defining drives in the library

All drives that you wish to use must be defined to the library. You can define
drives by issuing the DEFINE DRIVE command. You must also issue the DEFINE
PATH command to define the path for each of the drives.

About this task

See the following examples for defining drives in the library:

Manual device
define drive manual8mm drive01
define drive manual8mm drive02
define path server01 drive01 srctype=server desttype=drive
library=manual8mm device=mt1.0.0.0
define path server01 drive02 srctype=server desttype=drive
library=manual8mm device=mt2.0.0.0
Automated library device with one device type
define drive autodltlib dlt_mt4
define drive autodltlib dlt_mt5
define path server01 dlt_mt4 srctype=server desttype=drive
library=autodltlib device=mt4.0.0.0
define path server01 dlt_mt5 srctype=server desttype=drive
library=autodltlib device=mt5.0.0.0

For drives in SCSI libraries with more than one drive, the server requires
the element address for each drive. The element address indicates the
physical location of a drive within an automated library. The server
attempts to obtain the element address directly from the drive. If the drive
is not capable of supplying the information, you must supply the element
address in the drive definition.
Automated library device with two device types
define drive automixlib dlt_mt4
define drive automixlib lto_mt5
define path server01 dlt_mt4 srctype=server desttype=drive
library=automixlib device=mt4.0.0.0
define path server01 lto_mt5 srctype=server desttype=drive
library=automixlib device=mt5.0.0.0

For drives in SCSI libraries with more than one drive, the server requires
the element address for each drive. The element address indicates the
physical location of a drive within an automated library. The server
attempts to obtain the element address directly from the drive. If the drive
is not capable of supplying the information, you must supply the element
address in the drive definition.
Removable file system device (Iomega Jaz drive)
define drive manualjaz drive01

Chapter 6. Configuring storage devices 93

 define path server01 drive01 srctype=server desttype=drive
library=manualJAZ directory=e:\ device=e:

For additional information, see:

“Defining devices and paths” on page 173
“Defining drives” on page 174

Define the device classes that group together similar devices

Each Tivoli Storage Manager device must be a member of a Tivoli Storage
Manager device class. Device classes are collections of similar devices, for example
all 8 mm devices that use the same media format. You can define device classes by
issuing the DEFINE DEVCLASS command.

About this task

See the following examples of defining device classes that group together similar
devices:
Manual device
define devclass tape8mm_class devtype=8mm format=8500 library=manual8mm
Automated library device with one device type
define devclass autodlt_class devtype=dlt format=drive library=autodltlib
Automated library device with two device types
define devclass autodlt_class devtype=dlt format=dlt40 library=automixlib
define devclass autolto_class devtype=lto format=ultriumc library=automixlib

Important: Do not use the DRIVE format, which is the default. Because the
drives are different types, Tivoli Storage Manager uses the format
specification to select a drive. The results of using the DRIVE format in a
mixed media library are unpredictable.
Removable file system device (Iomega Jaz drive)
define devclass jazdisk_class devtype=removablefile library=manualjaz

For detailed information about defining Tivoli Storage Manager device classes, see
“Defining device classes” on page 177.

Creating a storage pool for the device added

Each Tivoli Storage Manager device must be associated with a Tivoli Storage
Manager storage pool to allow it to be used to store client data. Storage pools are
collections of media and like device classes.

About this task

They are organized for a grouping of specific types of media, for example a storage
pool named TAPE8MM_POOL for the device class TAPE8MM_CLASS, and
AUTODLT_POOL for the device class AUTODLT_CLASS. See the following
examples of how to create a storage pool for the added device:
Manual device
define stgpool tape8mm_pool tape8mm_class maxscratch=20
Automated library device with one device type
define stgpool autodlt_pool autodlt_class maxscratch=20
Automated library device with two device types
define stgpool autodlt_pool autodlt_class maxscratch=20

94 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 define stgpool autolto_pool autolto_class maxscratch=20
Removable file system device (Iomega Jaz drive)
define stgpool manualjaz_pool jazdisk_class

For detailed information about defining storage pools, see Chapter 10, “Managing
storage pools and volumes,” on page 239.

Determining backup strategies

Administrators are responsible for creating a backup strategy and implementing it
through Tivoli Storage Manager policy. Typically, a backup strategy determines the
device and media to which client data is written. It also determines if data is
backed up directly to tape or if data is backed up to disk and then later migrated
to tape.

Procedure

For disk-to-tape backups:

1. Set up a storage pool hierarchy
2. Use STANDARD default Tivoli Storage Manager policy

Results

For backups directly to tape, you must create new policy by copying default policy
and modifying it for the desired results.

See “Configuring policy for direct-to-tape backups” on page 522.

Determining the media and device type for client backups

Determine the type of media and the type of device to which the server backs up
client data by changing Tivoli Storage Manager policy.

About this task

See the following examples for how to determine the media and device type for
client backups:
Manual device
To assign client node astro to the direct-to-tape policy named dir2tape,
with the password cadet, enter:
register node astro cadet dir2tape
Automated library devices
To assign client node astro to a direct-to-tape policy domain named
dsk2tape, with the password cadet, enter:
register node astro cadet dsk2tape
Removable file system device (Iomega Jaz drive)
To assign client node astro to a removable media device policy domain
named rmdev, with the password cadet, enter:
register node astro cadet rmdev

Chapter 6. Configuring storage devices 95

Configuring IBM 3494 libraries
An IBM 3494 library can be added only by using Tivoli Storage Manager
commands. One or more Tivoli Storage Manager servers can use a single IBM 3494
library.

About this task

See the following sections:

v “Configuring an IBM 3494 library for use by one server” on page 97
v “Sharing an IBM 3494 library among servers” on page 102
v “Migrating a shared IBM 3494 library to a library manager” on page 104
v “Sharing an IBM 3494 library by static partitioning of drives” on page 105
v “Categories in an IBM 3494 library”

For detailed and current library support information, see the Supported Devices
Web site at: [Link]
IBM_TSM_Supported_Devices_for_AIXHPSUNWIN.html

Categories in an IBM 3494 library

The library manager built into the IBM 3494 library tracks the category number of
each volume in the library. A single category number identifies all volumes used
for the same purpose or application. Category numbers are useful when multiple
systems share the resources of a single library.

Attention: If other systems or other Tivoli Storage Manager servers connect to the
same 3494 library, each must use a unique set of category numbers. Otherwise, two
or more systems may try to use the same volume, and cause corruption or loss of
data.

Typically, a software application that uses a 3494 library uses volumes in one or
more categories that are reserved for that application. To avoid loss of data, each
application sharing the library must have unique categories. When you define a
3494 library to the server, you can use the PRIVATECATEGORY and
SCRATCHCATEGORY parameters to specify the category numbers for private and
scratch Tivoli Storage Manager volumes in that library. If the volumes are IBM
3592 WORM (write once, read many) volumes, you can use the
WORMSCRATCHCATEGORY parameter to specify category numbers for scratch
WORM volumes in the library. See “Tivoli Storage Manager volumes” on page 46
for more information on private, scratch, and scratch WORM volumes.

When a volume is first inserted into the library, either manually or automatically at
the convenience I/O station, the volume is assigned to the insert category
(X'FF00'). A software application such as Tivoli Storage Manager can contact the
library manager to change a volume's category number. For Tivoli Storage
Manager, you use the CHECKIN LIBVOLUME command (see “Checking media
into automated library devices” on page 130).

The Tivoli Storage Manager server only supports 3590 and 3592 tape drives in an
IBM 3494 library. The server reserves two different categories for each 3494 library
object. The categories are private and scratch.

When you define a 3494 library, you can specify the category numbers for volumes
that the server owns in that library by using the PRIVATECATEGORY,

96 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 SCRATCHCATEGORY, and if the volumes are IBM 3592 WORM volumes, the
WORMSCRATCHCATEGORY parameters. For example:
define library my3494 libtype=349x privatecategory=400 scratchcategory=401
wormscratchcategory=402

For this example, the server uses the following categories in the new my3494
library:
v 400 (X'190') Private volumes
v 401 (X'191') Scratch volumes
v 402 (X'192') WORM scratch volumes

Note: The default values for the categories may be acceptable in most cases.
However, if you connect other systems or Tivoli Storage Manager servers to a
single 3494 library, ensure that each uses unique category numbers. Otherwise, two
or more systems may try to use the same volume, and cause a corruption or loss
of data.

For a discussion regarding the interaction between library clients and the library
manager in processing Tivoli Storage Manager operations, see “Shared libraries” on
page 140.

Configuring an IBM 3494 library for use by one server

In the following example, an IBM 3494 library containing two drives is configured
for use by one Tivoli Storage Manager server.

Procedure

You must first set up the IBM 3494 library on the server system. This involves the
following tasks:
1. Set the symbolic name for the library in the configuration file for the library
device driver (c:\winnt\[Link]). This procedure is described in IBM Tape
Device Drivers Installation and User’s Guide.
2. Physically attach the devices to the server hardware or the SAN.
3. Install and configure the appropriate device drivers for the devices on the
server that will use the library and drives.
4. Determine the device names that are needed to define the devices to Tivoli
Storage Manager.

Results

For details, see the following topic:

v “Selecting a device driver” on page 78.

There are two possible configurations:

v In the first configuration, both drives in the library are the same device type. See
“Configuring a 3494 library with a single drive device type” on page 98.
v In the second configuration, the drives are different device types.
Drives with different device types (or different generations of drives) are
supported in a single physical library if you define one library to Tivoli Storage
Manager for each type of drive (or generation of drive). For example, if you
have two device types, such as 3590E and 3590H (or two generations of drives
of the same device type), define two libraries. Then, define drives and device
classes for each library. In each device class definition, you can use the FORMAT

Chapter 6. Configuring storage devices 97

 parameter with a value of DRIVE, if you choose. See “Configuring a 3494 library
with multiple drive device types” on page 99.

Configuring a 3494 library with a single drive device type

In this example, the 3494 library contains two IBM 3590 tape drives.

Procedure
1. Define a 3494 library named 3494LIB:
define library 3494lib libtype=349x
2. Define a path from the server to the library:
define path server1 3494lib srctype=server desttype=library
device=library1
See “Defining libraries” on page 173 and “SCSI libraries” on page 38.
For more information about paths, see “Defining paths” on page 176.
3. Define the drives in the library:
define drive 3494lib drive01
define drive 3494lib drive02
Both drives belong to the 3494LIB library.
See “Defining drives” on page 174.
4. Define a path from the server to each drive:
define path server1 drive01 srctype=server desttype=drive
library=3494lib device=mt1.0.0.0
define path server1 drive02 srctype=server desttype=drive
library=3494lib device=mt2.0.0.0
The DEVICE parameter gives the device alias name for the drive. For more
about device names, see “Device alias names” on page 75.
For more information about paths, see “Defining paths” on page 176.
5. Classify drives according to type by defining Tivoli Storage Manager device
classes. For example, for the two 3590 drives in the 3494LIB library, use the
following command to define a device class named 3494_CLASS:
define devclass 3494_class library=3494lib devtype=3590 format=drive
This example uses FORMAT=DRIVE as the recording format because both
drives associated with the device class use the same recording format; both are
3590 drives. If instead one drive is a 3590 and one is a 3590E, you need to use
specific recording formats when defining the device classes. See “Configuring a
3494 library with multiple drive device types” on page 99.
See also “Defining tape device classes” on page 180.
6. Verify your definitions by issuing the following commands:
query library
query drive
query path
query devclass
For details, see the following topics:
“Requesting information about drives” on page 152
“Obtaining information about device classes” on page 198
“Obtaining information about paths” on page 169
“Obtaining information about libraries” on page 150
7. Define a storage pool named 3494_POOL associated with the device class
named 3494_CLASS.
define stgpool 3494_pool 3494_class maxscratch=20

98 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 Key choices:
a. Scratch volumes are empty volumes that are labeled and available for use.
If you allow scratch volumes for the storage pool by specifying a value for
the maximum number of scratch volumes, the server can choose from the
scratch volumes available in the library, without further action on your part.
If you do not allow scratch volumes, you must perform the extra step of
explicitly defining each volume to be used in the storage pool.
b. The default setting for primary storage pools is collocation by group. The
default for copy storage pools and active-data pools is disablement of
collocation. Collocation is a process by which the server attempts to keep all
files belonging to a group of client nodes, a single client node, or a client
file space on a minimal number of volumes. If collocation is disabled for a
storage pool and clients begin storing data, you cannot easily change the
data in the pool so that it is collocated. To understand the advantages and
disadvantages of collocation, see “Keeping client files together using
collocation” on page 357 and “How collocation affects reclamation” on page
378.
For more information, see “Defining storage pools” on page 245.

Configuring a 3494 library with multiple drive device types

In this example, the 3494 library contains two IBM 3590E tape drives and two IBM
3590H tape drives.

Procedure
1. Define two libraries, one for each type of drive. For example, to define
3590ELIB and 3590HLIB enter the following commands:
define library 3590elib libtype=349x scratchcategory=301 privatecategory=300
define library 3590hlib libtype=349x scratchcategory=401 privatecategory=400
See “Defining libraries” on page 173.

Note: Specify scratch and private categories explicitly. If you accept the
category defaults for both library definitions, different types of media will be
assigned to the same categories.
2. Define a path from the server to each library:
define path server1 3590elib srctype=server desttype=library device=library1
define path server1 3590hlib srctype=server desttype=library device=library1
The DEVICE parameter specifies the symbolic name for the library, as defined
in the configuration file for the library device driver (c:\winnt\[Link]).
For more information about paths, see “Defining paths” on page 176.
3. Define the drives, ensuring that they are associated with the appropriate
libraries.
v Define the 3590E drives to 3590ELIB.
define drive 3590elib 3590e_drive1
define drive 3590elib 3590e_drive2
v Define the 3590H drives to 3590HLIB.
define drive 3590hlib 3590h_drive3
define drive 3590hlib 3590h_drive4

Note: Tivoli Storage Manager does not prevent you from associating a drive
with the wrong library.
See “Defining drives” on page 174.
4. Define a path from the server to each drive. Ensure that you specify the correct
library.

Chapter 6. Configuring storage devices 99

 v For the 3590E drives:
define path server1 3590e_drive1 srctype=server desttype=drive
library=3590elib device=mt1.0.0.0
define path server1 3590e_drive2 srctype=server desttype=drive
library=3590elib device=mt2.0.0.0
v For the 3590H drives:
define path server1 3590h_drive3 srctype=server desttype=drive
library=3590hlib device=mt3.0.0.0
define path server1 3590h_drive4 srctype=server desttype=drive
library=3590hlib device=mt4.0.0.0
The DEVICE parameter gives the device alias name for the drive. For more
about device names, see “Device alias names” on page 75.
For more information about paths, see “Defining paths” on page 176.
5. Classify the drives according to type by defining Tivoli Storage Manager device
classes, which specify the recording formats of the drives. Because there are
separate libraries, you can enter a specific recording format, for example 3590H,
or you can enter DRIVE.
define devclass 3590e_class library=3590elib devtype=3590 format=3590e

define devclass 3590h_class library=3590hlib devtype=3590 format=3590h

See “Defining tape device classes” on page 180.
6. To check what you have defined, enter the following commands:
query library
query drive
query path
query devclass
See the following topics:
v “Obtaining information about device classes” on page 198
v “Obtaining information about paths” on page 169
v “Requesting information about drives” on page 152
7. Create the storage pools to use the devices in the device classes you just
defined. For example, define a storage pool named 3590EPOOL associated with
the device class 3490E_CLASS, and 3590HPOOL associated with the device
class 3590H_CLASS:
define stgpool 3590epool 3590e_class maxscratch=20
define stgpool 3590hpool 3590h_class maxscratch=20

Key choices:
a. Scratch volumes are labeled, empty volumes that are available for use. If
you allow scratch volumes for the storage pool by specifying a value for the
maximum number of scratch volumes, the server can choose from the
scratch volumes available in the library, without further action on your part.
If you do not allow scratch volumes, you must perform the extra step of
explicitly defining each volume to be used in the storage pool.
b. The default setting for primary storage pools is collocation by group. The
default for copy storage pools and active-data pools is disablement of
collocation. Collocation is a process by which the server attempts to keep all
files belonging to a group of client nodes, a single client node, or a client
file space on a minimal number of volumes. If collocation is disabled for a
storage pool and clients begin storing data, you cannot easily change the
data in the pool so that it is collocated. To understand the advantages and
disadvantages of collocation, see “Keeping client files together using
collocation” on page 357 and “How collocation affects reclamation” on page
378.

100 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 For more information, see “Defining storage pools” on page 245.

Checking in and labeling 3494 library volumes

Ensure that enough volumes in the library are available to the server. Keep enough
labeled volumes on hand so that you do not run out during an operation such as
client backup. Label and set aside extra scratch volumes for any potential recovery
operations you might have later.

About this task

Each volume that is used by a server for any purpose must have a unique name.
This requirement applies to all volumes, whether the volumes are used for storage
pools, or used for operations such as database backup or export. The requirement
also applies to volumes that reside in different libraries.

The procedures for volume check-in and labeling are the same whether the library
contains drives of a single device type, or drives of multiple device types.

Note: If your library has drives of multiple device types, you defined two
libraries to the Tivoli Storage Manager server in the procedure in “Configuring a
3494 library with multiple drive device types” on page 99. The two Tivoli Storage
Manager libraries represent the one physical library. The check-in process finds all
available volumes that are not already checked in. You must check in media
separately to each defined library. Ensure that you check in volumes to the correct
Tivoli Storage Manager library.

Procedure
1. Check in the library inventory. The following shows two examples.
v Check in volumes that are already labeled:
checkin libvolume 3494lib search=yes status=scratch checklabel=no
v Label and check in volumes:
label libvolume 3494lib search=yes checkin=scratch
2. Depending on whether you use scratch volumes or private volumes, do one of
the following:
v If you use only scratch volumes, ensure that enough scratch volumes are
available. For example, you might need to label more volumes. As volumes
are used, you might also need to increase the number of scratch volumes
that are allowed in the storage pool that you defined for this library.
v If you want to use private volumes in addition to or instead of scratch
volumes in the library, define volumes to the storage pool you defined. The
volumes that you define must have been already labeled and checked in. See
“Defining storage pool volumes” on page 256.

Results

For more information about checking in volumes, see “Checking media into
automated library devices” on page 130.

Chapter 6. Configuring storage devices 101

 Sharing an IBM 3494 library among servers
Sharing an IBM 3494 library requires one of the following environments.

About this task

v The library must be on a SAN.
v By using the dual ports on 3590 drives in the library, the drives and the library
are connected to two systems on which Tivoli Storage Manager servers run.

Procedure

The following tasks are required for Tivoli Storage Manager servers to share library
devices over a SAN:
1. Ensure the server that is defined as the library manager is at the same or
higher version as the server or servers that are defined as library clients.
2. Set up server-to-server communications.
3. Set up the device on the server systems.
4. Set up the library on the library manager server. In the following example, the
library manager server is named MANAGER.
5. Set up the library on the library client server. In the following example, the
library client server is named CLIENT.

Results

For more information about configuring 3494 libraries, see “Categories in an IBM
3494 library” on page 96.

Setting up a 3494 library on the server system and SAN

You must first set up the device on the server system, which involves certain tasks.

Procedure
1. Set the symbolic name for the library in the configuration file for the library
device driver. This procedure is described in the IBM Tape Device Drivers
Installation and User’s Guide.
2. Physically attach the devices to the SAN or to the server hardware.
3. On each server system that will access the library and drives, install and
configure the appropriate device drivers for the devices.
4. Determine the device names that are needed to define the devices to Tivoli
Storage Manager.

Results

For details, see “Selecting a device driver” on page 78

Note: You can also configure a 3494 library so that it contains drives of multiple
device types or different generations of drives of the same device type. The
procedure for working with multiple drive device types is similar to the one
described for a LAN in “Configuring a 3494 library with multiple drive device
types” on page 99.

For details about mixing generations of drives, see “Defining 3592 device classes”
on page 183 and “Defining LTO device classes” on page 191.

102 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Setting up the 3494 library manager server
Use the following procedure as an example of how to set up a Tivoli Storage
Manager server as a library manager named MANAGER.

Procedure
1. Define a 3494 library named 3494SAN:
define library 3494san libtype=349x shared=yes
2. Define a path from the server to the library:
define path manager 3494san srctype=server desttype=library
device=library1
The DEVICE parameter specifies the symbolic name for the library, as defined
in the configuration file for the library device driver (c:\winnt\[Link]).
For more information about paths, see “Defining paths” on page 176.
3. Define the drives in the library:
define drive 3494san drivea
define drive 3494san driveb
4. Define a path from the server to each drive:
define path manager drivea srctype=server desttype=drive library=3494san
device=mt4.0.0.0
define path manager driveb srctype=server desttype=drive library=3494san
device=mt5.0.0.0
For more information about paths, see “Defining paths” on page 176.
5. Define all the device classes that are associated with the shared library.
define devclass 3494_class library=3494san devtype=3590
6. Check in the library inventory. The following shows two examples. In both
cases, the server uses the name on the barcode label as the volume name.
To check in volumes that are already labeled, use the following command:
checkin libvolume 3494san search=yes status=scratch checklabel=no

To label and check in the volumes, use the following command:

label libvolume 3494san checkin=scratch search=yes
7. Set any required storage pools for the shared library with a maximum of 50
scratch volumes.
define stgpool 3494_sanpool tape maxscratch=50

Setting up the 3494 library client servers

Use the following sample procedure for each Tivoli Storage Manager server that
will be a library client server.

Procedure
1. Define the server that is the library manager:
define server manager serverpassword=secret hladdress=[Link] lladdress=1580
crossdefine=yes
2. Define a shared library named 3494SAN, and identify the library manager:

Note: Ensure that the library name agrees with the library name on the library
manager.
define library 3494san libtype=shared primarylibmanager=manager
3. Perform this step from the library manager. Define a path from the library client
server to each drive that the library client server will be allowed to access. The
device name should reflect the way the library client system sees the device.

Chapter 6. Configuring storage devices 103

 There must be a path defined from the library manager to each drive in order
for the library client to use the drive. The following is an example of how to
define a path:
define path client drivea srctype=server desttype=drive
library=3494san device=mt3.0.0.0
define path client driveb srctype=server desttype=drive
library=3494san device=mt4.0.0.0
For more information about paths, see “Defining paths” on page 176.
4. Return to the library client for the remaining steps. Define all the device classes
that are associated with the shared library.
define devclass 3494_class library=3494san devtype=3590
Set the parameters for the device class the same on the library client as on the
library manager. Making the device class names the same on both servers is
also a good practice, but is not required.
The device class parameters specified on the library manager server override
those specified for the library client. This is true whether or not the device class
names are the same on both servers. If the device class names are different, the
library manager uses the parameters specified in a device class that matches the
device type specified for the library client.

Note: If a library client requires a setting that is different from what is

specified in the library manager's device class (for example, a different mount
limit), do the following:
a. Create an additional device class on the library manager server. Specify the
parameter settings you want the library client to use.
b. Create a device class on the library client with the same name and device
type as the new device class you created on the library server.
5. Define the storage pool, BACKTAPE, that will use the shared library.
define stgpool backtape 3494_class maxscratch=50
6. Repeat this procedure to define additional servers as library clients. For a
discussion regarding the interaction between library clients and the library
manager in processing Tivoli Storage Manager operations, see “Shared
libraries” on page 140

Migrating a shared IBM 3494 library to a library manager

If you have been sharing an IBM 3494 library among Tivoli Storage Manager
servers by using the 3494SHARED option in the [Link] file, you can migrate
to sharing the library by using a library manager and library clients.

About this task

To help ensure a smoother migration and to ensure that all tape volumes that are
being used by the servers get associated with the correct servers, perform the
following migration procedure.

Procedure
1. Do the following on each server that is sharing the 3494 library:
a. Update the storage pools using the UPDATE STGPOOL command. Set the
value for the HIGHMIG and LOWMIG parameters to 100%.
b. Stop the server by issuing the HALT command.
c. Edit the [Link] file and make the following changes:
1) Comment out the 3494SHARED YES option line

104 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 2) Activate the DISABLESCHEDS YES option line if it is not active
3) Activate the EXPINTERVAL X option line if it is not active and change
its value to 0, as follows:
EXPINTERVAL 0
d. Start the server.
e. Enter the following Tivoli Storage Manager command:
disable sessions
2. Set up the library manager on the Tivoli Storage Manager server of your choice
(see “Setting up server communications” on page 119 and “Setting up the
library manager server” on page 119).
3. Do the following on the remaining servers (the library clients):
a. Save the volume history file.
b. Check out all the volumes in the library inventory. Use the CHECKOUT
LIBVOLUME command with REMOVE=NO.
c. Follow the library client setup procedure (“Setting up the 3494 library client
servers” on page 103).
4. Do the following on the library manager server:
a. Check in each library client's volumes. Use the CHECKIN LIBVOLUME
command with the following parameter settings:
v STATUS=PRIVATE
v OWNER=<library client name>

Note: You can use the saved volume history files from the library clients
as a guide.
b. Check in any remaining volumes as scratch volumes. Use the CHECKIN
LIBVOLUME command with STATUS=SCRATCH.
5. Halt all the servers.
6. Edit the [Link] file and comment out the following lines in the file:
DISABLESCHEDS YES
EXPINTERVAL 0
7. Start the servers.

Sharing an IBM 3494 library by static partitioning of drives

If your IBM 3494 library is not on a SAN, you can use partitioning to share that
library among Tivoli Storage Manager servers.

About this task

Tivoli Storage Manager uses the capability of the 3494 library manager, which
allows you to partition a library between multiple Tivoli Storage Manager servers.
Library partitioning differs from library sharing on a SAN in that with
partitioning, there are no Tivoli Storage Manager library managers or library
clients.

When you partition a library on a LAN, each server has its own access to the same
library. For each server, you define a library with tape volume categories unique to
that server. Each drive that resides in the library is defined to only one server. Each
server can then access only those drives it has been assigned. As a result, library
partitioning does not allow dynamic sharing of drives or tape volumes because
they are pre-assigned to different servers using different names and category
codes.

Chapter 6. Configuring storage devices 105

 In the following example, an IBM 3494 library containing four drives is attached to
a Tivoli Storage Manager server named ASTRO and to another Tivoli Storage
Manager server named JUDY.

Note: Tivoli Storage Manager can also share the drives in a 3494 library with other
servers by enabling the 3494SHARED server option. When this option is enabled,
you can define all of the drives in a 3494 library to multiple servers, if there are
SCSI connections from all drives to the systems on which the servers are running.
This type of configuration is not recommended, however, because when this type
of sharing takes place there is a risk of contention between servers for drive usage,
and operations can fail.

Setting up the 3494 library on the server system

You must first set up the 3494 library on the server system.

Procedure

Setting up the 3494 library involves the following tasks:

1. Set the symbolic name for the library in the configuration file for the library
device driver. This procedure is described in IBM Tape Device Drivers Installation
and User’s Guide.
2. Physically attach the devices to the server hardware.
3. On each server system that accesses the library and drives, install and
configure the appropriate device drivers for the devices.
4. Determine the device names that are needed to define the devices to Tivoli
Storage Manager.

Results

For details, see “Selecting a device driver” on page 78.

Defining 3494 library devices to the Tivoli Storage Manager

server ASTRO
Complete the following steps to define the 3493 library.

Procedure
1. Define the 3494 library named 3494LIB:
define library 3494lib libtype=349x privatecategory=400 scratchcategory=600

The PRIVATECATEGORY and SCRATCHCATEGORY are set differently from

the default settings. See “Categories in an IBM 3494 library” on page 96.
2. Define the path from the server, ASTRO, to the library:
define path astro 3494lib srctype=server desttype=library
device=library1
The DEVICE parameter specifies the symbolic name for the library, as defined
in the configuration file for the library device driver (c:\winnt\[Link]).
See “Defining libraries” on page 173 and “SCSI libraries” on page 38.
For more information about paths, see “Defining paths” on page 176.
3. Define the drives that are partitioned to server ASTRO:
define drive 3494lib drive1
define drive 3494lib drive2
4. Define the path from the server, ASTRO, to each of the drives:

106 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 define path astro drive1 srctype=server desttype=drive library=3494lib
device=mt1.0.0.0
define path astro drive2 srctype=server desttype=drive library=3494lib
device=mt2.0.0.0
The DEVICE parameter gives the device alias name for the drive. For more
about device names, see “Device alias names” on page 75.
5. Classify drives according to type by defining Tivoli Storage Manager device
classes. For example, to classify the two drives in the 3494LIB library, use the
following command to define a device class named 3494_CLASS:
define devclass 3494_class library=3494lib devtype=3590 format=drive
This example uses FORMAT=DRIVE as the recording format because both
drives associated with the device class use the same recording format; both are
3590 drives. If instead one drive is a 3590 and one is a 3590E, you need to use
specific recording formats when defining the device classes. See “Configuring a
3494 library with multiple drive device types” on page 99.
See “Defining tape device classes” on page 180.
6. Verify your definitions by issuing the following commands:
query library
query drive
query path
query devclass
See the following topics:
v “Obtaining information about device classes” on page 198
v “Obtaining information about paths” on page 169
v “Requesting information about drives” on page 152
7. Define a storage pool named 3494_POOL associated with the device class
named 3494_CLASS:
define stgpool 3494_pool 3494_class maxscratch=20

Key choices:
a. Scratch volumes are empty volumes that are labeled and available for use.
If you allow scratch volumes for the storage pool by specifying a value for
the maximum number of scratch volumes, the server can choose from the
scratch volumes available in the library, without further action on your part.
If you do not allow scratch volumes, you must perform the extra step of
explicitly defining each volume to be used in the storage pool.
b. The default setting for primary storage pools is collocation by group. The
default for copy storage pools and active-data pools is disablement of
collocation. Collocation is a process by which the server attempts to keep all
files belonging to a group of client nodes, a single client node, or a client
file space on a minimal number of volumes. If collocation is disabled for a
storage pool and clients begin storing data, you cannot easily change the
data in the pool so that it is collocated. To understand the advantages and
disadvantages of collocation, see “Keeping client files together using
collocation” on page 357 and “How collocation affects reclamation” on page
378.
For more information, see “Defining storage pools” on page 245.

Chapter 6. Configuring storage devices 107

 Defining 3494 library devices to the Tivoli Storage Manager
server JUDY
The DEVICE parameter specifies the device special file for the library management
control points.

Procedure
1. Define the 3494 library named 3494LIB:
define library 3494lib libtype=3494 privatecategory=112 scratchcategory=300

The PRIVATECATEGORY and SCRATCHCATEGORY are defined differently

than the first server's definition. See “Categories in an IBM 3494 library” on
page 96.
2. Define the path from the server, JUDY, to the library:
define path judy 3494lib srctype=server desttype=library
device=library1
The DEVICE parameter specifies the symbolic name for the library, as defined
in the configuration file for the library device driver (c:\winnt\[Link]).
See “Defining libraries” on page 173 and “SCSI libraries” on page 38..
For more information about paths, see “Defining paths” on page 176
3. Define the drives that are partitioned to server JUDY:
define drive 3494lib drive3
define drive 3494lib drive4
4. Define the path from the server, JUDY, to each of the drives:
define path judy drive3 srctype=server desttype=drive library=3494lib
device=mt3.0.0.0
define path judy drive4 srctype=server desttype=drive library=3494lib
device=mt4.0.0.0
For more information about paths, see “Defining paths” on page 176.
5. Classify drives according to type by defining Tivoli Storage Manager device
classes. For example, to classify the two drives in the 3494LIB library, use the
following command to define a device class named 3494_CLASS:
define devclass 3494_class library=3494lib devtype=3590 format=drive
This example uses FORMAT=DRIVE as the recording format because both
drives associated with the device class use the same recording format; both are
3590 drives. If instead one drive is a 3590 and one is a 3590E, you need to use
specific recording formats when defining the device classes. See “Configuring a
3494 library with multiple drive device types” on page 99.
See “Defining tape device classes” on page 180.
6. Verify your definitions by issuing the following commands:
query library
query drive
query path
query devclass
See the following topics:
v “Obtaining information about device classes” on page 198
v “Requesting information about drives” on page 152
7. Define a storage pool named 3494_POOL associated with the device class
named 3494_CLASS.
define stgpool 3494_pool 3494_class maxscratch=20

Key choices:

108 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 a. Scratch volumes are empty volumes that are labeled and available for use.
If you allow scratch volumes for the storage pool by specifying a value for
the maximum number of scratch volumes, the server can choose from the
scratch volumes available in the library, without further action on your part.
If you do not allow scratch volumes, you must perform the extra step of
explicitly defining each volume to be used in the storage pool.
b. The default setting for primary storage pools is collocation by group. The
default for copy storage pools and active-data pools is disablement of
collocation. Collocation is a process by which the server attempts to keep all
files belonging to a group of client nodes, a single client node, or a client
file space on a minimal number of volumes. If collocation is disabled for a
storage pool and clients begin storing data, you cannot easily change the
data in the pool so that it is collocated. To understand the advantages and
disadvantages of collocation, see “Keeping client files together using
collocation” on page 357 and “How collocation affects reclamation” on page
378.
For more information, see “Defining storage pools” on page 245.

ACSLS-managed libraries
Tivoli Storage Manager supports tape libraries controlled by StorageTek Automated
Cartridge System Library Software (ACSLS). The ACSLS library server manages
the physical aspects of tape cartridge storage and retrieval.

The ACSLS client application communicates with the ACSLS library server to
access tape cartridges in an automated library. Tivoli Storage Manager is one of the
applications that gains access to tape cartridges by interacting with ACSLS through
its client, which is known as the control path. The Tivoli Storage Manager server
reads and writes data on tape cartridges by interacting directly with tape drives
through the data path. The control path and the data path are two different paths.

In order to utilize ACSLS functions, StorageTek Library Attach software must be

installed.

The ACSLS client daemon must be initialized before starting the server using
StorageTek Library Attach. For detailed installation, configuration, and system
administration of ACSLS, refer to the appropriate StorageTek documentation.

Configuring an ACSLS-managed library

The library ACSLS is attached to the ACSLS server, and the drives are attached to
the Tivoli Storage Manager server. The ACSLS server and the Tivoli Storage
Manager server must be on different systems. Refer to the ACSLS installation
documentation for details about how to set up the library.

About this task

There are two configurations described in this section:

v In the first configuration, both drives in the ACSLS library are the same device
type. See “Configuring an ACSLS library with a single drive device type” on
page 110.
v In the second configuration, the drives are different device types.
Drives with different device types (or different generations of drives) are
supported in a single physical library if you define one library to Tivoli Storage
Manager for each type of drive (or generation of drive). If you have two device

Chapter 6. Configuring storage devices 109

 types, such as 9840 and 9940 (or two generations of drives of the same device
type), define two libraries. Then define drives and device classes for each library.
In each device class definition, you can use the FORMAT parameter with a value
of DRIVE, if you choose. See “Configuring an ACSLS library with multiple drive
device type” on page 111.

Configuring an ACSLS library with a single drive device type

The parameter ACSID specifies the number that the Automatic Cartridge System
System Administrator (ACSSA) assigned to the library. Issue the QUERY ACS
command to your ACSLS system to determine the number for your library ID.

Procedure
1. Define an ACSLS library named ACSLIB:
define library acslib libtype=acsls acsid=1
2. Define the drives in the library:
define drive acslib drive01 acsdrvid=1,2,3,4
define drive acslib drive02 acsdrvid=1,2,3,5
The ACSDRVID parameter specifies the ID of the drive that is being accessed.
The drive ID is a set of numbers that indicate the physical location of a drive
within an ACSLS library. This drive ID must be specified as a, l, p, d, where a is
the ACSID, l is the LSM (library storage module), p is the panel number, and d
is the drive ID. The server needs the drive ID to connect the physical location
of the drive to the drive's SCSI address. See the StorageTek documentation for
details.
See “Defining drives” on page 174.
3. Define a path from the server to each drive:
define path server1 drive01 srctype=server desttype=drive
library=acslib device=mt1.0.0.0
define path server1 drive02 srctype=server desttype=drive
library=acslib device=mt2.0.0.0
The DEVICE parameter gives the device alias name for the drive. For more
about device names, see “Device alias names” on page 75.
For more information about paths, see “Defining paths” on page 176.
4. Classify drives according to type by defining Tivoli Storage Manager device
classes. For example, to classify the two drives in the ACSLIB library, issue the
following command to define a device class named ACS_CLASS:
define devclass acs_class library=acslib devtype=ecartridge format=drive
This example uses FORMAT=DRIVE as the recording format because both
drives associated with the device class use the same recording format; for
example, both are 9940 drives. If instead one drive is a 9840 and one is a 9940,
you must use specific recording formats when defining the device classes. See
“Configuring an ACSLS library with multiple drive device type” on page 111.
See “Defining tape device classes” on page 180.
5. To check what you have defined, issue the following commands:
query library
query drive
query path
query devclass
See the following topics:
v “Obtaining information about device classes” on page 198
v “Obtaining information about paths” on page 169
v “Requesting information about drives” on page 152

110 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 6. Create the storage pool to use the devices in the device class you just defined.
For example, define a storage pool named ACS_POOL associated with the
device class ACS_CLASS:
define stgpool acs_pool acs_class maxscratch=20

Key choices:
a. Scratch volumes are labeled, empty volumes that are available for use. If
you allow scratch volumes for the storage pool by specifying a value for the
maximum number of scratch volumes, the server can choose from the
scratch volumes available in the library, without further action on your part.
If you do not allow scratch volumes, you must perform the extra step of
explicitly defining each volume to be used in the storage pool.
b. The default setting for primary storage pools is collocation by group. The
default for copy storage pools and active-data pools is disablement of
collocation. Collocation is a process by which the server attempts to keep all
files belonging to a group of client nodes, a single client node, or a client
file space on a minimal number of volumes. If collocation is disabled for a
storage pool and clients begin storing data, you cannot easily change the
data in the pool so that it is collocated. To understand the advantages and
disadvantages of collocation, see “Keeping client files together using
collocation” on page 357 and “How collocation affects reclamation” on page
378.
For more information, see “Defining storage pools” on page 245.

Results

Configuring an ACSLS library with multiple drive device type

The following example shows how to set up and ACSLS library with a mix of two
9840 drives and two 9940 drives.

Procedure
1. Define two ACSLS libraries that use the same ACSID. For example to define
9840LIB and 9940LIB, enter the following commands:
define library 9840lib libtype=acsls acsid=1
define library 9940lib libtype=acsls acsid=1
The ACSID parameter specifies the number that the Automatic Cartridge
System System Administrator (ACSSA) assigned to the libraries. Issue the
QUERY ACS command to your ACSLS system to determine the number for
your library ID.
2. Define the drives, ensuring that they are associated with the appropriate
libraries.

Note: Tivoli Storage Manager does not prevent you from associating a drive
with the wrong library.
v Define the 9840 drives to 9840LIB.
define drive 9840lib 9840_drive1 acsdrvid=1,2,3,1
define drive 9840lib 9840_drive2 acsdrvid=1,2,3,2
v Define the 9940 drives to 9940LIB.
define drive 9940lib 9940_drive3 acsdrvid=1,2,3,3
define drive 9940lib 9940_drive4 acsdrvid=1,2,3,4
The ACSDRVID parameter specifies the ID of the drive that is being accessed.
The drive ID is a set of numbers that indicate the physical location of a drive
within an ACSLS library. This drive ID must be specified as a, l, p, d, where a is

Chapter 6. Configuring storage devices 111

 the ACSID, l is the LSM (library storage module), p is the panel number, and d
is the drive ID. The server needs the drive ID to connect the physical location
of the drive to the drive's SCSI address. See the StorageTek documentation for
details.
See “Defining drives” on page 174.
3. Define a path from the server to each drive. Ensure that you specify the correct
library.
v For the 9840 drives:
define path server1 9840_drive1 srctype=server desttype=drive
library=9840lib device=mt1.0.0.0

define path server1 9840_drive2 srctype=server desttype=drive

library=9840lib device=mt2.0.0.0
v For the 9940 drives:
define path server1 9940_drive3 srctype=server desttype=drive
library=9940lib device=mt3.0.0.0

define path server1 9940_drive4 srctype=server desttype=drive

library=9940lib device=mt4.0.0.0
The DEVICE parameter gives the device alias name for the drive. For more
about device names, see “Device alias names” on page 75.
For more information about paths, see “Defining paths” on page 176.
4. Classify the drives according to type by defining Tivoli Storage Manager device
classes, which specify the recording formats of the drives. Because there are
separate libraries, you can enter a specific recording format, for example 9840,
or you can enter DRIVE. For example, to classify the drives in the two libraries,
use the following commands to define one device class for each type of drive:
define devclass 9840_class library=9840lib devtype=ecartridge format=9840

define devclass 9940_class library=9940lib devtype=ecartridge format=9940

See “Defining tape device classes” on page 180.
5. To check what you have defined, enter the following commands:
query library
query drive
query path
query devclass
See the following topics:
v “Obtaining information about device classes” on page 198
v “Obtaining information about paths” on page 169
v “Requesting information about drives” on page 152
6. Create the storage pools to use the devices in the device classes that you just
defined. For example, define storage pools named 9840_POOL associated with
the device class 9840_CLASS and 9940_POOL associated with the device class
9940_CLASS:
define stgpool 9840_pool 9840_class maxscratch=20

define stgpool 9940_pool 9940_class maxscratch=20

Key choices:
a. Scratch volumes are labeled, empty volumes that are available for use. If
you allow scratch volumes for the storage pool by specifying a value for the
maximum number of scratch volumes, the server can choose from the
scratch volumes available in the library, without further action on your part.

112 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 If you do not allow scratch volumes, you must perform the extra step of
explicitly defining each volume to be used in the storage pool.
b. The default setting for primary storage pools is collocation by group. The
default for copy storage pools and active-data pools is disablement of
collocation. Collocation is a process by which the server attempts to keep all
files belonging to a group of client nodes, a single client node, or a client
file space on a minimal number of volumes. If collocation is disabled for a
storage pool and clients begin storing data, you cannot easily change the
data in the pool so that it is collocated. To understand the advantages and
disadvantages of collocation, see “Keeping client files together using
collocation” on page 357 and “How collocation affects reclamation” on page
378.
For more information, see “Defining storage pools” on page 245.

Setting up an ACSLS library manager server

Use the following procedure as an example of how to set up a Tivoli Storage
Manager server as a library manager named GLENCOE:

About this task

When upgrading multiple servers that participate in library sharing, upgrade all
the servers at once, or do the library manager servers and then the library client
servers. Library manager servers are compatible with earlier library clients.
However, library clients are not compatible with earlier library manager servers.

Note: An exception to this rule is when a fix or product enhancement requires

concurrent code changes to the server, storage agent, and library client.

Procedure
1. Verify that the server that is the library manager is running. Start it if it is not.
a. Start the Windows Services Management Console ([Link]).
b. Select the service. For example, TSM Server1.
c. Right-click and select Start.
2. Obtain the library and drive information for the shared library device:
a. Run the [Link] utility. The utility is in the \Program
Files\Tivoli\TSM\server directory.
3. Define a library whose library type is ACSLS. For example:
define library macgregor libtype=acsls shared=yes
4. Define the path from the server to the library:
define path glencoe macgregor srctype=server desttype=library
device=lb0.0.0.2
5. Define the drives in the library.
define drive macgregor drivea acsdrvid=1,0,1,0
define drive macgregor driveb acsdrvid=1,0,1,1
This example uses the acsdrvid value, which specifies the ID of the drive that
is being accessed in an ACSLS library. The drive ID is a set of numbers that
indicates the physical location of a drive within an ACSLS library. This drive ID
must be specified as a,l,p,d, where a is the ACSID, l is the LSM (library storage
module), p is the panel number, and d is the drive ID. The server needs the
drive ID to connect the physical location of the drive to the drive's SCSI
address. See the StorageTek documentation for details.
6. Define the path from the server to each of the drives.

Chapter 6. Configuring storage devices 113

 define path glencoe drivea srctype=server desttype=drive library=macgregor
device=mt0.1.0.2
define path glencoe driveb srctype=server desttype=drive library=macgregor
device=mt0.2.0.2
7. Define at least one device class.
define devclass tape devtype=dlt library=macgregor
8. Check in the library inventory. The following example checks all volumes into
the library inventory as scratch volumes. The server uses the name on the bar
code label as the volume name.
checkin libvolume macgregor search=yes status=scratch
checklabel=barcode
9. Set up a storage pool for the shared library with a maximum of 50 scratch
volumes.
define stgpool backtape tape
description=’storage pool for shared macgregor’ maxscratch=50

Setting up an ACSLS library client server

Use the following procedure as an example of how to set up a Tivoli Storage
Manager server named WALLACE as a library client.

About this task

You must define the library manager server before you set up the library client
server.

Procedure
1. Verify that the server that is the library client is running, and start it if it is not:
a. Start the Windows Services Management Console ([Link]).
b. Select the service. For example, TSM Server1.
c. Right-click and select Start.
2. Obtain the library and drive information for the shared library device:
a. Run the [Link] utility. The utility is in the \Program
Files\Tivoli\TSM\server directory.
3. Define the shared library, MACGREGOR, and identify the library manager.
Ensure that the library name is the same as the library name on the library
manager.
define library macgregor libtype=shared primarylibmanager=glencoe
4. Issue the following commands from an admin client: Define the paths from the
library client server to each of the drives.
define path wallace drivea srctype=server desttype=drive library=macgregor
device=mt0.1.0.3
define path wallace driveb srctype=server desttype=drive library=macgregor
device=mt0.2.0.3
5. Return to the library client for the remaining steps: Define at least one device class.
define devclass tape devtype=dlt mountretention=1 mountwait=10
library=macgregor

Set the parameters for the device class the same on the library client as on the
library manager. Making the device class names the same on both servers is
also a good practice, but is not required.
The device class parameters that are specified on the library manager server
override those specified for the library client. This is true whether or not the
device class names are the same on both servers. If the device class names are

114 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 different, the library manager uses the parameters specified in a device class
that matches the device type specified for the library client.

Note: If a library client requires a setting that is different from what is

specified in the library manager's device class (for example, a different mount
limit), do the following:
a. Create an additional device class on the library manager server. Specify the
parameter settings that you want the library client to use.
b. Create a device class on the library client with the same name and device
type as the new device class you created on the library server.
6. Define the storage pool, LOCHNESS, that will use the shared library.
define stgpool lochness tape
description=’storage pool for shared macgregor’ maxscratch=50
7. Update the copy group to set the destination to the storage pool, LOCHNESS
8. Repeat this procedure to define additional servers as library clients.

Checking in and labeling ACSLS library volumes

Ensure that enough volumes are available to the server in the library. You must
label volumes that do not already have a standard label. Keep enough labeled
volumes on hand so that you do not run out during an operation such as client
backup.

About this task

Each volume used by a server for any purpose must have a unique name. This
requirement applies to all volumes, whether the volumes are used for storage
pools, or used for operations such as database backup or export. The requirement
also applies to volumes that reside in different libraries.

Attention: If your library has drives of multiple device types, you defined two
libraries to the Tivoli Storage Manager server in the procedure in “Configuring an
ACSLS library with multiple drive device type” on page 111. The two Tivoli
Storage Manager libraries represent the one physical library. The check-in process
finds all available volumes that are not already checked in. You must check in
media separately to each defined library. Ensure that you check in volumes to the
correct Tivoli Storage Manager library.

Procedure
1. Check in the library inventory. The following shows examples for libraries with
a single drive device type and with multiple drive device types.
v Check in volumes that are already labeled:
checkin libvolume acslib search=yes status=scratch checklabel=no
v Label and check in volumes:
label libvolume acslib search=yes overwrite=no checkin=scratch
2. Depending on whether you use scratch volumes or private volumes, do one of
the following:
v If you use only scratch volumes, ensure that enough scratch volumes are
available. For example, you may need to label more volumes. As volumes are
used, you may also need to increase the number of scratch volumes allowed
in the storage pool that you defined for this library.
v If you want to use private volumes in addition to or instead of scratch
volumes in the library, define volumes to the storage pool you defined. The

Chapter 6. Configuring storage devices 115

 volumes you define must have been already labeled and checked in. See
“Defining storage pool volumes” on page 256.

Results

For more information about checking in volumes, see:

v “Checking media into automated library devices” on page 130
v “Checking media into automated library devices” on page 130

Managing virtual tape libraries

A virtual tape library (VTL) does not use physical tape media. When using VTL
storage, you can exceed the capabilities of a physical tape library. The ability to
define many volumes and drives can provide greater flexibility for the storage
environment.

Considerations for using virtual tape libraries

There are some considerations for defining a library as a virtual tape library (VTL),
including enhancements for performance and setup of your hardware.

About this task

Defining a VTL to the Tivoli Storage Manager server can help improve
performance because the server handles mount point processing for VTLs
differently than real tape libraries. The physical limitations for real tape hardware
are not applicable to a VTL, affording options for better scalability.

You can use a VTL for any virtual tape library when the following conditions are
true:
v There is no mixed media involved in the VTL. Only one type and generation of
drive and media is emulated in the library.
v Every server and storage agent with access to the VTL has paths that are defined
for all drives in the library.
If either of these conditions are not met, any mount performance advantage from
defining a VTL library to the Tivoli Storage Manager server can be reduced or
negated.

VTLs are compatible with earlier versions of both library clients and storage
agents. The library client or storage agent is not affected by the type of library that
is used for storage. If mixed media and path conditions are true for a SCSI library,
it can be defined or updated as LIBTYPE=VTL.

Storage capacity for virtual tape libraries

Because virtual tape libraries (VTLs) do not have the physical limitations that real
tape hardware does, their capacity for storage is more flexible.

The concept of storage capacity in a virtual tape library is different from capacity
in physical tape hardware. In a physical tape library, each volume has a defined
capacity, and the library's capacity is defined in terms of the total number of
volumes in the library. The capacity of a VTL, alternatively, is defined in terms of
total available disk space. You can increase or decrease the number and size of
volumes on disk.

116 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 This variability affects what it means to run out of space in a VTL. For example, a
volume in a VTL can run out of space before reaching its assigned capacity if the
total underlying disk runs out of space. In this situation, the server can receive an
end-of-volume message without any warning, resulting in backup failures.

When out-of-space errors and backup failures occur, disk space is usually still
available in the VTL. It is hidden in volumes that are not in use. For example,
volumes that are logically deleted or returned to scratch status in the Tivoli Storage
Manager server are only deleted in the server database. The VTL is not notified,
and the VTL maintains the full size of the volume as allocated in its capacity
considerations.

To help prevent out-of-space errors, ensure that any SCSI library that you update
to LIBTYPE=VTL is updated with the RELABELSCRATCH parameter set to YES. The
RELABELSCRATCH option enables the server to overwrite the label for any volume
that is deleted and to return the volume to scratch status in the library. The
RELABELSCRATCH parameter defaults to YES for any library defined as a VTL.

Drive configuration for virtual tape libraries

Drive configuration in a virtual tape library (VTL) is variable, depending on the
needs of your environment.

Most VTL environments use as many drives as possible to maximize the number
of concurrent tape operations. A single tape mount in a VTL environment is
typically faster than a physical tape mount. However, using many drives increases
the amount of time that the Tivoli Storage Manager server requires when a mount
is requested. The selection process takes longer as the number of drives that are
defined in a single library object in the server increases. Virtual tape mounts can
take as long or longer than physical tape mounts depending on the number of
drives in the VTL.

For best results when you create drives, check with your VTL vendor about
device-specific recommendations. If more than 300-500 drives for each VTL are
required, you can logically partition the VTL into multiple libraries and assign
drives to each library. Operating system and SAN hardware configurations could
impose limitations on the number of devices that can be utilized within the VTL
library.

Adding a virtual tape library to your environment

Define a virtual tape library (VTL) to take advantage of mount performance and
scalability advantages.

About this task

VTLs are identified by using the DEFINE LIBRARY command and specifying
LIBTYPE=VTL. Because a VTL library functionally interacts with the server in the
same way that a SCSI library does, it is possible to use the UPDATE LIBRARY
command to change the library type of a SCSI library that is already defined. You
do not have to redefine the library.

Example

The following examples show how to add a VTL library to your environment.

Add a new VTL library

Chapter 6. Configuring storage devices 117

 If you have a new VTL library and want to use the VTL enhancements that are
available in Tivoli Storage Manager Version 6.3, define the library as a VTL to the
server:
define library chester libtype=vtl

This sets up the new VTL library and enables the RELABELSCRATCH option to
relabel volumes that have been deleted and returned to scratch status.

Update an existing SCSI library to a VTL

If you have a SCSI library and you want to change it to a VTL, use the UPDATE
LIBRARY command to change the library type:
update library calzone libtype=vtl

You can only issue this command when the library being updated is defined with
LIBTYPE=SCSI.

Revert a real tape library from the VTL library type

If you define a SCSI tape library as a VTL and want to change it back to the SCSI
library type, update the library by issuing the UPDATE LIBRARY command:
update library chester libtype=scsi

Defining or deleting all drives and paths for a single library

Use the PERFORM LIBACTION command to set up a single SCSI or virtual tape library
(VTL) with one step.

About this task

If you are setting up or modifying your hardware environment and must create or
change large numbers of drive definitions, the PERFORM LIBACTION command can
make this task much simpler. You can define a new library and then define all
drives and paths to the drives. Or, if you have an existing library that you want to
delete, you can delete all existing drives and their paths in one step.

The PREVIEW parameter allows you to view the output of commands before they
are processed to verify the action that you want to perform. If you are defining a
library, a path to the library must already be defined if you want to specify the
PREVIEW parameter. You cannot use the PREVIEW and DEVICE parameters
together.

The PERFORM LIBACTION command can only be used for SCSI and VTL libraries. If
you are defining drives and paths for a library, the SANDISCOVERY option must be
supported and enabled. The tape library must be able to return the drive serial
number address association.

Procedure

To set up a VTL library named ODIN, complete these steps:

1. Define the library.
define library odin libtype=vtl
2. Define two drives and their paths for your new library, ODIN.
perform libaction odin action=define device=lb0.0.0.2 prefix=dr

118 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 The server then issues the following commands:
define path tsmserver odin srct=server destt=library device=lb0.0.0.2
define drive odin dr0
define path tsmserver dr0 srct=server destt=drive library=odin
device=mt0.1.0.2 define drive odin dr1
define path tsmserver dr1 srct=server destt=drive library=odin
device=mt0.2.0.2

Configuring Tivoli Storage Manager servers to share SAN-connected

devices
The steps to configure servers to share SAN-connected devices includes setting up:
server communications, the library manager server, and the library client servers.

About this task

The following tasks are required to share tape library devices over a SAN:

Task Required Privilege Class

“Setting up server communications” System or unrestricted storage
“Setting up the library manager server” System or unrestricted storage
“Setting up the library client servers” on System or unrestricted storage
page 121

Setting up server communications

Before Tivoli Storage Manager servers can share a storage device over a SAN, you
must set up server communications. This requires configuring each server as you
would for Enterprise Administration, which means you define the servers to each
other using the cross-define function.

About this task

Set up each server with a unique name.

For details, see “Setting up communications among servers” on page 692.

Setting up the library manager server

You must set up the library manager server in order to configure the Tivoli Storage
Manager servers to share SAN-connected devices.

About this task

Use the following procedure as an example of how to set up a Tivoli Storage

Manager server as a library manager named ASTRO:

Procedure
1. Verify that the server that is the library manager is running. Start it if it is not.
a. Start the Windows Services Management Console ([Link]).
b. Select the service. For example, TSM Server1.
c. Right-click and select Start.
2. Obtain the library and drive information for the shared library device:

Chapter 6. Configuring storage devices 119

 a. Run the [Link] utility. The utility is in the \Program
Files\Tivoli\TSM\server directory.
3. Define a library whose library type is SCSI. For example:
define library sangroup libtype=scsi shared=yes
This example uses the default for the library's serial number, which is to have
the server obtain the serial number from the library itself at the time that the
path is defined. Depending on the capabilities of the library, the server might
not be able to automatically detect the serial number. In this case, the server
does not record a serial number for the device, and is not able to confirm the
identity of the device when you define the path or when the server uses the
device.
4. Define the path from the server to the library.
define path astro sangroup srctype=server desttype=library
device=lb0.0.0.2
If you did not include the serial number when you defined the library, the
server now queries the library to obtain this information. If you did include the
serial number when you defined the library, the server verifies what you
defined and issues a message if there is a mismatch.
5. Define the drives in the library.
define drive sangroup drivea
define drive sangroup driveb
This example uses the default for the drive's serial number, which is to have
the server obtain the serial number from the drive itself at the time that the
path is defined. Depending on the capabilities of the drive, the server might not
be able to automatically detect the serial number. In this case, the server does
not record a serial number for the device, and is not able to confirm the
identity of the device when you define the path or when the server uses the
device.
This example also uses the default for the drive's element address, which is to
have the server obtain the element number from the drive itself at the time that
the path is defined.
The element address is a number that indicates the physical location of a drive
within an automated library. The server needs the element address to connect
the physical location of the drive to the drive's SCSI address. You can have the
server obtain the element number from the drive itself at the time that the path
is defined, or you can specify the element number when you define the drive.
Depending on the capabilities of the library, the server might not be able to
automatically detect the element address. In this case, you must supply the
element address when you define the drive. If you need the element numbers,
check the device worksheet that is filled out in step 6 on page 74. Element
numbers for many libraries are available at [Link]
entry/portal/Overview/Software/Tivoli/Tivoli_Storage_Manager.
6. Define the path from the server to each of the drives.
define path astro drivea srctype=server desttype=drive library=sangroup
device=mt0.1.0.2
define path astro driveb srctype=server desttype=drive library=sangroup
device=mt0.2.0.2
If you did not include the serial number or element address when you defined
the drive, the server now queries the drive or the library to obtain this
information.
7. Define at least one device class.
define devclass tape devtype=dlt library=sangroup

120 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 8. Check in the library inventory. The following example checks all volumes into
the library inventory as scratch volumes. The server uses the name on the bar
code label as the volume name.
checkin libvolume sangroup search=yes status=scratch
checklabel=barcode
9. Set up a storage pool for the shared library with a maximum of 50 scratch
volumes.
define stgpool backtape tape
description=’storage pool for shared sangroup’ maxscratch=50

Setting up the library client servers

You must set up the library client server in order to configure the Tivoli Storage
Manager servers to share SAN-connected devices.

About this task

You must define the library manager server. Use the following procedure as an
example of how to set up a Tivoli Storage Manager server named JUDY as a
library client.

Procedure
1. Verify that the server that is the library client is running. Start the server if it is
not running:
a. Start the Windows Services Management Console ([Link]).
b. Select the service. For example, TSM Server1.
c. Right-click and select Start.
2. Obtain the library and drive information for the shared library device:
a. Run the [Link] utility. The utility is in the \Program
Files\Tivoli\TSM\server directory.
3. Define the shared library, SANGROUP, and identify the library manager.
Ensure that the library name is the same as the library name on the library
manager.
define library sangroup libtype=shared primarylibmanager=astro
4. Issue the following commands from an admin client: Define the paths from the
library client server to each of the drives.
define path judy drivea srctype=server desttype=drive library=sangroup
device=mt0.1.0.3
define path judy driveb srctype=server desttype=drive library=sangroup
device=mt0.2.0.3
5. Return to the library client for the remaining steps: Define at least one device class.
define devclass tape devtype=dlt mountretention=1 mountwait=10
library=sangroup

Set the parameters for the device class the same on the library client as on the
library manager. Making the device class names the same on both servers is
also a good practice, but is not required.
The device class parameters that are specified on the library manager server
override those specified for the library client. This is true whether or not the
device class names are the same on both servers. If the device class names are
different, the library manager uses the parameters specified in a device class
that matches the device type specified for the library client.

Chapter 6. Configuring storage devices 121

 If a library client requires a setting that is different from what is specified in the
library manager's device class (for example, a different mount limit), do the
following steps:
a. Create an additional device class on the library manager server. Specify the
parameter settings that you want the library client to use.
b. Create a device class on the library client with the same name and device
type as the new device class you created on the library server.
6. Define the storage pool, BACKTAPE, that will use the shared library:
define stgpool backtape tape
description=’storage pool for shared sangroup’ maxscratch=50
7. Repeat this procedure to define additional servers as library clients.

Configuring Tivoli Storage Manager for LAN-free data movement

You can configure the Tivoli Storage Manager client and server so that the client,
through a storage agent, can move its data directly to storage on a Storage Area
Networks (SAN). This function, called LAN-free data movement, is provided by
Tivoli Storage Manager for the SAN.

About this task

As part of the configuration, a storage agent is installed on the client system. Tivoli
Storage Manager supports both tape libraries and FILE libraries. This feature
supports SCSI, 349X, and ACSLS tape libraries.

The configuration procedure you follow will depend on the type of environment
you implement; however in all cases you must perform the following steps:

Procedure
1. Install and configure the client.
2. Install and configure the storage agent.
3. Configure the libraries for LAN-free data movement.
4. Define the libraries and associated paths.
5. Define associated devices and their paths.
6. Configure Tivoli Storage Manager policy for LAN-free data movement for the
client. If you are using shared FILE storage, install and configure IBM
TotalStorage SAN File System or IBM General Parallel File System.

Restriction: If a General Parallel File System volume is formatted by an AIX

server, the Windows system uses TCP/IP to transfer data and not the storage
area network.

Results

For more information on configuring Tivoli Storage Manager for LAN-free data
movement see the Storage Agent User's Guide.

To help you tune the use of your LAN and SAN resources, you can control the
path that data transfers take for clients with the capability of LAN-free data
movement. For each client you can select whether data read and write operations
use:
v The LAN path only
v The LAN-free path only

122 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 v Any path
See the REGISTER NODE and UPDATE NODE commands in the Administrator's
Reference.

Validating your LAN-free configuration

After configured your Tivoli Storage Manager client for LAN-free data movement,
you can verify your configuration and server definitions by issuing the VALIDATE
LANFREE command. This command allows you to determine which destinations
for a given node, using a specific storage agent, are capable of LAN-free data
movement.

About this task

The VALIDATE LANFREE command can also be used to determine if there is a
problem with an existing LAN-free configuration. You can evaluate the policy,
storage pool, and path definitions for a given node using a given storage agent to
ensure that an operation is working properly.

To determine if there is a problem with the client node FRED using the storage
agent FRED_STA, issue the following:
validate lanfree fred fred_sta

The output will allow you to see which management class destinations for a given
operation type are not LAN-free capable. It will also report the total number of
LAN-free destinations.

See the VALIDATE LANFREE command in the Administrator's Reference for more
information.

Troubleshooting device configuration

Procedures are available for displaying device information and the event log.
Common sources of device problems are identified. The impact of device and
cabling changes in a SAN environment is a SAN are described.

Displaying device information

You can display information about devices that are connected to the server by
using the Device Information utility.

Procedure

Complete the following steps to display device information:

1. From a command prompt, change directory to the server subdirectory of the
server installation directory. For example, C:\Program Files\Tivoli\TSM\
server.
2. Run the [Link] utility.

Results

The information that is provided by this utility is from the Windows registry. Some
of the information is put into the registry by the Tivoli Storage Manager device
driver. To receive accurate information, ensure that the device driver is running. If
the device driver is not running, the information can be incorrect if device
attachments have changed since the last time the device driver was running.

Chapter 6. Configuring storage devices 123

 Troubleshooting problems with devices
You can troubleshoot errors that occur when you configure or use devices with
Tivoli Storage Manager.

About this task

Use Table 8 to find a solution to the device-related problem.

Table 8. Resolving device problems
Symptom Problem Solution
Conflicts with other Tivoli Storage Manager Set up a storage area
applications. requires a storage area network.
network to share devices.
Attention: Data loss can
occur if multiple Tivoli
Storage Manager servers use
the same device. Define or
use a device with only one
Tivoli Storage Manager
server.
Labeling fails. A device for labeling
volumes cannot be used at
the same time that the server
uses the device for other
processes.
Incorrect or incomplete Register the license for the
license registration. device support that was
purchased. For more
information about licensing,
see “Licensing IBM Tivoli
Storage Manager” on page
595.
Conflicts among device Tivoli Storage Manager Windows device drivers and
drivers. issues messages about I/O drivers that are provided by
errors when you define or other applications can
use a sequential access interfere with the Tivoli
device. Storage Manager device
driver if the Tivoli Storage
Manager driver is not started
first. To check on the order
that device drivers are
started by the system,
complete the following steps:
1. Click Control Panel.
2. Click Devices. Device
drivers and their startup
types are listed.

124 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 Table 8. Resolving device problems (continued)
Symptom Problem Solution
Device-driver conflicts often
result in I/O errors when
you try to define or use a
tape device. Windows device
drivers and drivers that are
provided by other
applications can interfere
with the Tivoli Storage
Manager device driver if it is
not started first. For a
procedure to ensure that the
Tivoli Storage Manager
device driver starts before
the Windows device driver,
see “Controlling devices with
the Tivoli Storage Manager
device driver” on page 82.

Impact of device changes on the SAN

The SAN environment can shift dramatically because of device or cabling changes.
Device IDs assigned by the SAN may be altered due to bus resets or other
environmental changes. This dynamically changing nature of the SAN can cause
the static definitions defined and known to the server (or storage agent) to fail or
become unpredictable.

The server may know a device as id=1 based on the original path specification to
the server and original configuration of the LAN. However, some event in the SAN
(new device added, cabling change) causes the device to be assigned id=2. When
the server tries to access the device with id=1, it will either get a failure or the
wrong target device. The server assists in recovering from changes to devices on
the SAN by using serial numbers to confirm the identity of devices it contacts.

When you define a device (drive or library) you have the option of specifying the
serial number for that device. If you do not specify the serial number when you
define the device, the server obtains the serial number when you define the path
for the device. In either case, the server then has the serial number in its database.
From then on, the server uses the serial number to confirm the identity of a device
for operations.

When the server uses drives and libraries on a SAN, the server attempts to verify
that the device it is using is the correct device. The server contacts the device by
using the device name in the path that you defined for it. The server then requests
the serial number from the device, and compares that serial number with the serial
number stored in the server database for that device.

If the serial numbers do not match, the server begins the process of discovery on
the SAN to attempt to find the device with the matching serial number. If the
server finds the device with the matching serial number, it corrects the definition
of the path in the server's database by updating the device name in that path. The
server issues a message with information about the change made to the device.
Then the server proceeds to use the device.

Chapter 6. Configuring storage devices 125

 You can monitor the activity log for messages if you want to know when device
changes on the SAN have affected Tivoli Storage Manager. The following are the
number ranges for messages related to serial numbers:
v ANR8952 through ANR8958
v ANR8961 through ANR8967

Restriction: Some devices do not have the capability of reporting their serial
numbers to applications such as the Tivoli Storage Manager server. If the server
cannot obtain the serial number from a device, it cannot assist you with changes to
that device's location on the SAN.

Increased block size for writing to tape

Tivoli Storage Manager provides the DSMMAXSG utility that can improve the rate at
which the server processes data for backup and restore operations, and for archive
and retrieve operations.

Actual results depend upon your system environment. The utility does not affect
the generation of backup sets.

The DSMMAXSG utility increases the maximum transfer length for some host bus
adapters (HBAs). The Tivoli Storage Manager server uses the following types of
tape drives:
v 3590
v 3592
v DLT
v ECARTRIDGE
v LTO
The maximum block size that you can use with this utility is 256 KB. When you
run the DSMMAXSG utility, it modifies one registry key for every HBA driver on the
system. The name of the key is MaximumSGList.

Typically, the utility is run automatically as part of the Tivoli Storage Manager
server or storage agent installation. However, you must run the utility manually to
take advantage of the larger block size if you:
v install a new HBA on your system after you install the server or storage agent
v install a new version of an existing HBA device driver that resets the value of
the maximum transfer size

Restriction: If you back up or archive to tape by using the 256 KB block size, you
cannot append to or read from the tape by using an HBA that is not compatible
with the 256 KB block size. For example, if you use a 256 KB Windows system to
back up client data to the Tivoli Storage Manager server, you cannot restore the
data by using a Windows system that uses a different transfer length. If you want
to append to or read from tape that is written to using a 256 KB transfer length,
you must install an HBA that supports 256 KB transfers.

For more information about the DSMMAXSG utility, see the Administrator's Reference

126 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Chapter 7. Managing removable media operations
Routine removable media operations include preparing media for use, controlling
how and when media are reused, and ensuring that sufficient media are available.
You also must respond to operator requests and manage libraries, drives, disks,
paths, and data movers.

About this task

“Preparing media for automated libraries” on page 129
“Managing media in automated libraries” on page 135
“Media management in manual libraries” on page 142
“Tivoli Storage Manager server requests” on page 143
“Tape rotation” on page 146
“Managing paths” on page 169
“Managing libraries” on page 150
“Managing drives” on page 152
“Managing data movers” on page 170
“Managing disks” on page 171

The examples in topics show how to perform tasks using the Tivoli Storage
Manager command-line interface. For information about the commands, see the
Administrator's Reference, or issue the HELP command from the command line of a
Tivoli Storage Manager administrative client.

Defining volumes
For each storage pool, decide whether to use scratch volumes or private volumes.
Private volumes require more human intervention than scratch volumes.

About this task

When you use commands to add devices, you specify the maximum number of
scratch volumes with the MAXSCRATCH parameter of the DEFINE STGPOOL or UPDATE
STGPOOL command. If the MAXSCRATCH parameter is 0, all the volumes in the storage
pool are private volumes that you must define.

For example, to create a storage pool named STORE1 that can use up to 500 scratch
volumes, issue the following command:
define stgpool store1 maxscratch=500

Scratch volumes are recommended for the following reasons:

v You need not explicitly define each storage pool volume.
v Scratch volumes are convenient to manage and they fully use the automation of
robotic devices.
v Different storage pools that share the same automated library can dynamically
acquire volumes from the library's collection of scratch volumes. The volumes
do not need to be preallocated to the different storage pools.

© Copyright IBM Corp. 1993, 2013 127

 Use private volumes to regulate the volumes that are used by individual storage
pools, and to manually control the volumes. Define each private volume with the
DEFINE VOLUME command. For database backups, dumps, or loads, or for server
import or export operations, you must list the private volumes.

Managing volumes
When Tivoli Storage Manager needs a new volume, it chooses a volume from the
storage pool available for client backups. If you set up private volumes, it selects a
specific volume. If you set up scratch volumes, it selects any scratch volume in the
library.

About this task

IBM 3494 Tape Library Dataservers use category numbers to identify volumes that
are used for the same purpose or application. For details, see “Category numbers
for IBM 3494 libraries” on page 141. For special considerations regarding
write-once, read-many (WORM) volumes, see “Write-once, read-many tape media”
on page 133.

Remember: Each volume used by a server for any purpose must have a unique
name. This requirement applies to all volumes, whether the volumes are used for
storage pools, or used for operations such as database backup or export. The
requirement also applies to volumes that reside in different libraries but that are
used by the same server.

Partially-written volumes
Partially-written volumes are always private volumes, even if their status was
scratch before Tivoli Storage Manager selects them to be mounted. Tivoli Storage
Manager tracks the original status of scratch volumes, so it can return them to
scratch status when they become empty.

Except for volumes in automated libraries, Tivoli Storage Manager is unaware of a

scratch volume until after the volume is mounted. Then, the volume status changes
to private, and the volume is automatically defined as part of the storage pool for
which the mount request was made.

For information about changing the status of a volume in an automated library, see
“Changing the status of automated library volumes” on page 135.

Volume inventory for automated libraries

Tivoli Storage Manager maintains a volume inventory for each automated library.
The volume inventory allows the device to provide maximum automation.

The volume inventory is created when you check media volumes into the library.
Tivoli Storage Manager tracks the status of volumes in the inventory as either
scratch or private.

A list of volumes in the library volume inventory will not necessarily be identical
to a list of volumes in the storage pool inventory for the device. For example,
scratch volumes may be checked in to the library but not defined to a storage pool
because they have not yet been selected for backup; private volumes may be
defined to a storage pool, but not checked into the device's volume inventory.

128 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 Changing the status of database-backup and database-export
volumes
When Tivoli Storage Manager backs up the database or exports server information,
it records information about the volumes that are used for these operations in the
volume history file.

About this task

To change the status of database-backup and database-export volumes, use the

DELETE VOLHISTORY command or the UPDATE LIBVOLUME command.

For details about the volume history file, see Chapter 29, “Protecting and
recovering the server infrastructure and client data,” on page 869.

Preparing media for automated libraries

You can prepare tape volumes by labeling them and checking them into the library
volume inventory:

About this task

Task Required Privilege Class
“Labeling media” System
“Checking media into automated library System
devices” on page 130
“Element addresses for library storage slots” Any Administrator or Operator
on page 132

Labeling media
All media require labels. Labeling media with an automated library requires you to
check media into the library. Checkin processing can be done at the same time that
the volume is labeled.

About this task

To label volumes with the LABEL LIBVOLUME command, specify the CHECKIN
parameter.

To automatically label tape volumes in SCSI-type libraries, use the AUTOLABEL

parameter on the DEFINE LIBRARY and UPDATE LIBRARY commands. Using this
parameter eliminates the need to pre-label a set of tapes. It is also more efficient
than using the LABEL LIBVOLUME command, which requires you to mount volumes
separately. If you use the AUTOLABEL parameter, you must check in tapes by
specifying CHECKLABEL=BARCODE on the CHECKIN LIBVOLUME command.

A label cannot include embedded blanks or periods and must be valid when used
as a file name on the media.

Note: You must label CD-ROM, Zip, or Jaz volumes with the device utilities from
the manufacturer or the Windows utilities because Tivoli Storage Manager does not
provide utilities to format or label these media types. The operating system utilities
include the Disk Administrator program (a graphical user interface) and the label
command.

Chapter 7. Managing removable media operations 129

 Checking media into automated library devices
After volumes have been labeled, make the volumes available to Tivoli Storage
Manager devices by checking the volumes into the library volume inventory using
the CHECKIN LIBVOLUME command.

About this task

The CHECKIN LIBVOLUME command involves device access, and may take a long
time to complete. For this reason, the command always executes as a background
process. Wait for the CHECKIN LIBVOLUME process to complete before defining
volumes or the defining process will fail. You can save time by checking in
volumes as part of the labeling operation. For details, see “Labeling media” on
page 129.

You can specify that Tivoli Storage Manager read media labels for the volumes you
are checking in. When label-checking is enabled, Tivoli Storage Manager mounts
each volume and reads the internal label before checking in the volume. Tivoli
Storage Manager checks in only volumes that are properly labeled. Checking labels
can prevent errors later, when Tivoli Storage Manager selects and mounts volumes,
but it also increases check in processing time.

Checking a single volume into an automated library

You can check in single volumes using the CHECKIN LIBVOLUME command with
the SEARCH=NO parameter.

About this task

Tivoli Storage Manager issues a mount request identifying a storage slot with an
element address. The media can be loaded directly into a single storage slot or into
one of the device's entry/exit ports, if it is equipped with them. For example,
check a scratch volume named VOL001 into a library named TAPELIB by entering
the following command:
checkin libvolume tapelib vol001 search=no status=scratch

Tivoli Storage Manager finds that the first empty slot is at element address 5, and
issues the following message:
ANR8306I 001: Insert 8MM volume VOL001 R/W in slot with element
address 5 of library TAPELIB within 60 minutes; issue ’REPLY’ along
with the request ID when ready.

If the library is equipped with entry/exit ports, the administrator can load the
volume into a port without knowing the element addresses of the device's storage
slots. After inserting the volume into an entry/exit port or storage slot, the
administrator responds to the preceding message at a Tivoli Storage Manager
command line by issuing the REPLY command with the request number (the
number at the beginning of the mount request):
reply 1

Tip: A REPLY command is not required if you specify a wait time of zero using the
optional WAITTIME parameter on the CHECKIN LIBVOLUME command. The default wait
time is 60 minutes.

130 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Checking in volumes using library bar code readers
You can save time checking volumes into libraries equipped with bar code readers
by using the characters on the bar code labels as names for the volumes being
checked in.

About this task

Tivoli Storage Manager reads the bar code labels and uses the information on the
labels to write the internal media labels. For volumes missing bar code labels,
Tivoli Storage Manager mounts the volumes in a drive and attempts to read the
internal, recorded label.

For example, to use a bar code reader to search a library named TAPELIB and
check in a scratch tape, enter:
checkin libvolume tapelib search=yes status=scratch
checklabel=barcode

Checking in volumes from library entry/exit ports

To search all slots of bulk entry/exit ports for labeled volumes that Tivoli Storage
Manager can check in automatically, issue the CHECKIN LIBVOLUME command,
specifying SEARCH=BULK. The server searches through all slots even if it
encounters an unavailable slot.

About this task

Issuing a REPLY command in response to a server request is not required if you

specify a wait time of zero using the optional WAITTIME parameter. Without the
requirement for a reply, the CHECKIN LIBVOLUME command is much easier to script
and requires less intervention. The default value for the WAITTIME parameter is 60
minutes.

To have Tivoli Storage Manager load a cartridge in a drive and read the label, you
must specify the CHECKLABEL=YES option. The CHECKLABEL=NO option is invalid with
the SEARCH=BULK option. After reading the label, Tivoli Storage Manager moves the
tape from the drive to a storage slot. When bar code reading is enabled with the
CHECKLABEL=BARCODE parameter, Tivoli Storage Manager reads the label and moves
the tape from the entry/exit port to a storage slot.

Checking in volumes from library storage slots

You can search storage slots for new volumes that have not yet been added to the
volume inventory and check those volumes into the library using the CHECKIN
LIBVOLUME command, specifying SEARCH=YES.

About this task

Issuing the SEARCH=YES parameter eliminates issuing an explicit CHECKIN LIBVOLUME

command for each volume. For example, for a SCSI device you can simply open
the library access door, place all of the new volumes in unused slots, close the
door, and issue the CHECKIN LIBVOLUME command with SEARCH=YES.

See “Element addresses for library storage slots” on page 132.

Chapter 7. Managing removable media operations 131

 Checkin of private volumes
Private volumes are volumes that are either predefined to a storage pool or
volumes that are partially-written. You can check in private volumes, but you must
assign a private status to them before checking them in.

Private volumes cannot be accidentally overwritten when a scratch mount is

requested. The server does not allow the administrator to check in a volume with
scratch status when that volume already belongs to a storage pool.

Partially-written volumes are always private volumes. Volumes begin with a status
of either scratch or private, but once Tivoli Storage Manager stores data on them,
their status becomes private. See “Returning partially-written volumes to
automated libraries” on page 136.

Checkin of volumes into full libraries

You can check volumes into devices that are fully populated and have no empty
storage slots by enabling tape swapping. Swapping allows Tivoli Storage Manager
to select and eject volumes to store in a different physical location.

Tivoli Storage Manager selects the volume to eject by checking first for any
available scratch volumes, then for the least frequently mounted volumes. Without
tape swapping, the checkin fails. See “Setting up volume overflow locations for
automated libraries” on page 138.

Checkin of volumes into IBM 3494 libraries

Volumes inserted into an IBM 3494 library are assigned to the insert category
(X'FF00').

When a volume is first inserted into an IBM 3494 library, either manually or
automatically at the convenience I/O station, the volume is assigned to the insert
category (X'FF00'). You can then change the category number when issuing the
CHECKIN LIBVOLUME command.

Element addresses for library storage slots

If a library has entry/exit ports, you can add and remove media by loading the
media into the ports. If there are no entry/exit ports, you must load tapes into
storage slots.

If you load tapes into storage slots, you must reply to mount requests that identify
storage slots with element addresses, unless you specify a wait time of zero when
issuing the CHECKIN LIBVOLUME or LABEL LIBVOLUME commands. (If the wait time is
zero, no reply is required.) An element address is a number that indicates the
physical location of a storage slot or drive within an automated library.

You need device names and element addresses when:

v Defining or updating drives in an automated library.
v Checking volumes into an automated library that has no entry/exit ports.
v Using a specific drive in an automated library to label volumes.
Element addresses are available in the device manufacturer's documentation or at
the following website: [Link]
Software/Tivoli/Tivoli_Storage_Manager.

132 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Write-once, read-many tape media
Write-once, read-many (WORM) media helps prevent accidental or deliberate
deletion of critical data. However, Tivoli Storage Manager imposes certain
restrictions and guidelines to follow when you use WORM media.

You can use the following types of WORM media with Tivoli Storage Manager:
v StorageTek VolSafe
v Sony AIT50 and AIT100
v IBM 3592
v IBM LTO-3 and LTO-4; HP LTO-3 and LTO-4; and Quantum LTO-3
v Quantum SDLT 600, Quantum DLT V4, and Quantum DLT S4

External and manual libraries use separate logical libraries to segregate their
media. You must ensure that the correct media is loaded.

Tips:
v A storage pool can consist of either WORM or RW media, but not both.
v To avoid wasting tape after a restore or import operation, avoid using WORM
tapes for database backup or export operations.
Related tasks:
“Selecting a device driver” on page 78

WORM-capable drives
To use WORM media in a library, all the drives in the library must be
WORM-capable. A mount will fail if a WORM cartridge is mounted in a read write
(RW) drive.

However, a WORM-capable drive can be used as a RW drive if the WORM

parameter in the device class is set to NO. Any type of library can have both
WORM and RW media if all of the drives are WORM enabled. The only exception
to this rule is NAS-attached libraries in which WORM tape media cannot be used.

Checkin of WORM media

The type of WORM media determines whether the media label needs to be read
during checkin.

Library changers cannot identify the difference between standard read-write (RW)
tape media and the following types of WORM tape media:
v VolSafe
v Sony AIT
v LTO
v SDLT
v DLT
To determine the type of WORM media that is being used, a volume must be
loaded into a drive. Therefore, when checking in one of these types of WORM
volumes, you must use the CHECKLABEL=YES option on the CHECKIN
LIBVOLUME command.

If they provide support for WORM media, IBM 3592 library changers can detect
whether a volume is WORM media without loading the volume into a drive.
Specifying CHECKLABEL=YES is not required. Verify with your hardware vendors
that your 3592 drives and libraries provide the required support.

Chapter 7. Managing removable media operations 133

 Restrictions on WORM media
You cannot use prelabeled WORM media with the LTO or ECARTRIDGE device
class. You cannot use WORM media in IBM LTO-4 drives, HP LTO-4, Oracle
StorageTek T10000B, or Oracle StorageTek T10000C drives with Tivoli Storage
Manager specified as the drive-encryption key manager.

Mount failures with WORM media

If WORM tape media are loaded into a drive for a read-write (RW) device-class
mount, it will cause a mount failure. Similarly, if RW tape media are loaded into a
drive for a WORM device-class mount, the mount will fail.

Relabeling WORM media

You cannot relabel a WORM cartridge if it contains data. This applies to Sony AIT
WORM, LTO WORM, SDLT WORM, DLT WORM, and IBM 3592 cartridges. The
label on a VolSafe volume should be overwritten only once and only if the volume
does not contain usable, deleted, or expired data.

Issue the LABEL LIBVOLUME command only once for VolSafe volumes. You can
guard against overwriting the label by using the OVERWRITE=NO option on the
LABEL LIBVOLUME command.

Removing private WORM volumes from a library

If you perform some action on a WORM volume (for example, if you delete file
spaces) and the server does not mark the volume as full, the volume is returned to
scratch status. If a WORM volume is not marked as full and you delete it from a
storage pool, the volume will remain private. To remove a private WORM volume
from a library, you must issue the CHECKOUT LIBVOLUME command.

Creation of DLT WORM volumes

DLT WORM volumes can be converted from read-write (RW) volumes.

If you have SDLT-600, DLT-V4, or DLT-S4 drives and you want to enable them for
WORM media, upgrade the drives using V30 or later firmware available from
Quantum. You can also use DLTIce software to convert unformatted read-write
(RW) volumes or blank volumes to WORM volumes.

In SCSI or automated-cartridge system-library software (ACSLS) libraries, the

Tivoli Storage Manager server creates scratch DLT WORM volumes automatically
when the server cannot locate any scratch WORM volumes in a library's inventory.
The server converts available unformatted or blank RW scratch volumes or empty
RW private volumes to scratch WORM volumes. The server also rewrites labels on
newly created WORM volumes using the label information on the existing RW
volumes.

In manual libraries, you can use the server to format empty volumes to WORM.

Support for short and normal 3592 WORM tapes

Tivoli Storage Manager supports both short and normal 3592 WORM tapes. For
best results, define them in separate storage pools

134 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 Querying a device class for the WORM-parameter setting
You can determine the setting of the WORM parameter for a device class by using
the QUERY DEVCLASS command. The output contains a field, labeled WORM,
and a value (YES or NO).

Managing media in automated libraries

Typically, automated libraries require little intervention after you set up a media
rotation. However, you might occasionally add, remove, or manually manage
media in automated libraries.

About this task

Tivoli Storage Manager tracks the media in the library volume inventory. The
library volume inventory is separate from the storage pool inventory for a library
device. To add volumes to the volume inventory for a device, check volumes into
the library device. For details on the CHECKIN procedure, see “Checking media into
automated library devices” on page 130. To add volumes to a storage pool for a
library, see “Adding scratch volumes to automated library devices” on page 138.

You can manage media in automated libraries by reviewing the following tasks:

Task Required Privilege Class

“Changing the status of automated library System or Unrestricted Storage
volumes”
“Removing volumes from automated System or Unrestricted Storage
libraries” on page 136
“Returning partially-written volumes to System or Unrestricted Storage
automated libraries” on page 136
“Auditing volume inventories in libraries” System or Unrestricted Storage
on page 137
“Adding scratch volumes to automated System or Unrestricted Storage
library devices” on page 138
“Category numbers for IBM 3494 libraries” System or Unrestricted Storage
on page 141
“Media reuse in automated libraries” on System or Unrestricted Storage
page 141

Changing the status of automated library volumes

You can change the status of a volume from private to scratch or from scratch to
private.

About this task

To change the status of volumes, issue the UPDATE LIBVOLUME command.

Private volumes must be administrator-defined volumes with either no data or
invalid data. They cannot be partially-written volumes containing active data.
Volume statistics are lost when volume statuses are modified.

Chapter 7. Managing removable media operations 135

 Removing volumes from automated libraries
You can remove volumes from automated libraries by issuing the CHECKOUT
LIBVOLUME command.

About this task

Tivoli Storage Manager mounts each volume and verifies its internal label before
checking it out of the volume inventory. After a volume has been checked out,
Tivoli Storage Manager moves the media to the entry/exit port of the device if it
has one, or Tivoli Storage Manager requests that the operator remove the volume
from a drive within the device.

For automated libraries with multiple entry/exit ports, you can issue the
CHECKOUT LIBVOLUME command with the SEARCH=BULK parameter. Tivoli
Storage Manager ejects the volume to the next available entry/exit port.

Partially-written volumes that are removed from the device will need to be
checked in again if Tivoli Storage Manager attempts to access them. See
“Partially-written volumes” on page 128.

Messages: When a volume is dismounted, TapeAlert information will be reported

in four possible messages. TapeAlert has only three severity levels, Critical,
Warning and Informational. Some Critical messages will result in ANR8481S, while
others will use ANRxxxxE, depending on the text. Examples of each warning type
are:
ANRxxxxS Device /dev/rmt1, volume VOL123 has issued the following
Critical TapeAlert: Your Data is at risk:
1. copy any data you require from this tape;
2. Do not use the tape again;
3. Restart the operation with a different tape.

ANRxxxxE Device /dec/lb0, volume NONE has issued the following

Critical TapeAlert: The library has a problem with the host interface:
1. Check the cables and cable connections;
2. Restart the operation.

ANRxxxxW Device /dev/lb0, volume NONE has issued the following

Warning TapeAlert: A hardware failure of the library is predicted.
Call the library supplier helpline.

ANRxxxxI Device /dev/mto, volume MYVOL1 has issued the following

Informational TapeAlert: You have tried to load a cartridge of
a type which is not supported by this drive

These messages indicate a hardware error, and not a Tivoli Storage Manager
application error.

Returning partially-written volumes to automated libraries

Partially-written volumes that are checked out of a library continue to be defined
to a storage pool and have a status of private.

About this task

To return partially-written volumes:

136 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 Procedure
1. Check in the volume by issuing the CHECKIN LIBVOLUME command with
STATUS=PRIVATE parameter.
2. Change the volume access from unavailable to read/write or read-only by
issuing the UPDATE VOLUME command with the ACCESS parameter.

Returning reclaimed volumes to a library

Tivoli Storage Manager can reuse volumes after valid data is reclaimed.

About this task

Scratch volumes are automatically returned to the library as scratch volumes. To

reuse private volumes, check them into the library.

Auditing volume inventories in libraries

Auditing the volume inventory ensures that the information maintained by the
Tivoli Storage Manager server is consistent with the physical media in the library.
Audits are useful when the inventory was manually manipulated.

About this task

To audit the volume inventories of automated libraries, issue the AUDIT LIBRARY
command . Tivoli Storage Manager deletes missing volumes and updates the
locations of volumes that have moved since the last audit. Tivoli Storage Manager
cannot add new volumes during an audit.

Unless devices are equipped with bar code readers, the server mounts each volume
during the audit process to verify the label. After the label has been verified, the
volume remains in a wait state until the mount retention interval times out. You
can save time by issuing the DISMOUNT VOLUME command to force idle
volumes to be dismounted.

Auditing volume inventories using bar code readers

You can save time when auditing volume inventories for devices equipped with
bar code readers by using the bar code reader to verify the identity of volumes.

About this task

If a volume has a bar code label with six characters or less, Tivoli Storage Manager
reads the volume name from the bar code label during the audit. The volume is
not mounted to verify that the external bar code name matches the internal,
recorded volume name.

If a volume has no bar code label or the bar code label does not meet Tivoli
Storage Manager label requirements, Tivoli Storage Manager mounts the volume in
a drive and attempts to read the internal label. See “Labeling media” on page 129.
For example, to audit the TAPELIB library using its bar code reader, issue the
following command:
audit library tapelib checklabel=barcode

Chapter 7. Managing removable media operations 137

 Adding scratch volumes to automated library devices
As the demand for media grows, you can add scratch volumes to libraries.

About this task

To increase the maximum number of scratch volumes:

Procedure
1. Check volumes into the library. Label them if necessary. You might need to
temporarily store volumes in an overflow location in order to make room for
additional scratch volumes. See “Setting up volume overflow locations for
automated libraries.”
2. Increase the maximum number of scratch volumes. You can issue the UPDATE
STGPOOL command to increase the number of scratch volumes that can be
added to a storage pool.

Results
The initial maximum number of scratch volumes for a library is determined when
the library storage pool is created. See “Defining volumes” on page 127.

Setting up volume overflow locations for automated libraries

As the demand for media grows, the number of volumes needed for a storage pool
may exceed the physical capacity of an automated library. To make room for new
volumes while keeping track of existing volumes, you can define a physical
location as an overflow area.

About this task

Tivoli Storage Manager tracks the volumes moved to the overflow area thus
allowing you to make storage slots available for new volumes. To set up and
manage an overflow location:

Procedure
1. Create a volume overflow location. Define or update the storage pool
associated with the automated library by issuing the DEFINE STGPOOL or
UPDATE STGPOOL command with the OVERFLOW parameter. For example,
to create an overflow location named ROOM2948 for a storage pool named
ARCHIVEPOOL, issue the following:
update stgpool archivepool ovflocation=Room2948
2. Move media to the overflow location as required. Issue the MOVE MEDIA
command to move media from the library to the overflow location. For
example, to move all full volumes in the specified storage pool out of the
library.
move media * stgpool=archivepool

All full volumes are checked out of the library, and Tivoli Storage Manager
records the location of the volumes as Room2948.
Use the DAYS parameter to specify the number of days that must elapse before
the volume is eligible for processing by the MOVE MEDIA command.
3. Check in new scratch volumes (if required). See “Checking media into
automated library devices” on page 130. If a volume has an entry in volume
history, you cannot check it in as a scratch volume.

138 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 4. Identify the empty scratch tapes in the overflow location. For example, enter
this command:
query media * stg=* whereovflocation=Room2948 wherestatus=empty
move media * stg=* wherestate=mountablenotinlib wherestatus=empty
cmd="checkin libvol autolib &vol status=scratch"
cmdfilename=\storage\move\media\[Link]
5. Check in volumes from the overflow area when Tivoli Storage Manager
requests them. Operators must check volumes in from the overflow area when
Tivoli Storage Manager needs them. Tivoli Storage Manager issues mount
requests that include the location of the volumes.
Operators can locate volumes in an overflow location by issuing the QUERY
MEDIA command. This command can also be used to generate commands. For
example, you can issue a QUERY MEDIA command to list the volumes in the
overflow location, and at the same time generate the commands to check those
volumes into the library. For example, enter this command:
query media format=cmd stgpool=archivepool whereovflocation=Room2948
cmd="checkin libvol autolib &vol status=private"
cmdfilename="\storage\move\media\[Link]"
Use the DAYS parameter to specify the number of days that must elapse before
the volumes are eligible for processing by the QUERY MEDIA command.
The file that contains the generated commands can be run using the Tivoli
Storage Manager MACRO command. For this example, the file may look like
this:
checkin libvol autolib TAPE13 status=private
checkin libvol autolib TAPE19 status=private

Modifying volume access modes

Occasionally, you might need to manipulate the access mode for volumes, for
example, when removing partially-written volumes from or returning them to
libraries.

About this task

To change the access mode of a volume, issue the UPDATE VOLUME command,
specifying ACCESS=UNAVAILABLE.

If you want to make volumes unavailable in order to send the data they contain
offsite for safekeeping, consider using copy storage pools or active-data pools
instead. You can back up primary storage pools to a copy storage pool and then
send the copy storage pool volumes offsite. You can also copy active versions of
client backup data to active-data pools, and then send the volumes offsite. You can
track copy storage pool volumes and active-data pool volumes by changing their
access mode to offsite, and updating the volume history to identify their location.
For more information, see “Backing up primary storage pools” on page 883.

Chapter 7. Managing removable media operations 139

 Shared libraries
Shared libraries are logical libraries that are represented physically by SCSI, 349X,
or ACSLS libraries. The Tivoli Storage Manager server is configured as a library
manager and controls the physical library. Tivoli Storage Manager servers using
the SHARED library type are library clients to the library manager server.

The library client contacts the library manager, when the library manager starts
and the storage device initializes, or after a library manager is defined to a library
client. The library client confirms that the contacted server is the library manager
for the named library device. The library client also compares drive definitions
with the library manager for consistency. The library client contacts the library
manager for each of the following operations:
Volume Mount
A library client sends a request to the library manager for access to a
particular volume in the shared library device. For a scratch volume, the
library client does not specify a volume name. If the library manager
cannot access the requested volume, or if scratch volumes are not available,
the library manager denies the mount request. If the mount is successful,
the library manager returns the name of the drive where the volume is
mounted.
Volume Release (free to scratch)
When a library client no longer needs to access a volume, it notifies the
library manager that the volume should be returned to scratch. The library
manager's database is updated with the volume's new location. The
volume is deleted from the volume inventory of the library client.

Table 9 shows the interaction between library clients and the library manager in
processing Tivoli Storage Manager operations.
Table 9. How SAN-enabled servers process Tivoli Storage Manager operations
Operation Library Manager Library Client

(Command)
Query library volumes Displays the volumes that Not applicable.
are checked into the library.
(QUERY LIBVOLUME) For private volumes, the
owner server is also
displayed.
Check in and check out Performs the commands to Not applicable.
library volumes the library device.
When a checkin operation
(CHECKIN LIBVOLUME, must be performed because
CHECKOUT LIBVOLUME) of a client restore, a request
is sent to the library manager
server.
Audit library inventory Performs the inventory Performs the inventory
synchronization with the synchronization with the
(AUDIT LIBRARY) library device. library manager server.
Label a library volume Performs the labeling and Not applicable.
checkin of media.
(LABEL LIBVOLUME)
Dismount a volume Sends the request to the Requests that the library
library device. manager server perform the
(DISMOUNT VOLUME) operation.

140 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 Table 9. How SAN-enabled servers process Tivoli Storage Manager operations (continued)
Operation Library Manager Library Client

(Command)
Query a volume Checks whether the volume Requests that the library
is owned by the requesting manager server perform the
(QUERY VOLUME) library client server and operation.
checks whether the volume
is in the library device.

Category numbers for IBM 3494 libraries

Category numbers for IBM 3494 Tape Library Dataservers identify volumes that
are used for the same purpose or application. To avoid data loss, ensure that each
application sharing the library uses unique category numbers.

A 3494 library has an intelligent control unit that tracks the category number of
each volume in the volume inventory. The category numbers are useful when
multiple systems share the resources of a single library. Typically, a software
application that uses a 3494 uses only volumes in categories that are reserved for
that application.

Media reuse in automated libraries

Reusing media in automated libraries is essentially the same as reusing media in
manual libraries except that less intervention is required for automated devices
than for manual devices.

You can set up expiration processing and reclamation processing and tune the
media rotation to achieve the desired results.
v Setting up expiration processing
Expiration processing is the same, regardless of the type of device and media on
which backups are stored. See “Running expiration processing to delete expired
files” on page 512.
v Setting up reclamation processing
For a storage pool associated with a library that has more than one drive, the
reclaimed data is moved to other volumes in the same storage pool. See
“Reclaiming space in sequential-access storage pools” on page 367.
v Returning reclaimed media to the storage pool
Most media can be returned to a storage pool after it has been reclaimed but
media containing database backups and database export data require you to
perform an additional step. For these volumes, you must issue the DELETE
VOLHISTORY command or the UPDATE LIBVOLUME command to change the
status of the volume.
When Tivoli Storage Manager backs up the database or exports server
information, Tivoli Storage Manager records information about the volumes used
for these operations in the volume history file. Volumes that are tracked in the
volume history file require the administrator to delete the volume information
from the volume history file. The volume history file is a key component of
server recovery and is discussed in detail in Chapter 29, “Protecting and
recovering the server infrastructure and client data,” on page 869.

Chapter 7. Managing removable media operations 141

 Tip: If your server uses the disaster recovery manager function, the volume
information is automatically deleted during MOVE DRMEDIA command
processing. For additional information about DRM, see Chapter 32, “Disaster
recovery manager,” on page 1005.
v Ensuring media are available
See “Tape rotation” on page 146.

Media management in manual libraries

Media for manually operated devices are stored outside of the device, for example,
in a filing cabinet. Operators must therefore mount and dismount media manually.

If you use a media manager for stand-alone devices can introduce unnecessary
administrative overhead.

The following table lists the required privilege classes you need to complete the
specified task.
Table 10. Privilege classes for library-specific tasks
Task Required privilege class
Modifying the status of manual device System or unrestricted storage
volumes
Removing volumes from a manual library Not applicable
device
Returning volumes to a manual library Not applicable
device
Adding more volumes to a manual library Not applicable
device
Reusing media in manual libraries Not applicable

The following list explains the tasks from Table 10.

Modifying the status of manual device volumes
You can modify the status of volumes, regardless of the type of device that
uses them, by using the UPDATE LIBVOLUME command. When you issue the
command, you can assign a private status to scratch volumes or assign a
scratch status to private volumes. Private volumes cannot be partially
written to with active data.
Removing volumes from a manual library device
You can remove volumes from manual devices at any time because the
server maintains no volume inventory for manually operated devices. No
checkout processing is required for manual devices.
Returning volumes to a manual library device
You can return volumes to manual devices at any time because the server
maintains no volume inventory for manual libraries. No check in
processing is required for manual libraries.
Adding volumes to a manual library device
See “Tape rotation” on page 146.
Reusing media in manual libraries
Reusing media in manual libraries is similar to reusing media in
automated libraries except that more human intervention is required for
manual devices. See “Media reuse in automated libraries” on page 141.

142 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Tivoli Storage Manager server requests
Tivoli Storage Manager displays requests and status messages to all administrative
clients that are started in console mode. These request messages often have a time
limit. If the request is not fulfilled within the time limit, the operation times out
and fails.

For manual libraries, Tivoli Storage Manager detects when there is a cartridge
loaded in a drive, so no operator reply is necessary. For automated libraries, the
CHECKIN LIBVOLUME and LABEL LIBVOLUME commands involve inserting
cartridges into slots and, depending on the value of the WAITTIME parameter,
issuing a reply message. (If the value of the parameter is zero, no reply is
required.) The CHECKOUT LIBVOLUME command involves inserting cartridges
into slots and, in all cases, issuing a reply message.

Task Required Privilege Class

“Starting the administrative client as a Any Administrator
server console monitor”
“Displaying information about volumes that Any Administrator
are currently mounted” on page 144
“Displaying information about mount Operator
requests that are pending” on page 144
“Replying to mount requests” on page 144 Operator
“Canceling mount requests” on page 144 Operator
“Responding to requests for volume System or Unrestricted Storage
checkin” on page 145
“Dismounting idle volumes” on page 145 Operator
“Dismounting volumes from stand-alone Operator
removable-file devices” on page 146

Starting the administrative client as a server console monitor

You can display mount requests and status messages by starting the administrative
client in console mode. However, if the server is started as a Windows service,
which is recommended, a server console is required to see messages.

About this task

To start a server console monitor from an operating system command line, enter
this command:

Procedure

dsmadmc -consolemode

Chapter 7. Managing removable media operations 143

 Displaying information about volumes that are currently
mounted
To display the volumes currently mounted by Tivoli Storage Manager, issue the
QUERY MOUNT command. The information lists mounted volumes, the drives on
which they are mounted, and whether the volumes are currently in use.

Displaying information about mount requests that are pending

You can display information about pending mount requests either by checking the
mount message queue on a server console monitor or by issuing the QUERY
REQUEST command.

About this task

Tivoli Storage Manager displays a message similar to the following:

ANR8352I Requests outstanding:
ANR8326I 001: Mount DLT volume VOL1 R/W in drive D1 (MT4) of library
MYMANLIB within 60 minutes.

Tivoli Storage Manager displays a three-digit request ID number as part of the

message. The request ID number can also be obtained by issuing a QUERY
REQUEST command. If the request requires the operator to provide a device to be
used for the mount, the second parameter for this command is a device name.

Replying to mount requests

Unless the specified wait time is zero, you must issue a REPLY command in
response to mount requests from automated libraries. Manual libraries do not
require a reply because Tivoli Storage Manager detects when there is a cartridge
loaded in the drive.

About this task

If a wait time greater than zero was specified, the server waits the specified
number of minutes before resuming processing.

The first parameter for the REPLY command is the three-digit request ID number
that indicates which of the pending mount requests has been completed. For
example, an operator can issue the following command to respond to request 001
in the previous code sample.
reply 001

Canceling mount requests

If a mount request for a manual library cannot be satisfied, you can issue the
CANCEL REQUEST command. Tivoli Storage Manager cancels the request and the
operation that required the volume fails.

About this task

The CANCEL REQUEST command must include the request identification number.
This number is included in the request message, or it can be obtained by issuing a
QUERY REQUEST command, as described in “Displaying information about
mount requests that are pending.”

144 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 Canceling mount requests for volumes that were removed from
library devices
You might occasionally remove media from a library with the intention of storing
or destroying the media. If, after the media have been removed, Tivoli Storage
Manager requests the volumes, then you can cancel the request with the CANCEL
REQUEST command.

About this task

To ensure that the server does not try to mount the requested volume again,
specify the PERMANENT parameter to mark the volume as unavailable.

For most of the requests associated with automated libraries, the Tivoli Storage
Manager CANCEL REQUEST command is not accepted by the server. An operator
must perform a hardware or system action to cancel the requested mount.

Responding to requests for volume checkin

The procedure for responding to request for volume checkin vary, depending on
whether the requested volume is available or unavailable.

About this task

Operators may occasionally need to check additional volumes into an automated

library, for example, when Tivoli Storage Manager cannot locate a volume it
requires from the volume inventory. If the requested volume is available, place the
volume in the device and check in the volume. See “Checking media into
automated library devices” on page 130.

If the volume requested is unavailable (lost or destroyed):

Procedure
1. Update the access mode of the volume to UNAVAILABLE by using the
UPDATE VOLUME command.
2. Cancel the server's request for checkin by using the CANCEL REQUEST
command. (Do not cancel the client process that caused the request.) To get the
ID of the request to cancel, issue the QUERY REQUEST command.

Results

If operators do not respond to checkin requests within the mount-wait period,

Tivoli Storage Manager marks the volume as unavailable. The mount-wait period
is set in the device class of the storage pool.

Dismounting idle volumes

After a volume becomes idle, it remains mounted for a time specified by the
mount retention parameter for the device class.

About this task

To explicitly request that an idle volume be dismounted, use the DISMOUNT

VOLUME command.

Using mount retention can reduce the access time if volumes are used repeatedly.
For information about setting mount retention times, see “Controlling the amount

Chapter 7. Managing removable media operations 145

 of time that a volume remains mounted” on page 182.

Dismounting volumes from stand-alone removable-file devices

For manual libraries, operators must respond to messages that require media (for
example, JAZ, DVD, and CD media) to be manually ejected from removable file
devices.

About this task

Tivoli Storage Manager checks the drive every seven seconds to see if the medium
has been ejected. A volume dismount is not considered complete until Tivoli
Storage Manager detects that the medium has been ejected from the drive or that a
different medium has been inserted into the drive.

Obtaining tape alert messages

Tape alert messages are generated by tape and library devices to report hardware
errors. These messages help to determine problems that are not related to the IBM
Tivoli Storage Manager server.

About this task

A log page is created and can be retrieved at any given time or at a specific time
such as when a drive is dismounted.

There are three severity levels of tape alert messages:

v Informational (for example, you may have tried to load a cartridge type that is
not supported)
v Warning (for example, a hardware failure is predicted)
v Critical (for example, there is a problem with the tape and your data is at risk)

Tape alert messages are turned off by default. To set tape alert messages to ON,
issue the SET TAPEALERTMSG command. To query tape alert messages, issue the
QUERY TAPEALERTMSG command.

Tape rotation
Over time, media ages, and certain backup data might no longer be needed. You
can reclaim useful data on media and then reclaim and reuse the media.

Tivoli Storage Manager policy determines how many backup versions are retained
and how long they are retained. Policy also determines how many inactive
versions of files the server keeps, and for how long. For more information about
policies to help you manage client data, see “Basic policy planning” on page 476.
Deleting data - expiration processing
Expiration processing deletes data that is no longer valid either because it
exceeds the retention specifications in policy or because users or
administrators deleted the active versions of the data.
Reusing media - reclamation processing
Data on tapes might expire, move, or be deleted. Reclamation processing
consolidates any unexpired data by moving it from multiple volumes onto
fewer volumes. The media can then be returned to the storage pool and
reused.

146 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 You can set a reclamation threshold that allows Tivoli Storage Manager to
reclaim volumes whose valid data drops below a threshold. The threshold
is a percentage of unused space on the volume and is set for each storage
pool. The amount of data on the volume and the reclamation threshold for
the storage pool affects when the volume is reclaimed.
Determining when media have reached end of life
You can use Tivoli Storage Manager to display statistics about volumes,
which include the number of write operations that are completed on the
media and the number of write errors. Tivoli Storage Manager overwrites
this statistical data for media that are initially defined as scratch volumes
each time the media are reclaimed. For media initially defined as private
volumes, Tivoli Storage Manager maintains this statistical data, even as the
volume is reclaimed. You can compare the information with the number of
write operations and write errors that are recommended by the
manufacturer.
Reclaim any valid data from volumes that have reached the end of life. If
the volumes are in automated libraries, check them out of the volume
inventory. Delete private volumes from the database.
Ensuring media are available for the tape rotation
Over time, the demand for volumes might cause the storage pool to run
out of space. You can set the maximum number of scratch volumes high
enough to meet demand by doing one or both of the following:
v Increase the maximum number of scratch volumes by updating the
storage pool definition. Label and check in new volumes to be used as
scratch volumes if needed.
v Make volumes available for reuse by running expiration processing and
reclamation to consolidate data onto fewer volumes.
Write-once-read-many (WORM) drives can waste media when Tivoli
Storage Manager cancels transactions because volumes are not available to
complete the backup. When Tivoli Storage Manager writes to WORM
volumes, the space on the volumes cannot be reused, even if the
transactions are canceled (for example, if a backup is canceled because of a
shortage of media in the device).
To minimize wasted WORM media:
1. Ensure that the maximum number of scratch volumes for the device
storage pool is at least equal to the number of storage slots in the
library.
2. Check enough volumes into the volume inventory for the expected
load.
If most backups are small files, controlling the transaction size can affect
how WORM media is used. Smaller transactions mean that less space is
wasted if a transaction such as a backup must be canceled. Use the
TXNGROUPMAX server option and the TXNBYTELIMIT client option to control
transaction size.
Related concepts:
“Media reuse in automated libraries” on page 141
“Media management in manual libraries” on page 142
Related tasks:
“Reclaiming space in sequential-access storage pools” on page 367
“Setting up volume overflow locations for automated libraries” on page 138

Chapter 7. Managing removable media operations 147

 Related reference:
“File expiration and expiration processing” on page 479
“Running expiration processing to delete expired files” on page 512

Labeling volumes using commands

All media require labels. You can label volumes with the LABEL LIBVOLUME
command.

About this task

The following example demonstrates using the LABEL LIBVOLUME command to

label tapes for a manual library and for an automated library. Assume the
automated device is attached to SCSI address 4, and the manual device is attached
to SCSI address 5. You want to insert media into the device's entry/exit ports and
you want the device's bar code reader to read bar code labels and overwrite
existing labels with the information on the bar code label.

To automatically label tape volumes in SCSI-type libraries, you can use the
AUTOLABEL parameter on the DEFINE LIBRARY and UPDATE LIBRARY
commands. Using this parameter eliminates the need to pre-label a set of tapes. It
is also more efficient than using the LABEL LIBVOLUME command, which
requires you to mount volumes separately. If you use the AUTOLABEL parameter,
you must check in tapes by specifying CHECKLABEL=BARCODE on the
CHECKIN LIBVOLUME command.

Automated library device:

label libvolume storagelibname overwrite=yes labelsource=barcode

Manual library device:

label libvolume storagelibname overwrite=yes labelsource=barcode
label libvolume storagelibname volname

Tip: To automatically label tape volumes in SCSI-type libraries, you can use the
AUTOLABEL parameter on the DEFINE LIBRARY and UPDATE LIBRARY
commands. Using this parameter eliminates the need to pre-label a set of tapes. It
is also more efficient than using the LABEL LIBVOLUME command, which
requires you to mount volumes separately. If you use the AUTOLABEL parameter,
you must check in tapes by specifying CHECKLABEL=BARCODE on the
CHECKIN LIBVOLUME command.

Tivoli Storage Manager media-manager support

While Tivoli Storage Manager tracks and manages client data, the removable
media manager labels, catalogs, and tracks physical volumes. The media manager
also controls libraries, drives, slots, and doors.

Tivoli Storage Manager works with removable media managers to control storage.
Media managers help Tivoli Storage Manager make better use of media resources.
To use a media manager with Tivoli Storage Manager, you must define a Tivoli
Storage Manager library that represents the media manager. Defining the library is
similar to defining any other type of library to Tivoli Storage Manager, except the
library does not represent a physical device. Different library types are required for
External Media Management Interface control.

148 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 External Media Management Interface
The External Media Management Interface uses the EXTERNAL library
type. The EXTERNAL library type does not map to a device or media type,
but instead specifies the installed path of the external media manager. See
“Using external media managers to control media.”

Using external media managers to control media

The External Media Management API enables you to use external media manager
software to control your media.

About this task

For details about the interface, see Appendix A, “External media management
interface description,” on page 1085.

The following sample procedure, describes how to set up an 8 mm automated tape

library to use the External Media Management Interface with a media manager.

Procedure
1. Set up the media manager to interface with Tivoli Storage Manager. For more
information, see Appendix A, “External media management interface
description,” on page 1085.
2. Define a library whose library type is EXTERNAL, and define a path to the
media manager executable. For example:
define library medman libtype=external
For example:
define devclass class1 devtype=8mm library=medman mountretention=5 mountlimit=2
define path server1 medman srctype=server desttype=library
externalmanager=c:\server\[Link]

The MOUNTLIMIT parameter specifies the number of drives in the library. The
MOUNTRETENTION parameter determines how long an idle volume remains in a
drive. If the library is shared among applications, this setting is especially
important. Some media managers do not dismount an allocated drive to satisfy
pending requests. You must set the mount retention period to balance
competing mount requests and system performance.
3. Define a device class for the EXTERNAL library with a device type of 8MM.
4. Define a storage pool for the device class. For example:
define stgpool pool1 class1 maxscratch=500
5. Associate client nodes with the new storage pool by defining a new policy
domain or by updating an existing policy domain

Requirements for managing media in external libraries

There are certain unique requirements for managing media in external libraries

When managing media in external libraries, consider the following guidelines:

v You do not need to check in and label media in external libraries. Those media
are not tracked in the Tivoli Storage Manager volume inventory, and the media
manager handles labeling. However, you must ensure that an adequate supply
of blank media are available.
v If you are using disaster recovery manager, you can issue the MOVE DRMEDIA
command to issue an operator request to remove the media from the library. For
more information, see Chapter 32, “Disaster recovery manager,” on page 1005.

Chapter 7. Managing removable media operations 149

 v You should not migrate media from a Tivoli Storage Manager SCSI library to an
external library. Instead, use external media management on a new Tivoli
Storage Manager configuration or when defining externally managed devices to
Tivoli Storage Manager.
v Deleting externally managed storage pools requires that you delete any volumes
associated with the Tivoli Storage Manager library. When the library is deleted,
the externally managed storage pool associated with that library is also deleted.
For more information, see “Deleting storage pool volumes that contain data” on
page 415.

Troubleshooting database errors

Error conditions can cause the Tivoli Storage Manager volume database and the
media manager's volume database to become unsynchronized.

About this task

The most likely symptom of this problem is that the volumes in the media
manager's database are not known to Tivoli Storage Manager, and thus not
available for use. Verify the Tivoli Storage Manager volume list and any disaster
recovery media. If volumes not identified to Tivoli Storage Manager are found, use
the media manager interface to deallocate and delete the volumes.

Managing libraries
Using Tivoli Storage Manager commands, you can query and delete libraries. You
can also update automated libraries.

Obtaining information about libraries

Standard and detailed reports provide information about libraries.

About this task

Task Required Privilege Class
Request information about libraries Any administrator

To obtain information about libraries, use the QUERY LIBRARY command. The
default is a standard report. For example, to display information about all libraries
in a standard report, issue the following command:
query library

The following shows an example of output from this command:

Library Library Private Scratch WORM Scratch External
Name Type Category Category Category Manager
------- ------- -------- -------- ------------ --------
MANLIB MANUAL
EXB SCSI
3494LIB 349X 300 301 302

150 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Updating automated libraries
You can update an existing automated library by issuing the UPDATE LIBRARY
command. To update the device names of a library, issue the UPDATE PATH
command. You cannot update a MANUAL library.

About this task

Task Required Privilege Class
Update libraries System or unrestricted storage

If your system or device is reconfigured, and the device name changes, you may
need to update the device name. The examples below show how you can issue the
UPDATE LIBRARY and UPDATE PATH commands for the following library types:
v SCSI
v 349X
v ACSLS
v External

Examples:
v SCSI Library
Update the path from SERVER1 to a SCSI library named SCSILIB:
update path server1 scsilib srctype=server desttype=library device=lb4.0.0.0
Update the definition of a SCSI library named SCSILIB defined to a library client
so that a new library manager is specified:
update library scsilib primarylibmanager=server2
v 349X Library
Update the path from SERVER1 to an IBM 3494 library named 3494LIB with
new device names.
update path server1 3494lib srctype=server desttype=library
device=lb2.0.0.0,lb3.0.0.0,lb4.0.0.0
Update the definition of an IBM 3494 library named 3494LIB defined to a library
client so that a new library manager is specified:
update library 3494lib primarylibmanager=server2
v ACSLS Library
Update an automated cartridge system library software (ACSLS) library named
ACSLSLIB with a new ID number.
update library acslslib ascid=1
v External Library
Update an external library named EXTLIB with a new media manager path
name.
update path server1 extlib srctype=server desttype=library
externalmanager=c:\server\[Link]
Update an EXTERNAL library named EXTLIB in a LAN-free configuration so
that the server uses the value set for mount retention in the device class
associated with the library:
update library extlib obeymountretention=yes

Chapter 7. Managing removable media operations 151

 Deleting libraries
Before you delete a library with the DELETE LIBRARY command, you must delete
all of the drives and drive paths that have been defined as part of the library and
delete the path to the library.

About this task

Task Required Privilege Class
Delete libraries System or unrestricted storage

For information about deleting drives, see “Deleting drives” on page 168.

For example, suppose that you want to delete a library named 8MMLIB1. After
deleting all of the drives defined as part of this library and the path to the library,
issue the following command to delete the library itself:
delete library 8mmlib1

Managing drives
You can query, update, and delete drives.

Requesting information about drives

You can request information about drives by using the QUERY DRIVE command.

About this task

Task Required Privilege Class
Request information about drives Any administrator

The QUERY DRIVE command accepts wildcard characters for both a library name
and a drive name. See the Administrator's Reference for information about using
wildcard characters.

For example, to query all drives associated with your server, issue the following
command:
query drive

The following shows an example of the results of this command.

Library Drive Device On Line
Name Name Type
-------- ------- --------- -------
MANLIB 8MM.0 8MM Yes
AUTOLIB 8MM.2 8MM Yes

152 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Updating drives
You can change the attributes of a drive by issuing the UPDATE DRIVE command.

About this task

Task Required Privilege Class
Update drives System or unrestricted storage

You can change the following attributes of a drive by issuing the UPDATE DRIVE
command.
v The element address, if the drive is in a SCSI or virtual tape library (VTL).
v The ID of a drive in an automated cartridge system library software (ACSLS)
library
v The cleaning frequency
v Change whether the drive is online or offline
For example, to change the element address of a drive named DRIVE3 to 119, issue
the following command:
update drive auto drive3 element=119

Note: You cannot change the element number if a drive is in use. If a drive has a
volume mounted, but the volume is idle, it can be explicitly dismounted as
described in “Dismounting idle volumes” on page 145.

If you are reconfiguring your system, you can change the device name of a drive
by issuing the UPDATE PATH command. For example, to change the device name of a
drive named DRIVE3, issue the following command:
update path server1 drive3 srctype=server desttype=drive library=scsilib
device=mt3.0.0.0

You can change a drive to offline status while the drive is in use. Tivoli Storage
Manager finishes with the current tape in the drive, and then does not use the
drive anymore. By changing a drive to offline, you can drain work off a drive.
However, if the tape that was in use was part of a series of tapes for a single
transaction, the drive is not available to complete the series. If no other drives are
available, the transaction might fail. If all drives in a library are made offline, any
attempts by Tivoli Storage Manager to write to the storage pool associated with the
library fails.

The ONLINE parameter specifies the value of the drive's online state, even if the
drive is in use. ONLINE=YES indicates that the drive is available for use (online).
ONLINE=NO indicates that the drive is not available for use (offline). This
parameter is optional. Do not specify other optional parameters along with
ONLINE=YES or ONLINE=NO. If you do, the drive is not updated, and the
command fails when the drive is in use. This command can be issued when the
drive is involved in an active process or session, but this action is not
recommended.

The ONLINE parameter allows drives to be taken offline and used for another
activity, such as maintenance. If you make the drive offline while it is in use, the
drive is marked offline. However, the mounted volume completes its current
process. If this volume was part of a series of volumes for a given transaction, the
drive is no longer available to complete mounting the series. If no other drives are
available, the active process might fail. The updated state is retained even when

Chapter 7. Managing removable media operations 153

 the server is halted and restarted. If a drive is marked offline when the server is
restarted, a warning is issued noting that the drive must be manually brought
online. If all the drives in a library are updated to be offline, processes requiring a
library mount point fail, rather than queue up for one.

Using drive encryption

You can use drive encryption to protect tapes that contain critical or sensitive data
(for example, tapes that contain sensitive financial information). Drive encryption
is particularly beneficial for tapes that you move from the Tivoli Storage Manager
server environment to an offsite location.

About this task

Tivoli Storage Manager supports encryption for the following drives:

v IBM 3592 generation 2 and later
v IBM LTO generation 4 and later
v HP LTO-4 generation 4 and later
v Oracle StorageTek T10000B
v Oracle StorageTek T10000C

Drives must be able to recognize the correct format. With Tivoli Storage Manager,
you can use the following encryption methods:
Table 11. Encryption methods supported
Application method Library method System method
3592 generation 2 Yes Yes Yes
and later
IBM LTO-4 Yes Yes, but only if your Yes
system hardware (for
example, 3584)
supports it
HP LTO-4 Yes No No
Oracle StorageTek Yes No No
T10000B
Oracle StorageTek Yes No No
T10000C

To enable drive encryption with IBM LTO-4, you must have the IBM RMSS
Ultrium device driver installed. SCSI drives do not support IBM LTO-4 encryption.
To enable encryption with HP LTO-4, you must have the Tivoli Storage Manager
device driver installed.

To enable drive encryption, specify the DRIVEENCRYPTION parameter on the

DEFINE DEVCLASS and UPDATE DEVCLASS commands for the 3592, LTO, and
ECARTRIDGE device types.

A library can contain a mixture of drives, some of which support encryption and
some which do not. (For example, a library might contain two LTO-2 drives, two
LTO-3 drives, and two LTO-4 drives.) You can also mix media in a library using,
for example, a mixture of encrypted and non-encrypted device classes having
different tape and drive technologies. However, all LTO-4 drives must support
encryption if Tivoli Storage Manager is to use drive encryption. In addition, all

154 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 drives within a logical library must use the same method of encryption. Tivoli
Storage Manager does not support an environment in which some drives use the
Application method and some drives use the Library or System methods of
encryption.

When using encryption-capable drives with a supported encryption method, a new

format will be used to write encrypted data to tapes. If data is written to volumes
using the new format and if the volumes are then returned to scratch, they will
contain labels that are only readable by encryption-enabled drives. To use these
scratch volumes in a drive that is not enabled for encryption, either because the
hardware is not capable of encryption or because the encryption method is set to
NONE, you must relabel the volumes.

For more information on setting up your hardware environment to use drive

encryption, refer to your hardware documentation. For details about drive
encryption, see:
v “Encrypting data with drives that are 3592 generation 2 and later” on page 185
v “Encrypting data using LTO generation 4 tape drives” on page 193
v “Enabling ECARTRIDGE drive encryption” on page 196 and “Disabling
ECARTRIDGE drive encryption” on page 197

Validating data during read/write operations to tape

To validate data and identify data that is corrupted, you can use a feature called
logical block protection. If you use logical block protection, Tivoli Storage Manager
inserts a cyclic redundancy check (CRC) value at the end of each logical block of
data to be written to tape.

About this task

With logical block protection, you can identify errors that occur while data is being
written to tape and while data is transferred from the tape drive to Tivoli Storage
Manager through the storage area network. Drives that support logical block
protection validate data during read and write operations. The Tivoli Storage
Manager server validates data during read operations.

If validation by the drive fails during write operations, it can indicate that data
was corrupted while being transferred to tape. The Tivoli Storage Manager server
fails the write operation. You must restart the operation to continue. If validation
by the drive fails during read operations, it can indicate that the tape media is
corrupted. If validation by the Tivoli Storage Manager server fails during read
operations, it can indicate that data was corrupted while being transferred from the
tape drive and the server tries the operation again. If validation fails consistently,
the Tivoli Storage Manager server issues an error message that indicates hardware
or connection problems.

If logical block protection is disabled on a tape drive, or the drive does not support
logical block protection, the Tivoli Storage Manager server can read protected data.
However, the data is not validated.

Logical block protection is superior to the CRC validation that you can specify
when you define or update a storage pool definition. When you specify CRC
validation for a storage pool, data is validated only during volume auditing
operations. Errors are identified after data is written to tape.

Chapter 7. Managing removable media operations 155

 Restriction: You cannot use logical block protection for sequential data such as
backup sets and database backups.

Drives that support logical block protection

Logical block protection is available only for 3592, LTO, and ECARTRIDGE device
types. Capable 3592 drives include IBM TS1130, TS1140, and later generations.
Capable LTO drives include IBM LTO-5 and supported LTO-6 drives. Capable
Oracle StorageTek drives include those with the T10000C format.

The following table shows the media and the formats that you can use with drives
that support logical block protection.

Drive Tape media Drive formats

IBM TS1130 3592 Generation 2 3592-3 and 3592-3C
IBM TS1140 3592 Generation 2 Generation 2: 3592-3 and
3592-3C
3592 Generation 3
Generation 3: 3592-4 and
3592-4C
IBM LTO-5 LTO-5 ULTRIUM5 and ULTRIUM5C
IBM LTO-6 LTO-6 ULTRIUM6 and ULTRIUM6C

LTO-5 ULTRIUM5 and ULTRIUM5C

Oracle T10000C Oracle StorageTek T10000 T2 T10000C and T10000C-C

Tip: If you have a 3592, LTO, or Oracle StorageTek drive that is not capable of
logical block protection, you can upgrade the drive with firmware that provides
logical block protection.

Logical block protection is only available for drives that are in MANUAL, SCSI,
349x, and ACSLS libraries. Logical block protection is not available for drives that
are in external libraries. For the most current information about support for logical
block protection, see [Link]
[Link]?uid=swg21568108.

To use logical block protection for write operations, all the drives in a library must
support logical block protection. If a drive is not capable of logical block
protection, volumes that have read/write access are not mounted. However, the
server can use the drive to mount volumes that have read-only access. The
protected data is read and validated by the Tivoli Storage Manager server if logical
block protection is enabled for read/write operations.

Enabling and disabling logical block protection

You can specify logical block protection for read and write operations, or only for
write operations. You can also disable logical block protection if it is enabled. By
default, logical block protection is disabled because of performance effects that
result from CRC validation on the server and the tape drive.

About this task

Read/write operations to empty or filling volumes depend on whether the

volumes have logical block protection. Protected and unprotected data blocks
cannot be mixed on the same volume. If you change the setting for logical block
protection, the change applies only to empty volumes. Filling and full volumes
maintain their status of logical block protection until they are empty and ready to

156 IBM Tivoli Storage Manager for Windows: Administrator's Guide

be refilled. For example, suppose that you disable logical block protection. If the
server selects a volume that is associated with a device class that has logical block
protection, the server continues writing protected data to the volume.

To enable logical block protection, specify the LBPROTECT parameter on the DEFINE
DEVCLASS or the UPDATE DEVCLASS command for the 3592, LTO, and ECARTRIDGE
device types:

Procedure
v To enable logical block protection, specify a value of READWRITE or
WRITEONLY for the LBPROTECT parameter.
For example, to specify logical block protection during read/write operations for
a 3592 device class named 3592_lbprotect, issue the following command:
define devclass 3592_lbprotect library=3594 lbprotect=readwrite

Tips:
– If you update the value of the LBPROTECT parameter from NO to READWRITE
or WRITEONLY and the server selects a filling volume without logical block
protection for write operations, the server issues a message each time the
volume is mounted. The message indicates that data will be written to the
volume without logical block protection. To prevent this message from
displaying or to have Tivoli Storage Manager only write data with logical
block protection, update the access of filling volumes without logical block
protection to read-only.
– To reduce the performance effects, do not specify the CRCDATA parameter on
the DEFINE STGPOOL or UPDATE STGPOOL command.
– When data is validated during read operations by both the drive and by the
Tivoli Storage Manager server, it can slow server performance during restore
and retrieval operations. If the time that is required for restore and retrieval
operations is critical, you can change the setting of the LBPROTECT parameter
from READWRITE to WRITEONLY to increase the restore or retrieval speed.
After data is restored or retrieved, you can reset the LBPROTECT parameter to
READWRITE.
v To disable logical block protection, specify a value of NO for the LBPROTECT
parameter.

Restriction: If logical block protection is disabled, the server does not write to
an empty tape with logical block protection. However, if a filling volume with
logical block protection is selected, the server continues to write to the volume
with logical block protection. To prevent the server from writing to tapes with
logical block protection, change access of filling volumes with logical block
protection to read-only. When data is read, the CRC on each block is not
checked by either drive or the server.

If a disaster occurs and the disaster recovery site does not have drives that
support logical block protection, you must set the LBPROTECT parameter to NO. If
the tape drives are used for write operations, you must change the volume
access for volumes with protected data to read-only to prevent the server from
using the volumes.

If the server is to enable logical block protection, the server issues an error
message that indicates that the drive does not support logical block protection.

Chapter 7. Managing removable media operations 157

 What to do next

To determine whether a volume has logical block protection, issue the QUERY
VOLUME command and verify the value in the field Logical Block Protection.

Read/write operations to volumes with logical block protection

Read/write operations to empty or filling volumes depend on whether the
volumes have logical block protection. Protected and unprotected data blocks
cannot be mixed on the same volume.

If you use the UPDATE DEVCLASS command to change the setting for logical block
protection, the change applies only to empty volumes. Filling and full volumes
maintain their status of logical block protection until they are empty and ready to
be refilled.

For example, suppose that you change the value of the LBPROTECT parameter from
READWRITE to NO. If the server selects a volume that is associated with the
device class and that has logical block protection, the server continues writing
protected data to the volume.

Remember:
v Before you select the volume, the Tivoli Storage Manager server does not verify
whether the volume has logical block protection.
v If a drive does not support logical block protection, the mounts of volumes with
logical block protection for write operations fail. To prevent the server from
mounting the protected volumes for write operations, change the volume access
to read-only. Also, disable logical block protection to prevent the server from
enabling the feature on the tape drive.
v If a drive does not support logical block protection, and logical block protection
is disabled, the server reads data from protected volumes. However, the data is
not validated by the server and the tape drive.

To determine whether a volume has logical block protection, issue the QUERY
VOLUME command and verify the value in the field Logical Block Protection.

Tip: Consider updating the access of filling volumes to read-only if you update the
value of the LBPROTECT parameter in one of the following ways:
v READWRITE or WRITEONLY to NO
v NO to READWRITE or WRITEONLY
For example, suppose that you change the setting of the LBPROTECT parameter from
NO to READWRITE. If the server selects a filling volume without logical block
protection for write operations, the server issues a message each time the volume
is mounted. The message indicates that data will be written to the volume without
logical block protection. To prevent this message from being displayed or to have
Tivoli Storage Manager only write data with logical block protection, update the
access of filling volumes without logical block protection to read-only.

158 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 Storage pool management
To mix protected and unprotected data in a library, you must create different
device classes and different storage pools to separate the data. The device classes
associated with the protected data can specify logical block protection for
read/write operations or only for write operations.

Suppose, for example, that you have a 3584 library that has LTO-5 drives and that
you want to use for protected and unprotected data. To define the required device
classes and storage pools, you can issue the following commands.
define library 3584 libtype=scsi
define devclass lbprotect library=3584 devicetype=lto lbprotect=readwrite
define devclass normal library=3584 devicetype=lto lbprotect=no
define stgpool lbprotect_pool lbprotect maxscratch=10
define stgpool normal_pool normal maxscratch=10

Replacing tape drives

If you replace a drive in a tape library that is defined to IBM Tivoli Storage
Manager, you must delete the drive and path definitions for the old drive and
define the new drive and path.

Replacing drive and path definitions is required even if you are exchanging one
drive for another of the same type, with the same logical address, physical address,
SCSI ID, and port number. Device alias names can change when you change your
drive connections.

If the new drive is an upgrade that supports a new media format, you might be
required to define a new logical library, device class, and storage pool. Procedures
for setting up policy for a new drive in a multiple-drive library varies, depending
on the types of drives and media in the library.

Preventing errors caused by media incompatibility

Understanding media compatibility issues can prevent errors. Sometimes a new
drive has a limited ability to use media formats supported by a previous version of
the drive. Often, a new drive can read but not write to the old media.

About this task

By default, existing volumes with a status of FILLING will remain in that state
after a drive upgrade. In some cases, you might want to continue using an older
drive to fill these volumes. This will preserve read/write capability for the existing
volumes until they have been reclaimed. If you choose to upgrade all of the drives
in a library, pay attention to the media formats supported by the new hardware.
Unless you are planning to use only the latest media with your new drive, you
will need to be aware of any compatibility issues. For migration instructions, see
“Migrating to upgraded drives” on page 163.

To use a new drive with media it can read but not write to, issue the UPDATE
VOLUME command to set the access for those volumes to read-only. This will
prevent errors caused by read/write incompatibility. For example, a new drive
may eject media written in a density format it does not support as soon as the
media is loaded into the drive. Or a new drive may fail the first write command to
media partially written in a format it does not support.

When data on the read-only media expires and the volume is reclaimed, replace it
with media that is fully compatible with the new drive. Errors can be generated if
a new drive is unable to correctly calibrate a volume written using an older

Chapter 7. Managing removable media operations 159

 format. To avoid this problem, ensure that the original drive is in good working
order and at current microcode levels.

Removing drives
Drive removal requires a new drive and path definition.

About this task

To remove a drive:

Procedure
1. Stop the IBM Tivoli Storage Manager server and shut down the operating
system.
2. Remove the old drive and follow the manufacturer's instructions to install the
new drive.
3. Restart the operating system and the IBM Tivoli Storage Manager server.
4. Delete the path from the server to the drive. For example:
delete path server1 lib1 srctype=server desttype=drive
5. Delete the drive definition. For example, to delete a drive named DLT1 from a
library device named LIB1, enter:
delete drive lib1 dlt1
6. Define the new drive and path. This procedure will vary, depending on the
configuration of drives in your library. See “Defining new drives.”

Defining new drives

How you define a new drive depends on several factors, including whether the
new drive is a replacement or an upgrade, whether you plan to use different drive
types within the same library, and whether you plan to use different media types
within the same library.

Replacing drives with others of the same type:

To add a drive that supports the same media formats as the drive it replaces, you
need to define a new drive and path.

About this task

For example, to define a new drive and name it DRIVE1 and a path to it from
SERVER1, enter the following commands:
define drive lib1 drive1

define path server1 drive1 srctype=server desttype=drive library=lib1

device=mt3.0.0.1

You can use your existing library, device class, and storage pool definitions.

160 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Upgrading all of the drives in a library that contained only one type of drive:

To upgrade all the drives in a library that contained only one type of drive, you
need to define a new drive and path. You also need to update device class and
storage pool definitions.

About this task

You must decide how to manage any new types of media supported by the new
drives. See “Preventing errors caused by media incompatibility” on page 159 for
more information.

The following scenario assumes you already have a library device defined as
follows:
Library Library Private Scratch WORM Scratch External
Name Type Category Category Category Manager
------- ------- -------- -------- ------------ --------
LIB1 349X 200 201

Define each new drive and path

For example, to define a new drive and name it DRIVE1, enter:
define drive lib1 drive1

define path server1 drive1 srctype=server desttype=drive library=lib1

device=mt3.0.0.1
Update device class and storage pool definitions
v If you plan to use only one type of media in the library, you can use
your existing device class and storage pool definitions.
v If you plan to use media with different capacities in the same library,
you can define separate device classes and storage pools for each type of
media. This will provide accurate capacity reporting for each type of
media.
For example, if you plan to use both 3590E and 3590H tapes, you can
define two device classes:
define devclass 3590E_class devtype=3590 format=drive library=lib1
estcapacity=20g
define devclass 3590H_class devtype=3590 format=drive library=lib1
estcapacity=40g

Note: You must specify FORMAT=DRIVE for the new device classes.
You can then define two storage pools to divide the tapes within the
library:
define stgpool 3590E_pool 3590E_class maxscratch=number_of_3590E_tapes
define stgpool 3590H_pool 3590H_class maxscratch=number_of_3590H_tapes

Finally, you can issue the DEFINE VOLUME command to associate

media with the appropriate storage pool.

Chapter 7. Managing removable media operations 161

 Upgrading some of the drives in a library that contained only one type of
drive:

To upgrade some of the drives in a library that contained only one type of drive,
you need to define a separate logical library for each type of drive.

About this task

If an automated cartridge system library software (ACSLS), 349X, Manual, or

External library contains only one type of drive and you upgrade only a subset of
those drives, you must define an additional logical library. For SCSI libraries, we
do not support upgrading one type of drive if the new drives cannot read and
write in the format of the existing media. If the new drives can only read some of
the media, they must upgrade all of the drives.

The following scenario assumes you already have a library device defined as
follows:
Library Library Private Scratch WORM Scratch External
Name Category Category Category Manager
------- ------- -------- -------- ------------ --------
LIB1 349X 200 201

Define a new logical library and path for each new type of drive
For example, to add a logical library named LIB2 for the same physical
device already defined as LIB1, enter:
define library lib2 libtype=349X privatecategory=300 scratchcategory=301
wormscratchcategory=302

define path server1 lib2 srctype=server desttype=library

device=lb3.0.0.0
Define each new drive and path to the new library
To define a new drive named DRIVE2 to the logical library named LIB2
and a new path to the drive, enter:
define drive lib2 drive2

define path server1 drive1 srctype=server desttype=drive library=lib2

device=mt3.0.0.1
Update device class and storage pool definitions
To define a new device class, enter:
define devclass new_dev_class devtype=3592 worm=yes format=drive
library=lib2 estcapacity=40G

For accurate reporting of capacity information, you must specify the

ESTCAPACITY parameter.
To define a new storage pool, enter:
define stgpool new_stg_pool new_dev_class maxscratch=number_of_new_tapes

You can then issue the CHECKIN LIBVOLUME command to check the
new media into the logical library LIB2.

162 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Upgrading all of the drives in a library that contained more than one type of
drive:

To upgrade all the drives in a library that contained more than one type of drive,
you need to update the drive and path definitions for each logical library.

About this task

The following scenario assumes you already have two logical libraries defined. For
example:

Library Library Private Scratch WORM Scratch External

Name Category Category Category Manager
------- ------- -------- -------- ------------ --------
LIB1 349X 200 201
LIB2 349X 300 301 302

Update drive and path definitions for each logical library

For each library, follow the guidelines in “Upgrading all of the drives in a
library that contained only one type of drive” on page 161. For accurate
reporting of capacity information, you cannot use a global scratch pool
with this configuration.

Upgrading some of the drives in a library that contained more than one type of
drive:

To upgrade some ofl the drives in a library that contained more than one type of
drive, you need to update the drive and path definitions for each logical library.

About this task

The following scenario assumes you already have two logical libraries defined, for
example:

Library Library Private Scratch WORM Scratch External

Name Category Category Category Manager
------- ------- -------- -------- ------------ --------
LIB1 349X 200 201
LIB2 349X 300 301 302

You must update the drive and path definitions for each logical library. Follow the
guidelines in “Upgrading some of the drives in a library that contained only one
type of drive” on page 162. For accurate reporting of capacity information, you
cannot use a global scratch pool with this configuration.

Migrating to upgraded drives

If you upgrade all of the drives in a library, you can preserve your existing policy
definitions to migrate and expire existing data, while using the new drives to store
new data.

About this task

Define a new DISK storage pool and set it up to migrate its data to a storage pool
created for the new drives. Then update your existing management-class
definitions to begin storing data in the new DISK storage pool.

Chapter 7. Managing removable media operations 163

 Cleaning drives
You can use the server to manage tape-drive cleaning. The server can control how
tape drives in SCSI libraries are cleaned and offers partial support for cleaning
tape drives in manual libraries.

About this task

Task Required privilege class
Clean drives System or unrestricted storage

For automated libraries, you can automate cleaning by specifying the frequency of
cleaning operations and checking a cleaner cartridge into the library volume
inventory. Tivoli Storage Manager mounts the cleaner cartridge as specified. For
manual libraries, Tivoli Storage Manager issues a mount request for the cleaner
cartridge. There are special considerations if you plan to use server-controlled
drive cleaning with a SCSI library that provides automatic drive cleaning support
in its device hardware.

Tip: Use library-drive cleaning for automated tape libraries that use this function.

Device manufacturers that include library-cleaning functions recommend its use to

prevent premature wear on the read/write heads of the drives.

Drives and libraries from manufacturers differ in how they manage cleaner
cartridges, and how they report the presence of a cleaner cartridge in a drive. The
device driver might not be able to open a drive that contains a cleaner cartridge.
Sense codes and error codes that are issued by devices for drive cleaning vary.
Library-drive cleaning is usually not known to applications, therefore, Tivoli
Storage Manager might not always detect the cleaner cartridges in drives and
might not be able to determine when cleaning begins.

Some devices require a small amount of idle time between mount requests to start
drive cleaning. However, Tivoli Storage Manager tries to minimize the idle time for
a drive. The result might be to prevent the library drive cleaning from functioning
effectively. If this happens, use Tivoli Storage Manager to control drive cleaning.
You can set the frequency to match the cleaning recommendations from the
manufacturer.

Drive-cleaning methods
Cleaning of drives can be enabled from your drive or from Tivoli Storage Manager.

You can choose to use the library-drive cleaning method or the Tivoli Storage
Manager drive cleaning method, but not both. Some SCSI libraries provide
automatic drive cleaning.

Over time, the read heads on tapes can get dirty, which can cause read and write
failures. You can select a drive-cleaning method to keep the tapes in a library
clean. The library-drive cleaning method provides several advantages for
automated tape libraries that use this function:
v Lowers the burden on the Tivoli Storage Manager administrator to physically
manage cleaning the cartridges.
v Improves cleaning cartridge usage rates. Most tape libraries track the number of
times that drives can be cleaned based on the hardware indicators. Tivoli Storage
Manager uses a raw count.

164 IBM Tivoli Storage Manager for Windows: Administrator's Guide

v Reduces unnecessary cleaning frequency. Modern tape drives do not need to be
cleaned at fixed intervals and can detect and request when cleaning is required.

If you have Tivoli Storage Manager control-drive cleaning, disable the library
drive-cleaning function to prevent problems. If the library drive-cleaning function
is enabled, some devices automatically move any cleaner cartridge that is found in
the library to slots in the library that are dedicated for cleaner cartridges. You
cannot check a cleaner cartridge into the Tivoli Storage Manager library inventory
until you disable the library drive-cleaning function.

Cleaning drives in an automated library

When you configure a server-controlled drive cleaning in an automated library,
you can specify how often you want the drives that are cleaned.

About this task

Complete the following steps to configure server-controlled drive cleaning in an

automated library:

Procedure
1. Define or update the drives in a library, by using the CLEANFREQUENCY parameter
in the DEFINE DRIVE and UPDATE DRIVE command.
The CLEANFREQUENCY parameter sets how often you want the drive to be
cleaned. Consult the documentation that accompanies the drives for
recommendations on cleaning frequency. For example, to have DRIVE1 cleaned
after 100 GB are processed on the drive, issue the following command:
update drive autolib1 drive1 cleanfrequency=100
Consult the drive manufacturer documentation for cleaning recommendations.
If the documentation gives recommendations for cleaning frequency in terms of
hours of use, convert to a gigabytes value by completing the following steps:
a. Use the bytes-per-second rating for the drive to determine a
gigabytes-per-hour value.
b. Multiply the gigabytes-per-hour value by the recommended hours of use
between cleanings.
c. Use the result as the cleaning frequency value.

Restrictions:
a. For IBM 3590, and 3592 drives, specify a value for the CLEANFREQUENCY
parameter rather than specify ASNEEDED. Using the cleaning frequency that is
recommended by the product documentation will not overclean the drives.
b. The CLEANFREQUENCY=ASNEEDED parameter value does not work for all tape
drives. To determine whether a drive supports this function, see the
website: [Link]
IBM_TSM_Supported_Devices_for_AIXHPSUNWIN.html. Click the drive to
view detailed information. If ASNEEDED is not supported, you can use the
gigabytes value for automatic cleaning.
2. Check a cleaner cartridge into the library volume inventory with the CHECKIN
LIBVOLUME command.
checkin libvolume autolib1 cleanv status=cleaner cleanings=10 checklabel=no

After the cleaner cartridge is checked in, the server mounts the cleaner
cartridge in a drive when the drive needs cleaning. The server uses that cleaner

Chapter 7. Managing removable media operations 165

 cartridge for the number of cleanings specified. See “Checking in cleaner
volumes” and “Operations with cleaner cartridges in a library” on page 167 for
more information.

Checking in cleaner volumes:

To allow server to control drive cleaning without operator intervention, you must
check a cleaner cartridge into the automated library's volume inventory.

About this task

It is recommended that you check in cleaner cartridges one at a time and do not
use the search function of checkin for a cleaner cartridge.

Attention: When checking in a cleaner cartridge to a library, ensure that it is

correctly identified to the server as a cleaner cartridge. Also use caution when a
cleaner cartridge is already checked in and you are checking in data cartridges.
Ensure that cleaner cartridges are in their correct home slots, or errors and delays
can result.

When checking in data cartridges with SEARCH=YES, ensure that a cleaner

cartridge is not in a slot that will be detected by the search process. Errors and
delays of 15 minutes or more can result from a cleaner cartridge being improperly
moved or placed. For best results, check in the data cartridges first when you use
the search function. Then check in the cleaner cartridge separately.

For example, if you need to check in both data cartridges and cleaner cartridges,
put the data cartridges in the library and check them in first. You can use the
search function of the CHECKIN LIBVOLUME command (or the LABEL
LIBVOLUME command if you are labeling and checking in volumes). Then check
in the cleaner cartridge to the library by using one of the following methods.
v Check in without using search:
checkin libvolume autolib1 cleanv status=cleaner cleanings=10
checklabel=no

The server then requests that the cartridge be placed in the entry/exit port, or
into a specific slot.
v Check in using search, but limit the search by using the VOLRANGE or
VOLLIST parameter:
checkin libvolume autolib1 status=cleaner cleanings=10
search=yes checklabel=barcode vollist=cleanv

The process scans the library by using the bar code reader, looking for the
CLEANV volume.

166 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Manual drive cleaning in an automated library:

If your library has limited capacity and you do not want to use a slot in your
library for a cleaner cartridge, the server can issue messages telling you that a
drive needs to be cleaned.

Set the cleaning frequency for the drives in the library. When a drive needs
cleaning based on the frequency setting, the server issues the message, ANR8914I.
For example:
ANR89141I Drive DRIVE1 in library AUTOLIB1 needs to be cleaned.

You can use that message as a cue to manually insert a cleaner cartridge into the
drive. However, the server cannot track whether the drive has been cleaned.

Operations with cleaner cartridges in a library:

Guidelines include monitoring cleaning messages and verifying that cleaner

cartridges are in the correct storage slots.

When a drive needs to be cleaned, the server runs the cleaning operation after
dismounting a data volume if a cleaner cartridge is checked in to the library. If the
cleaning operation fails or is canceled, or if no cleaner cartridge is available, then
the indication that the drive needs cleaning is lost. Monitor cleaning messages for
these problems to ensure that drives are cleaned as needed. If necessary, issue the
CLEAN DRIVE command to have the server try the cleaning again, or manually
load a cleaner cartridge into the drive.

The server uses a cleaner cartridge for the number of cleanings that you specify
when you check in the cleaner cartridge. If you check in two or more cleaner
cartridges, the server uses only one of the cartridges until the designated number
of cleanings for that cartridge has been reached. Then the server begins to use the
next cleaner cartridge. If you check in two or more cleaner cartridges and issue
two or more CLEAN DRIVE commands concurrently, the server uses multiple
cartridges at the same time and decrements the remaining cleanings on each
cartridge.

Visually verify that cleaner cartridges are in the correct storage slots before issuing
any of the following commands:
v AUDIT LIBRARY
v CHECKIN LIBVOLUME with SEARCH specified
v LABEL LIBVOLUME with SEARCH specified
To find the correct slot for a cleaner cartridge, issue the QUERY LIBVOLUME
command.

Chapter 7. Managing removable media operations 167

 Drive cleaning in a manual library
The server can issue messages telling you that a drive in a manual library needs to
be cleaned.

Cleaning a drive in a manual library is the same as setting up drive cleaning

without checking in a cleaner cartridge for an automated library. The server issues
the ANR8914I message when a drive needs cleaning. For example:
ANR89141I Drive DRIVE1 in library MANLIB1 needs to be cleaned.

Monitor the activity log or the server console for these messages and load a cleaner
cartridge into the drive as needed. The server cannot track whether the drive has
been cleaned.

Error checking for drive cleaning

Occasionally you might move some cartridges around within a library and put a
data cartridge where Tivoli Storage Manager shows that there is a cleaner
cartridge. Tivoli Storage Manager can recover from the error.

When a drive needs cleaning, the server loads what its database shows as a cleaner
cartridge into the drive. The drive then moves to a READY state, and Tivoli
Storage Manager detects that the cartridge is a data cartridge. The server then
performs the following steps:
1. The server attempts to read the internal tape label of the data cartridge.
2. The server ejects the cartridge from the drive and moves it back to the home
slot of the “cleaner” cartridge within the library. If the eject fails, the server
marks the drive offline and issues a message that the cartridge is still in the
drive.
3. The server checks out the “cleaner” cartridge to avoid selecting it for another
drive cleaning request. The “cleaner” cartridge remains in the library but no
longer appears in the Tivoli Storage Manager library inventory.
4. If the server was able to read the internal tape label, the server checks the
volume name against the current library inventory, storage pool volumes, and
the volume history file.
v If there is not a match, you probably checked in a data cartridge as a cleaner
cartridge by mistake. Now that the volume is checked out, you do not need
to do anything else.
v If there is a match, the server issues messages that manual intervention and a
library audit are required. Library audits can take considerable time, so you
should issue the command when sufficient time permits. See “Auditing
volume inventories in libraries” on page 137.

Deleting drives
You can delete a drive if it is not currently in use. If a drive has a volume
mounted, but the volume is currently idle, it can be dismounted.

About this task

Task Required Privilege Class
Delete drives System or unrestricted storage

To delete a drive definition, issue the DELETE DRIVE command.

168 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 Note: A drive cannot be deleted until the defined path to the drive has been
deleted. Also, a library cannot be deleted until all of the drives defined within it
are deleted.

For details about dismounting, see “Dismounting idle volumes” on page 145.

Managing paths
You can use Tivoli Storage Manager commands to query, update, and delete paths.

Obtaining information about paths

You can use the QUERY PATH command to obtain information about paths.

About this task

You can request either a standard or a detailed report. For example, to display
information about all paths, issue the following command:
query path

The following shows an example of the output from this command.

Source Name Source Type Destination Name Destination Type Online
----------- ----------- ---------------- ---------------- ------
NETAPP1 Data mover DRIVE1 Drive Yes
NETAPP1 Data mover NASLIB Library Yes
datamover2 Data mover drive4 Drive Yes

Updating paths
You can use the UPDATE PATH command to update the attributes of an existing
path definition.

About this task

The examples below show how you can use the UPDATE PATH commands for the
following path types:
v Library Paths
Update the path from SERVER1 to a SCSI library named SCSILIB:
update path server1 scsilib srctype=server desttype=library device=lb4.0.0.0
v Drive Paths
Update the path from SERVER1 to a SCSI library named SCSILIB:
update path nas1 scsilib srctype=datamover desttype=drive
library=naslib device=mt3.0.0.0

Chapter 7. Managing removable media operations 169

 Deleting paths
You can use the DELETE PATH command to delete an existing path definition.

About this task

Task Required Privilege Class
Delete paths System or unrestricted storage

A path cannot be deleted if the destination is currently in use.

To delete a path from a NAS data mover NAS1 to the library NASLIB:
delete path nas1 naslib srctype=datamover desttype=library

Attention: If you delete the path to a device or make the path offline, you disable
access to that device.

Managing data movers

You can use Tivoli Storage Manager commands to query, update, and delete data
movers.

Obtaining information about data movers

You can use the QUERY DATAMOVER command to obtain information about SCSI
and NAS data movers.

About this task

You can request either a standard or a detailed report. For example, to display a
standard report about all data movers, issue the following command:
query datamover *

The following shows an example of the output from this command.

Data Mover Name Type Online
------------- ---------- -------
NASMOVER1 NAS Yes
NASMOVER2 NAS No
DATAMOVER1 SCSI Yes

Updating data movers

You can use the UPDATE DATAMOVER command to update the attributes of a
data mover definition.

About this task

For example, to update the data mover for the node named NAS1 to change the IP
address, issue the following command:
update datamover nas1 hladdress=[Link]

170 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 Deleting data movers
You can use the DELETE DATAMOVER command to delete an existing datamover.

About this task

Before you can delete a data mover, you must delete all paths defined for the data
mover.

To delete a data mover named NAS1, issue the following command:

delete datamover nas1

Managing disks
You can query, update, and delete client-owned disks that reside in a storage area
network.

Obtaining information about disks

You can use the QUERY DISK command to obtain information about client-owned
disks that reside in a SAN environment.

About this task

You can request either a standard or a detailed report. For example, to display a
standard report about all defined disks, issue the following command:
query disk *

The following shows an example of the output from this command.

Node Disk Online
Name Name
------------ ------ ------
NODE1 Harddisk1 Yes
NODE2 Harddisk2 Yes

Updating disks
You can use the UPDATE DISK command to update the attributes of an existing
disk definition.

About this task

The example below shows how you can use the UPDATE DISK command to
change the world wide name, serial number, and status of a disk.

Update a disk named Harddisk1 owned by NODE1. The world wide name to
20020060450d00e2 and the serial number to 100047. Change the ONLINE status to
YES.
update disk node1 Harddisk1 wwn=20020060450d00e2 serial=100047 online=yes

Chapter 7. Managing removable media operations 171

 Deleting disks
You can use the DELETE DISK command to delete an existing disk definition.

About this task

All paths related to a disk must be deleted before the disk itself can be deleted.

Delete a disk named Harddisk1 that is owned by the node NODE1.

delete disk node1 Harddisk1

172 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Chapter 8. Tips for defining devices
Learn more details about defining the Tivoli Storage Manager storage objects.

About this task

Defining devices and paths

The following topics describe how to define libraries, drives, and paths to Tivoli
Storage Manager.

About this task

See “Managing libraries” on page 150 and “Managing drives” on page 152 for
information about displaying library and drive information, and updating and
deleting libraries and drives.

Defining libraries
Before you can use a drive, you must first define the library to which the drive
belongs.

About this task

Task Required Privilege Class
Define or update libraries System or unrestricted storage

For both manually mounted drives and drives in automated libraries, the library
must be defined before the drives can be used. For example, you have several
stand-alone tape drives. You can define a library named MANUALMOUNT for
these drives by using the following command:
define library manualmount libtype=manual

For all libraries other than manual libraries, you define the library and then define
a path from the server to the library. For example, if you have an IBM 3583 device,
you can define a library named ROBOTMOUNT using the following command:
define library robotmount libtype=scsi

Next, you use the DEFINE PATH command. In the path, you must specify the
DEVICE parameter. The DEVICE parameter is required and specifies the device
alias name by which the library's robotic mechanism is known.
define path server1 robotmount srctype=server desttype=library
device=lb3.0.0.0

For more information about paths, see “Defining paths” on page 176.

© Copyright IBM Corp. 1993, 2013 173

 Defining SCSI libraries on a SAN
For a library type of SCSI on a SAN, the server can track the library's serial
number. With the serial number, the server can confirm the identity of the device
when you define the path or when the server uses the device.

About this task

If you choose, you can specify the serial number when you define the library to
the server. For convenience, the default is to allow the server to obtain the serial
number from the library itself at the time that the path is defined.

If you specify the serial number, the server confirms that the serial number is
correct when you define the path to the library. When you define the path, you can
set AUTODETECT=YES to allow the server to correct the serial number if the
number that it detects does not match what you entered when you defined the
library.

Depending on the capabilities of the library, the server may not be able to
automatically detect the serial number. Not all devices are able to return a serial
number when asked for it by an application such as the server. In this case, the
server will not record a serial number for the device, and will not be able to
confirm the identity of the device when you define the path or when the server
uses the device. See “Impact of device changes on the SAN” on page 125.

Defining drives
To inform the server about a drive that can be used to access storage volumes,
issue the DEFINE DRIVE command, followed by the DEFINE PATH command.

About this task

When issuing the DEFINE DRIVE command, you must provide some or all of the
following information:
Library name
The name of the library in which the drive resides.
Drive name
The name assigned to the drive.
Serial number
The serial number of the drive. The serial number parameter applies only
to drives in SCSI libraries. With the serial number, the server can confirm
the identity of the device when you define the path or when the server
uses the device.
You can specify the serial number if you choose. The default is to allow the
server to obtain the serial number from the drive itself at the time that the
path is defined. If you specify the serial number, the server confirms that
the serial number is correct when you define the path to the drive. When
you define the path, you can set AUTODETECT=YES to allow the server to
correct the serial number if the number that it detects does not match what
you entered when you defined the drive.
Depending on the capabilities of the drive, the server may not be able to
automatically detect the serial number. In this case, the server will not
record a serial number for the device, and will not be able to confirm the
identity of the device when you define the path or when the server uses
the device.

174 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 Element address
The element address of the drive. The ELEMENT parameter applies only
to drives in SCSI libraries. The element address is a number that indicates
the physical location of a drive within an automated library. The server
needs the element address to connect the physical location of the drive to
the drive's SCSI address. You can allow the server to obtain the element
number from the drive itself at the time that the path is defined, or you
can specify the element number when you define the drive.
Depending on the capabilities of the library, the server may not be able to
automatically detect the element address. In this case you must supply the
element address when you define the drive, if the library has more than
one drive. Element numbers for many libraries are available at
[Link]
IBM_TSM_Supported_Devices_for_AIXHPSUNWIN.html.

For example, to define a drive that belongs to the manual library named MANLIB,
enter this command:
define drive manlib mandrive

Next, you define the path from the server to the drive, using the device name used
to access the drive:
define path server1 mandrive srctype=server desttype=drive library=manlib
device=mt3.0.0.0

For more information about paths, see:

“Defining paths” on page 176
“Impact of device changes on the SAN” on page 125

Defining data movers

Data movers are SAN-attached devices that, through a request from Tivoli Storage
Manager, transfer client data for backup, archiving or restore purposes. Data
movers are defined as unique objects to Tivoli Storage Manager.

When issuing the DEFINE DATAMOVER command, you must provide some or all
of the following information:
Data mover name
The name of the defined data mover.
Type The type of data mover (SCSI or NAS).
World wide name
The Fibre Channel world wide name for the data mover device.
Serial number
Specifies the serial number of the data mover.
High level address
The high level address is either the numerical IP address or the domain
name of a NAS file server.
Low level address
The low level address specifies the TCP port number used to access a NAS
file server.
User ID
The user ID specifies the ID for a user when initiating a Network Data
Management Protocol (NDMP) session with a NAS file server.

Chapter 8. Tips for defining devices 175

 Password
The password specifies the password associated with a user ID when
initiating an NDMP session with a NAS file server. Check with your NAS
file server vendor for user ID and password conventions.
Copy threads
The number of concurrent copy operations that the SCSI data mover can
support.
Online
The online parameter specifies whether the data mover is online.
Data format
The data format parameter specifies the data format used according to the
type of data mover device used.

For example, to define a NAS data mover named NAS1, enter the following:
define datamover nas1 type=nas hladdress=[Link] lladdress=10000
userid=root password=admin dataformat=netappdump

Defining paths
Before a device can be used, a path must be defined between the device and the
server or the device and the data mover responsible for outboard data movement.

About this task

This command must be used to define the following path relationships:

v Between a server and a drive or a library.
v Between a storage agent and a drive.
v Between a data mover and a drive, a disk, or a library.

When issuing the DEFINE PATH command, you must provide some or all of the
following information:
Source name
The name of the server, storage agent, or data mover that is the source for
the path.
Destination name
The assigned name of the device that is the destination for the path.
Source type
The type of source for the path. (A storage agent is considered a type of
server for this purpose.)
Destination type
The type of device that is the destination for the path.
Library name
The name of the library that a drive is defined to if the drive is the
destination of the path.
GENERICTAPE device class
If you plan to use a device that is not supported by the Tivoli Storage
Manager server and you want to use the GENERICTAPE device class,
specify GENERICTAPE=Yes when defining the path.
Device
The alias name of the device (or for an IBM 3494 library, the symbolic
name). This parameter is used when defining a path between a server or a

176 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 storage agent and a library, drive, or disk. This parameter should not be
used when defining a data mover as the source type, except when the data
mover is a NAS data mover. NAS data movers always require a device
parameter. For shared FILE drives, this value is always “FILE.”
Directory
The directory location or locations of the files used in the FILE device
class. The default is the current working directory of the server at the time
the command is issued. Windows registry information is used to determine
the default directory.
Automatic detection of serial number and element address
For devices on a SAN, you can specify whether the server should correct
the serial number or element address of a drive or library, if it was
incorrectly specified on the definition of the drive or library. The server
uses the device name to locate the device and compares the serial number
(and the element address for a drive) that it detects with that specified in
the definition of the device. The default is to not allow the correction.
LUN Logical Unit Number. An identifier used on a SCSI bus to distinguish
between devices with the same target ID. On a Fibre Channel bus it is used
to distinguish between devices with the same world wide name. If the
LUN of the device, as identified by the source of the path, differs from the
LUN in the base definition of the device, you must use the LUN as
identified by the source of the path. This parameter should not be used
when defining a server as the source type.
Initiator ID
The SCSI initiator ID that the source will use when accessing the
destination. The parameter should not be used when defining a server as
the source type.

For example, if you had a SCSI type library named AUTODLTLIB that had a
device name of lb3.0.0.0, and you wanted to define it to a server named ASTRO1,
you would issue the following command:
define path astro1 autodltlib srctype=server desttype=library
device=lb3.0.0.0

If you had a drive, DRIVE01, that resided in library AUTODLTLIB, and had a
device name of mt3.0.0.0, and you wanted to define it to server ASTRO1, you
would issue the following command:
define path astro1 drive01 srctype=server desttype=drive library=autodltlib
device=mt3.0.0.0

Defining device classes

A device class defines a device type for a set of volumes that can be created in a
storage pool. Device classes are also important for storing database backups and
for exporting and importing data.

About this task

Sequential-access device types include tape, and sequential-access disk. For

random access storage, Tivoli Storage Manager supports only the DISK device
class, which is defined by Tivoli Storage Manager.

To define a device class, use the DEFINE DEVCLASS command and specify the
DEVTYPE parameter. The DEVTYPE parameter assigns a device type to the device

Chapter 8. Tips for defining devices 177

 class. You can define multiple device classes for each device type. For example,
you might want to specify different attributes for different storage pools that use
the same type of tape drive. Variations might be required that are not specific to
the device, but rather to how you want to use the device (for example, mount
retention or mount limit). For all device types other than FILE or SERVER, you must
define libraries and drives to Tivoli Storage Manager before you define device
classes.

To update an existing device class definition, use the UPDATE DEVCLASS command.
You can also delete a device class and query a device class by using the DELETE
DEVCLASS and QUERY DEVCLASS commands.

Task Required privilege class

Define, update, or delete device classes System or unrestricted storage
Request information about device classes Any administrator

Remember:
v One device class can be associated with multiple storage pools, but each storage
pool is associated with only one device class.
v If you include the DEVCONFIG option in the [Link] file, the files that you
specify with that option are automatically updated with the results of the DEFINE
DEVCLASS, UPDATE DEVCLASS, and DELETE DEVCLASS commands.
v Tivoli Storage Manager allows SCSI libraries to include tape drives of more than
one device type. When you define the device class in this environment, you
must declare a value for the FORMAT parameter.

See the following topics:

Tasks
“Defining tape device classes” on page 180
“Defining 3592 device classes” on page 183
“Device classes for devices not supported by the Tivoli Storage Manager server” on page
186
“Defining device classes for removable media devices” on page 186
“Defining sequential-access disk (FILE) device classes” on page 187
“Defining LTO device classes” on page 191
“Defining SERVER device classes” on page 194
“Defining device classes for StorageTek VolSafe devices” on page 195
“Defining device classes for CENTERA devices” on page 197
“Obtaining information about device classes” on page 198
“How Tivoli Storage Manager fills volumes” on page 199

For details about commands and command parameters, see the Administrator's
Reference.

For the most up-to-date list of supported devices and valid device class formats,
see the Tivoli Storage Manager Supported Devices website:
[Link]
IBM_TSM_Supported_Devices_for_AIXHPSUNWIN.html

178 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 The examples in topics show how to perform tasks using the Tivoli Storage
Manager command-line interface. For information about the commands, see the
Administrator's Reference, or issue the HELP command from the command line of a
Tivoli Storage Manager administrative client.

Sequential-access device types

With Tivoli Storage Manager you can use tape devices, magnetic disk devices,
removable media devices, and virtual volumes.

The following tables list the valid devices, media types, and Tivoli Storage
Manager device types.

For details and updates, see the following website: [Link]

software/sysmgmt/products/support/
IBM_TSM_Supported_Devices_for_AIXHPSUNWIN.html
Table 12. Tape devices
Examples Media type Device type
IBM 3590, 3590E drives IBM 3590 cartridges 3590
IBM 3592 drives IBM 3592 cartridges 3592
IBM 7206-005 4 mm cartridges 4MM
IBM 7208-001 and 7208-011 8 mm cartridges 8MM
Oracle StorageTek SD-3, Tape cartridges ECARTRIDGE
9490, 9840, 9940, and T10000
drives
Tape drives that can be used Tape cartridges GENERICTAPE
by the operating system
device drivers
IBM 3580 LTO Ultrium cartridges LTO
Tape drives that can be used Unknown NAS
by the NAS file server for
backups
Oracle StorageTek 9840 Write-once read-many VOLSAFE
drives (WORM) tape cartridges

Table 13. Magnetic disk devices

Examples Media type Device type
Sequential-access disk File system or storage FILE
volumes
EMC Centera File system or storage CENTERA
volumes

Table 14. Removable media (file system) devices

Examples Media type Device Type
Removable media devices Iomega Zip or Jaz, DVD, or REMOVABLEFILE
that are attached as local, CD media
removable file systems

Chapter 8. Tips for defining devices 179

 Table 15. Virtual volumes
Examples Media type Device type
Tivoli Storage Manager Storage volumes or files that SERVER
target server are archived on another
Tivoli Storage Manager
server

Defining tape device classes

Device class definitions for tapes include parameters that allow you to control
storage operations.

Specifying the estimated capacity of tape volumes

Tivoli Storage Manager also uses estimated capacity to determine when to begin
reclamation storage pool volumes.

About this task

For tape device classes, the default values selected by the server depend on the
recording format that is used to write data to the volume. You can either accept the
default for a device type or specify a value.

To specify estimated capacity for tape volumes, use the ESTCAPACITY parameter
when you define the device class or update its definition.

For more information about how Tivoli Storage Manager uses the estimated
capacity value, see “How Tivoli Storage Manager fills volumes” on page 199.

Specifying recording formats for tape media

You can specify the recording format that is used by Tivoli Storage Manager when
writing data to tape media.

About this task

To specify a recording format, use the FORMAT parameter when you define the
device class or update its definition.

If all drives associated with that device class are identical, specify FORMAT=DRIVE.
The server selects the highest format that is supported by the drive on which a
volume is mounted.

If some drives associated with the device class support a higher density format
than others, specify a format that is compatible with all drives. If you specify
FORMAT=DRIVE, mount errors can occur. For example, suppose a device class uses
two incompatible devices such as an IBM 7208-2 and an IBM 7208-12. The server
might select the high-density recording format of 8500 for each of two new
volumes. Later, if the two volumes are to be mounted concurrently, one fails
because only one of the drives is capable of the high-density recording format.

If drives in a single SCSI library use different tape technologies (for example, DLT
and LTO Ultrium), specify a unique value for the FORMAT parameter in each device
class definition.

The recording format that Tivoli Storage Manager uses for a volume is selected
when the first piece of data is written to the volume. Updating the FORMAT

180 IBM Tivoli Storage Manager for Windows: Administrator's Guide

parameter does not affect media that already contain data until those media are
rewritten from the beginning. This process might happen after a volume is
reclaimed or deleted, or after all of the data on the volume expires.

Associating library objects with device classes

A library contains the drives that can be used to mount the volume. Only one
library can be associated with a given device class. However, multiple device
classes can reference the same library.

About this task

To associate a device class with a library, use the LIBRARY parameter when you
define a device class or update its definition.

Controlling media-mount operations for tape devices

Using device class definitions, you can control the number of mounted volumes,
the amount of time a volume remains mounted, and the amount of time that the
Tivoli Storage Manager server waits for a drive to become available.

Controlling the number of simultaneously mounted volumes:

When setting a mount limit for a device class, you need to consider the number of
storage devices connected to your system, whether you are using the
simultaneous-write function, whether you are associating multiple device classes
with a single library, and the number of processes that you want to run at the
same time.

About this task

When selecting a mount limit for a device class, consider the following issues:
v How many storage devices are connected to your system?
Do not specify a mount limit value that is greater than the number of associated
available drives in your installation. If the server tries to mount as many
volumes as specified by the mount limit and no drives are available for the
required volume, an error occurs and client sessions may be terminated. (This
does not apply when the DRIVES parameter is specified.)
v Are you using the simultaneous-write function to primary storage pools, copy
storage pools, and active-data pools?
Specify a mount limit value that provides a sufficient number of mount points to
support writing data simultaneously to the primary storage pool and all
associated copy storage pools and active-data pools.
v Are you associating multiple device classes with a single library?
A device class associated with a library can use any drive in the library that is
compatible with the device class' device type. Because you can associate more
than one device class with a library, a single drive in the library can be used by
more than one device class. However, Tivoli Storage Manager does not manage
how a drive is shared among multiple device classes.
v How many Tivoli Storage Manager processes do you want to run at the same
time, using devices in this device class?
Tivoli Storage Manager automatically cancels some processes to run other,
higher priority processes. If the server is using all available drives in a device
class to complete higher priority processes, lower priority processes must wait
until a drive becomes available. For example, Tivoli Storage Manager cancels the
process for a client backing up directly to tape if the drive being used is needed

Chapter 8. Tips for defining devices 181

 for a server migration or tape reclamation process. Tivoli Storage Manager
cancels a tape reclamation process if the drive being used is needed for a client
restore operation. For additional information, see “Preempting operations” on
page 611.
If processes are often canceled by other processes, consider whether you can
make more drives available for Tivoli Storage Manager use. Otherwise, review
your scheduling of operations to reduce the contention for drives.
This consideration also applies to the simultaneous-write function. You must
have enough drives available to allow for a successful simultaneous-write
operation.

Best Practice: If the library associated with this device class is EXTERNAL type,
explicitly specify the mount limit instead of using MOUNTLIMIT=DRIVES.

To specify the maximum number of volumes that can be simultaneously mounted,

use the MOUNTLIMIT parameter when you define the device class or update its
definition.

Controlling the amount of time that a volume remains mounted:

You can control the amount of time that a mounted volume remains mounted after
its last I/O activity. If a volume is used frequently, you can improve performance
by setting a longer mount retention period to avoid unnecessary mount and
dismount operations.

About this task

If mount operations are being handled by manual, operator-assisted activities, you

might want to specify a long mount retention period. For example, if only one
operator supports your entire operation on a weekend, then define a long mount
retention period so that the operator is not being asked to mount volumes every
few minutes.

To control the amount of time a mounted volume remains mounted, use the
MOUNTRETENTION parameter when you define the device class or update its
definition. For example, if the mount retention value is 60, and a mounted volume
remains idle for 60 minutes, then the server dismounts the volume.

While Tivoli Storage Manager has a volume mounted, the drive is allocated to
Tivoli Storage Manager and cannot be used for anything else. If you need to free
the drive for other uses, you can cancel Tivoli Storage Manager operations that are
using the drive and then dismount the volume. For example, you can cancel server
migration or backup operations. For information on how to cancel processes and
dismount volumes, see:
v “Canceling server processes” on page 610
v “Dismounting idle volumes” on page 145

182 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Controlling the amount of time that the server waits for a drive:

You can specify the maximum amount of time, in minutes, that the Tivoli Storage
Manager server waits for a drive to become available for the current mount
request.

About this task

To control the wait time for a drive to become available for a mount request, use
the MOUNTWAIT parameter when you define or update a device class.

This parameter is not valid for EXTERNAL library types.

Write-once, read-many devices

The WORM parameter specifies whether the drive being defined is a write-once,
read-many WORM device. This parameter is not supported for all device classes.
You cannot change the value of the WORM parameter using the UPDATE
DEVCLASS command.

For an example that shows how to configure a VolSafe device using the WORM
parameter, see “Defining device classes for StorageTek VolSafe devices” on page
195

Defining 3592 device classes

Device class definitions for 3592 devices include parameters for faster
volume-access speeds and drive encryption. Particular methods are required to
prevent or minimize problems when mixing different generations of 3592 drives in
a library.

Mixing generations of 3592 media in a single library:

For optimal performance, do not mix generations of 3592 media in a single library.
Media problems can result when different drive generations are mixed. For
example, Tivoli Storage Manager might not be able to read a volume's label.

About this task

The following table shows read-and-write interoperability for drive generations.

Generation 1 Generation 2 Generation 3 Generation 4

Drives format format format format
Generation 1 Read and write n/a n/a n/a
Generation 2 Read and write Read and write n/a n/a
Generation 3 Read only Read and write Read and write n/a
Generation 4 n/a Read only Read and write Read and write

If you must mix generations of drives in a library, use one of the methods in the
following table to prevent or minimize the potential for problems.

Important: Mixed drive generations is not the optimal configuration. If possible,

limit the drives in a library to a single generation.

Chapter 8. Tips for defining devices 183

 Mixing generations of drives
(349X, ACSLS, SCSI libraries) If your library contains two drive generations, force all drives
to use the format of the earlier generation. For example, if your library contains generation
3 and generation 4 drives, force all the generation 4 drives to use the generation 3 format.
The result is that all the drives can read and write to all the media.
Remember: If you force a drive generation to write in the format of an earlier drive
generation, both drive generations can verify labels and read media written in the format
of the earlier drive generation. For example, if you library contains generation 2 and
generation 3 drives, both drive generations can verify labels and read media written in the
generation 2 format. However, this configuration does not allow the generation 3 drives to
read or write in their optimal format.

If your library contains three drive generations, the latest drive generation in your library
can only read media from the earliest format, but cannot write with it. For example, if your
library contains generation 2, generation 3, and generation 4 drives, the generation 4 drives
can only read the generation 2 format. In this configuration, mark all media previously
written in generation 2 format to read-only.
(349X and ACSLS libraries only) Logically partition the generations without partitioning the
hardware. Define two or three new library objects for each drive generation that the
physical library contains. For example, if you have a physical library with 3592-2 drives
and 3592-3 drives, define two new library objects.

Specify a path with the same special file name for each new library object. In addition, for
349X libraries, specify disjoint scratch categories (including the WORMSCRATCH category,
if applicable) for each library object. Specify a new device class and a new storage pool
that points to each new library object.
(SCSI libraries only) Define a new storage pool and device class for the latest drive
generation. For example, suppose you have a storage pool and device class for 3592-2. The
storage pool will contain all the media written in generation 2 format. Suppose that the
value of the FORMAT parameter in the device class definition is set to 3952-2 (not DRIVE).
You add generation 3 drives to the library. Complete the following steps:
1. In the new device-class definition for the generation 3 drives, set the value of the
FORMAT parameter to 3592-3 or 3592-3C. Do not specify DRIVE.
2. In the definition of the storage pool associated with generation 2 drives, update the
MAXSCRATCH parameter to 0, for example:
update stgpool genpool2 maxscratch=0

This method allows both generations to use their optimal format and minimizes potential
media problems that can result from mixing generations. However, it does not resolve all
media issues. For example, competition for mount points and mount failures might result.
(To learn more about mount point competition in the context of LTO drives and media, see
“Defining LTO device classes” on page 191.) The following list describes media restrictions:
v CHECKIN LIBVOL: The issue resides with using the CHECKLABEL=YES option. If the label is
currently written in a generation 3 or later format, and you specify the
CHECKLABEL=YES option, drives of previous generations fail using this command. As
a best practice, use CHECKLABEL=BARCODE.
v LABEL LIBVOL: When the server tries to use drives of a previous generation to read the
label written in a generation 3 or later format, the LABEL LIBVOL command fails unless
OVERWRITE=YES is specified. Verify that the media being labeled with OVERWRITE=YES
does not have any active data.
v CHECKOUT LIBVOL: When Tivoli Storage Manager verifies the label (CHECKLABEL=YES),
as a generation 3 or later format, and read operations by drives of previous generations,
the command fails. As a best practice, use CHECKLABEL=NO.

184 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Controlling data-access speeds for 3592 volumes:

Tivoli Storage Manager lets you reduce media capacity to create volumes with
faster data-access speeds. The benefit is that can partition data into storage pools
that have volumes with faster data-access speeds.

About this task

To reduce media capacity, use the SCALECAPACITY parameter when you define
the device class or update its definition.

Specify a percentage value of 20, 90 or 100. A value of 20 percent provides the

fastest access time, and 100 percent provides the largest storage capacity. For
example, If you specify a scale capacity of 20 for a 3592 device class without
compression, a 3592 volume in that device class would store 20 percent of its full
capacity of 300 GB, or about 60 GB.

Scale capacity only takes effect when data is first written to a volume. Updates to
the device class for scale capacity do not affect volumes that already have data
written to them until the volume is returned to scratch status.

For information about setting up storage pool hierarchies, see “Setting up a storage
pool hierarchy” on page 261.

Encrypting data with drives that are 3592 generation 2 and later:

With Tivoli Storage Manager, you can use the following types of drive encryption
with drives that are 3592 generation 2 and later: Application, System, and Library.
These methods are defined through the hardware.

Enabling 3592 drive encryption:

The DRIVEENCRYPTION parameter specifies whether drive encryption is enabled

or can be enabled for drives that are 3592 generation 2 and later. Use this
parameter to ensure Tivoli Storage Manager compatibility with hardware
encryption settings for empty volumes.

About this task

v To use the Application method, in which Tivoli Storage Manager generates and
manages encryption keys, set the DRIVEENCRYPTION parameter to ON. This
permits the encryption of data for empty volumes. If the parameter is set to ON
and if the hardware is configured for another encryption method, backup
operations will fail.
v To use the Library or System methods of encryption, set the parameter to
ALLOW. This specifies that Tivoli Storage Manager is not the key manager for
drive encryption, but will allow the hardware to encrypt the volume's data
through one of the other methods. Specifying this parameter does not
automatically encrypt volumes. Data can only be encrypted by specifying the
ALLOW parameter and configuring the hardware to use one of these methods.

The following simplified example shows how to permit the encryption of data for
empty volumes in a storage pool, using Tivoli Storage Manager as the key
manager:

Chapter 8. Tips for defining devices 185

 Procedure
1. Define a library. For example:
define library 3584 libtype=SCSI
2. Define a device class, 3592_ENCRYPT, and specify the value ON for the
DRIVEENCRYPTION parameter. For example:
define devclass 3592_encrypt library=3584 devtype=3592 driveencryption=on
3. Define a storage pool. For example:
define stgpool 3592_encrypt_pool 3592_encrypt

Results

The DRIVEENCRYPTION parameter is optional. The default value is to allow the

Library or System methods of encryption.

For more information about using drive encryption, refer to “Encrypting data on
tape” on page 536.

Disabling 3592 drive encryption:

To disable any method of encryption on new volumes, set the

DRIVEENCRYPTION parameter to OFF. If the hardware is configured to encrypt
data through either the Library or System method and DRIVEENCRYPTION is set
to OFF, backup operations will fail.

Device classes for devices not supported by the Tivoli Storage

Manager server
To use a tape device that is not supported by the Tivoli Storage Manager server,
you must define a device class whose device type is GENERICTAPE.

For Windows systems, you must also define a drive path with GENERICTAPE=Yes
to use a tape device.

For a manual library with multiple drives of device type GENERICTAPE, ensure
that the device types and recording formats of the drives are compatible. Because
the devices are controlled by the operating system device driver, the Tivoli Storage
Manager server is not aware of the following:
v The actual type of device: 4 mm, 8 mm, digital linear tape, and so forth. For
example, if you have a 4 mm device and an 8 mm device, you must define
separate manual libraries for each device.
v The actual cartridge recording format. For example, if you have a manual library
defined with two device classes of GENERICTAPE, ensure the recording formats
are the same for both drives.

Defining device classes for removable media devices

To access volumes that belong to this device class, the server requests that the
removable media be mounted in drives. The server then opens a file on the media
and reads or writes the file data.

About this task

Removable file devices include:

Iomega Zip drives, Iomega Jaz drives, CD drives, and DVD drives

186 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 To define a device class for removable media, use the
DEVTYPE=REMOVABLEFILE parameter in the device class definition.

Tivoli Storage Manager REMOVABLEFILE device class supports only single-sided

media. Therefore, if a data cartridge that is associated with a REMOVABLEFILE
device class has two sides, the Tivoli Storage Manager server treats each side as a
separate Tivoli Storage Manager volume.

When using CD-ROM media for the REMOVABLEFILE device type, the library
type must be specified as MANUAL. Access this media through a drive letter, for
example, E:.

For more information, see:

“Configuring removable media devices” on page 89

Defining sequential-access disk (FILE) device classes

FILE device classes are used for storing data on disk in simulated storage volumes.
The storage volumes are actually files. Data is written sequentially into the file
system of the server machine. Because each volume in a FILE device class is
actually a file, a volume name must be a fully qualified file name.

About this task

To define a FILE device class, use the DEVTYPE=FILE parameter in the device
class definition.

Do not use raw partitions with a device class type of FILE.

Concurrent access to FILE volumes

Concurrent access improves restore performance by allowing two or more clients
to access the same volume at the same time.

The Tivoli Storage Manager server allows for multiple client sessions (archive,
retrieve, backup, and restore) or server processes. For example, storage pool
backup, to concurrently read a volume in a storage pool that is associated with a
FILE-type device class. In addition, one client session or one server process can
write to the volume while it is being read.

The following server processes are allowed shared read access to FILE volumes:
v BACKUP DB
v BACKUP STGPOOL
v COPY ACTIVEDATA
v EXPORT/IMPORT NODE
v EXPORT/IMPORT SERVER
v GENERATE BACKUPSET
v RESTORE STGPOOL
v RESTORE VOLUME

The following server processes are not allowed shared read access to FILE
volumes:
v AUDIT VOLUME
v DELETE VOLUME
v MIGRATION
Chapter 8. Tips for defining devices 187
 v MOVE DATA
v MOVE NODEDATA
v RECLAMATION

Mitigating performance degradation when backing up or

archiving to FILE volumes
The minimum I/O to a volume associated with a FILE device class is 256 KB,
regardless how much data is being written to the volume. For example, if you are
backing up one 500-byte object, it takes 256 KB of I/O to store it on the volume.
The size of the I/O for a volume associated with a FILE device class has the
greatest impact when backing up or archiving a large number of small objects, for
example, small files or small directories.

About this task

To reduce the potential for performance degradation, increase the size of

aggregates created by the server. (An aggregate is an object that contains multiple
logical files that are backed up or archived from a client in a single transaction.) To
increase the size of aggregates, do one of the following
v Increase the value of the TXNGROUPMAX option in the server options file
([Link]).
v Increase the value of the TXNGROUPMAX parameter on the REGISTER NODE
or UPDATE NODE server commands.

In addition to increasing the TXNGROUPMAX value, you might also need to

increase the values for the following options:
v The client option TXNBYTELIMIT in the client options file ([Link])
v The server options MOVEBATCHSIZE and MOVESIZETHRESH
For details about the client option TXNBYTELIMIT, refer to the Backup-Archive
Clients Installation and User's Guide. For details about server commands and
options, refer to the Administrator's Reference.

Specifying directories in FILE device-class definitions

The directory name in a FILE device-class definition identifies the location where
the server places the files that represent storage volumes for the device class. When
processing the DEFINE DEVCLASS command, the server expands the specified
directory name into its fully qualified form, starting from the root directory.

About this task

You can specify one or more directories as the location of the files used in the FILE
device class. The default is the current working directory of the server at the time
the command is issued.

Attention: Do not specify multiple directories from the same file system. Doing
so can cause incorrect space calculations. For example, if the directories /usr/dir1
and /usr/dir2 are in the same file system, the space check, which does a
preliminary evaluation of available space during store operations, will count each
directory as a separate file system. If space calculations are incorrect, the server
could commit to a FILE storage pool, but not be able to obtain space, causing the
operation to fail. If the space check is accurate, the server can skip the FILE pool in
the storage hierarchy and use the next storage pool if one is available.

188 IBM Tivoli Storage Manager for Windows: Administrator's Guide

If the server needs to allocate a scratch volume, it creates a new file in the
specified directory or directories. (The server can choose any of the directories in
which to create new scratch volumes.) To optimize performance, ensure that
multiple directories correspond to separate physical volumes.

The following table lists the file name extension created by the server for scratch
volumes depending on the type of data that is stored.

For scratch volumes used to store this data: The file extension is:
Client data .BFS
Export .EXP
Database backup .DBV

Avoiding data-integrity problems when using disk subsystems and file

systems:

Tivoli Storage Manager supports the use of remote file systems or drives for
reading and writing storage pool data, database backups, and other data
operations. Disk subsystems and file systems must not report successful write
operations when they can fail after a successful write report to Tivoli Storage
Manager.

About this task

A write failure after a successful notification constitutes a data-integrity problem

because the data that was reported as successfully written is unavailable for
retrieval. In this situation, all data subsequently written is also at risk due to
positioning mismatches within the target file. To avoid these problems, ensure that
disk subsystems and file systems, whatever implementation you use, are always
able to return data when the data is requested.

For important disk-related information, see “Requirements for disk systems” on

page 63.

Giving storage agents access to FILE volumes:

You must ensure that storage agents can access newly created FILE volumes. To
access FILE volumes, storage agents replace names from the directory list in the
device class definition with the names in the directory list for the associated path
definition.

About this task

The following example illustrates the importance of matching device classes and
paths to ensure that storage agents can access newly created FILE volumes.
Suppose you want to use these three directories for a FILE library:
c:\server
d:\server
e:\server

Procedure
1. Use the following command to set up a FILE library named CLASSA with one
drive named CLASSA1 on SERVER1:

Chapter 8. Tips for defining devices 189

 define devclass classa devtype=file
directory="c:\server,d:\server,e:\server"
shared=yes mountlimit=1
2. You want the storage agent STA1 to be able to use the FILE library, so you
define the following path for storage agent STA1:
define path server1 sta1 srctype=server desttype=drive device=file
directory="\\[Link]\c\server,\\[Link]\d\server,
\\[Link]\e\server" library=classa
In this scenario, the storage agent, STA1, will replace the directory name
c:\server with the directory name \\[Link]\c\server to access FILE
volumes that are in the c:\server directory on the server.

Results

File volume c:\server\[Link] is created by SERVER1. If you later change the

first directory for the device class with the following command:
update devclass classa directory="c:\otherdir,d:\server,e:\server"

SERVER1 will still be able to access file volume c:\server\[Link], but the
storage agent STA1 will not be able to access it because a matching directory name
in the PATH directory list no longer exists. If a directory name is not available in
the directory list associated with the device class, the storage agent can lose access
to a FILE volume in that directory. Although the volume will still be accessible
from the Tivoli Storage Manager server for reading, failure of the storage agent to
access the FILE volume can cause operations to be retried on a LAN-only path or
to fail.

Controlling the size of FILE volumes

You can specify a maximum capacity value that controls the size of volumes (that
is, files) associated with a FILE device class.

About this task

To restrict the size of volumes, use the MAXCAPACITY parameter when you
define a device class or update its definition. When the server detects that a
volume has reached a size equal to the maximum capacity, it treats the volume as
full and stores any new data on a different volume.

Controlling the number of concurrently open FILE volumes

Tivoli Storage Manager lets you restrict the number of mount points (volumes or
files) that can be concurrently opened for access by server storage and retrieval
operations. Attempts to access more volumes than the number indicated causes the
requester to wait.

About this task

When selecting a mount limit for this device class, consider how many Tivoli
Storage Manager processes you want to run at the same time.

Tivoli Storage Manager automatically cancels some processes to run other, higher
priority processes. If the server is using all available mount points in a device class
to complete higher priority processes, lower priority processes must wait until a
mount point becomes available. For example, Tivoli Storage Manager cancels the
process for a client backup if the mount point being used is needed for a server
migration or reclamation process. Tivoli Storage Manager cancels a reclamation

190 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 process if the mount point being used is needed for a client restore operation. For
additional information, see “Preempting operations” on page 611.

If processes are often canceled by other processes, consider whether you can make
more mount points available for Tivoli Storage Manager use. Otherwise, review
your scheduling of operations to reduce the contention for resources.

To specify the number of concurrently opened mount points, use the

MOUNTLIMIT parameter when you define the device class or update its
definition.

Defining LTO device classes

Special consideration is required to prevent or minimize problems when mixing
different generations of LTO drives and media in a single library. LTO drive
encryption might also be a consideration.

Mixing LTO drives and media in a library

When mixing different generations of LTO drives and media, you need to consider
the read-write capabilities of each generation. As a best practice, configure a
different device class for each generation of media.

About this task

If you are considering mixing different generations of LTO media and drives, be
aware of the following restrictions:
Table 16. Read - write capabilities for different generations of LTO drives
Generation 1 Generation 2 Generation 3 Generation 4 Generation 5
Drives media media media media media
Generation 1 Read and n/a n/a n/a n/a
write
Generation 2 Read and Read and n/a n/a n/a
write write
Generation 3 Read only Read and Read and n/a n/a
write write
Generation 4 n/a Read only Read and Read and Read and
write write write
Generation 5 n/a n/a Read only Read and Read and
write write

If you are mixing different types of drives and media, configure different device
classes: one for each type of media. To specify the exact media type, use the
FORMAT parameter in each of the device class definitions. (Do not specify
FORMAT=DRIVE). For example, if you are mixing Ultrium Generation 1 and
Ultrium Generation 2 drives, specify FORMAT=ULTRIUMC (or ULTRIUM) for the
Ultrium Generation 1 device class, and FORMAT=ULTRIUM2C (or ULTRIUM2) for
the Ultrium Generation 2 device class.

Both device classes can point to the same library in which there can be Ultrium
Generation 1 and Ultrium Generation 2 drives. The drives will be shared between
the two storage pools. One storage pool will use the first device class and Ultrium
Generation 1 media exclusively. The other storage pool will use the second device
class and Ultrium Generation 2 media exclusively. Because the two storage pools

Chapter 8. Tips for defining devices 191

 share a single library, Ultrium Generation 1 media can be mounted on Ultrium
Generation 2 drives as they become available during mount point processing.

Remember:
v If you are mixing Ultrium Generation 1 with Ultrium Generation 3 drives and
media in a single library, you must mark the Generation 1 media as read-only,
and all Generation 1 scratch volumes must be checked out.
v If you are mixing Ultrium Generation 2 with Ultrium Generation 4 or
Generation 5 drives and media in a single library, you must mark the Generation
2 media as read-only, and all Generation 2 scratch volumes must be checked out.

Mount limits in LTO mixed-media environments

In a mixed-media library, in which multiple device classes point to the same
library, compatible drives are shared between storage pools. You must pay special
attention to setting an appropriate value for the MOUNTLIMIT parameter in each
of the device classes. In a mixed media library containing Ultrium Generation 1
and Ultrium Generation 2 drives and media, for example, Ultrium Generation 1
media can get mounted in Ultrium Generation 2 drives.

Consider the example of a mixed library: that consists of the following drives and
media:
v Four LTO Ultrium Generation 1 drives and LTO Ultrium Generation 1 media
v Four LTO Ultrium Generation 2 drives and LTO Ultrium Generation 2 media

You created the following device classes:

v LTO Ultrium Generation 1 device class LTO1CLASS specifying
FORMAT=ULTRIUMC
v LTO Ultrium Generation 2 device class LTO2CLASS specifying
FORMAT=ULTRIUM2C

You also created the following storage pools:

v LTO Ultrium Generation 1 storage pool LTO1POOL based on device class
LTO1CLASS
v LTO Ultrium Generation 2 storage pool LTO2POOL based on device class
LTO2CLASS

The number of mount points available for use by each storage pool is specified in
the device class using the MOUNTLIMIT parameter. The MOUNTLIMIT parameter
in the LTO2CLASS device class should be set to 4 to match the number of available
drives that can mount only LTO2 media. The MOUNTLIMIT parameter in the
LTO1CLASS device class should be set to a value higher (5 or possibly 6) than the
number of available drives to adjust for the fact that Ultrium Generation 1 media
can be mounted in Ultrium Generation 2 drives. The optimum value for
MOUNTLIMIT will depend on workload and storage pool access patterns.

Monitor and adjust the MOUNTLIMIT setting to suit changing workloads. If the
MOUNTLIMIT for LTO1POOL is set too high, mount requests for the LTO2POOL
might be delayed or fail because the Ultrium Generation 2 drives have been used
to satisfy Ultrium Generation 1 mount requests. In the worst scenario, too much
competition for Ultrium Generation 2 drives might cause mounts for Generation 2
media to fail with the following message:
ANR8447E No drives are currently available in the library.

192 IBM Tivoli Storage Manager for Windows: Administrator's Guide

If the MOUNTLIMIT for LTO1POOL is not set high enough, mount requests that
could potentially be satisfied LTO Ultrium Generation 2 drives will be delayed.

Some restrictions apply when mixing Ultrium Generation 1 with Ultrium

Generation 2 or Generation 3 drives because of the way in which mount points are
allocated. For example, processes that require multiple mount points that include
both Ultrium Generation 1 and Ultrium Generation 2 volumes might try to reserve
Ultrium Generation 2 drives only, even when one mount can be satisfied by an
available Ultrium Generation 1 drive. Processes that behave in this manner include
the MOVE DATA and BACKUP STGPOOL commands. These processes will wait
until the needed number of mount points can be satisfied with Ultrium Generation
2 drives.

Encrypting data using LTO generation 4 tape drives

Tivoli Storage Manager supports the three types of drive encryption available with
LTO generation 4 drives: Application, System, and Library. These methods are
defined through the hardware.

About this task

For more information about using drive encryption, refer to “Encrypting data on
tape” on page 536.

Enabling LTO drive encryption:

The DRIVEENCRYPTION parameter specifies whether drive encryption is enabled

or can be enabled for IBM and HP LTO generation 4, Ultrium4, and Ultrium4C
formats. This parameter ensures Tivoli Storage Manager compatibility with
hardware encryption settings for empty volumes.

About this task

Tivoli Storage Manager supports the Application method of encryption with IBM
and HP LTO-4 drives. Only IBM LTO-4 supports the System and Library methods.
The Library method of encryption is supported only if your system hardware (for
example, IBM 3584) supports it.

Remember: You cannot use drive encryption with write-once, read-many (WORM)
media.

The Application method is defined through the hardware. To use the Application
method, in which Tivoli Storage Manager generates and manages encryption keys,
set the DRIVEENCRYPTION parameter to ON. This permits the encryption of data
for empty volumes. If the parameter is set to ON and the hardware is configured
for another encryption method, backup operations will fail.

The following simplified example shows the steps you would take to permit the
encryption of data for empty volumes in a storage pool:

Procedure
1. Define a library:
define library 3584 libtype=SCSI
2. Define a device class, LTO_ENCRYPT, and specify Tivoli Storage Manager as
the key manager:
define devclass lto_encrypt library=3584 devtype=lto driveencryption=on

Chapter 8. Tips for defining devices 193

 3. Define a storage pool:
define stgpool lto_encrypt_pool lto_encrypt

Disabling LTO drive encryption:

To disable encryption on new volumes, set the DRIVEENCRYPTION parameter to

OFF. The default value is ALLOW. Drive encryption for empty volumes is
permitted if another method of encryption is enabled.

Defining SERVER device classes

SERVER device classes let you create volumes for one Tivoli Storage Manager
server that exist as archived files in the storage hierarchy of another server, called a
target server. These virtual volumes have the characteristics of sequential-access
volumes such as tape.

About this task

To define a SERVER device class, use the DEFINE DEVCLASS command with the
DEVTYPE=SERVER parameter. For information about how to use a SERVER device
class, see “Using virtual volumes to store data on another server” on page 732.

Controlling the size of files created on a target server

You can specify a maximum capacity value that controls the size of files that are
created on the target server to store data for the source server.

About this task

To specify a file size, use the MAXCAPACITY parameter when you define the
device class or update its definition.

The storage pool volumes of this device type are explicitly set to full when the
volume is closed and dismounted.

Controlling the number of simultaneous sessions between

source and target servers
You can control the number of simultaneous sessions between the source server
and the target server. Any attempts to access more sessions than indicated by the
mount limit causes the requester to wait.

About this task

To control the number of simultaneous sessions, use the MOUNTLIMIT parameter

when you define the device class or update its definition.

When specifying a mount limit, consider your network load balancing and how
many Tivoli Storage Manager processes you want to run at the same time.

Tivoli Storage Manager automatically cancels some processes to run other, higher
priority processes. If the server is using all available sessions in a device class to
complete higher priority processes, lower priority processes must wait until a
session becomes available. For example, Tivoli Storage Manager cancels the process
for a client backup if a session is needed for a server migration or reclamation
process. Tivoli Storage Manager cancels a reclamation process if the session being
used is needed for a client restore operation.

194 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 When specifying a mount limit, also consider the resources available on the target
server when setting mount limits. Do not set a high mount limit value if the target
cannot move enough data or access enough data to satisfy all of the requests.

If processes are often canceled by other processes, consider whether you can make
more sessions available for Tivoli Storage Manager use. Otherwise, review your
scheduling of operations to reduce the contention for network resources.

Controlling the amount of time a SERVER volume remains

mounted
You can improve response time for SERVER media mounts by leaving previously
mounted volumes online.

About this task

To specify the amount of time, in minutes, to retain an idle sequential access

volume before dismounting it, use the MOUNTRETENTION parameter when you
define the device class or update its definition.

A value of 1 to 5 minutes is recommended.

Defining device classes for StorageTek VolSafe devices

StorageTek VolSafe brand Ultrium drives use media that cannot be overwritten. Do
not use this media for short-term backups of client files, the server database, or
export tapes.

About this task

There are two methods for using VolSafe media and drives: This technology uses
media that cannot be overwritten; therefore, do not use this media for short-term
backups of client files, the server database, or export tapes.
v Define a device class using the DEFINE DEVCLASS command and specify
DEVTYPE=VOLSAFE. You can use this device class with EXTERNAL, SCSI, and
ACSLS libraries. All drives in a library must be enabled for VolSafe use.
v Define a device class using the DEFINE DEVCLASS command, and specify
DEVTYPE=ECARTRIDGE and WORM=YES. For VolSafe devices, WORM=YES is
required and must be specified when the device class is defined. You cannot
update the WORM parameter using the UPDATE DEVCLASS command. You
cannot specify DRIVEENCRYPTION=ON if your drives are using WORM
media.

To enable the VolSafe function, consult your StorageTek hardware documentation.

Attempting to write to VolSafe media without a VolSafe-enabled drive results in
errors.

To configure a VolSafe device in a SCSI library using the DEVTYPE-ECARTRIDGE

parameter, enter the following series of commands. (The values you select for the
library variable, the drive variable, and so on might be different for your
environment.)

Procedure
1. Define a library:
define library volsafelib libtype=scsi
2. Define a drive:

Chapter 8. Tips for defining devices 195

 define drive volsafelib drive01
3. Define a path:
define path server01 drive01 srctype=server destype=drive device=mt4.0.0.1
library=volsafelib
4. Define a device class:
define devclass volsafeclass library=volsafelib devtype=ecartridge
format=drive worm=yes

Results

For more information about VolSafe media, see “Write-once, read-many tape
media” on page 133.

Enabling ECARTRIDGE drive encryption

The DRIVEENCRYPTION parameter specifies whether drive encryption is enabled
or can be enabled for DRIVE, T10000B, T10000B-C, T10000C, and T10000C-C
formats. This parameter ensures Tivoli Storage Manager compatibility with
hardware encryption settings for empty volumes.

About this task

Tivoli Storage Manager supports the Application method of encryption with Oracle
StorageTek T10000B or T10000C drives. The Library method of encryption is
supported only if your system hardware supports it.

Remember: You cannot use drive encryption with write-once, read-many (WORM)
media or VolSafe media.

The Application method, in which Tivoli Storage Manager generates and manages
encryption keys, is defined through the hardware. To use the Application method,
set the DRIVEENCRYPTION parameter to ON. This setting permits the encryption
of data for empty volumes. If the parameter is set to ON and the hardware is
configured for another encryption method, backup operations fail.

The following simplified example shows the steps you would take to permit data
encryption for empty volumes in a storage pool:

Procedure
1. Define a library:
define library sl3000 libtype=scsi
2. Define a device class, ECART_ENCRYPT, and specify Tivoli Storage Manager
as the key manager:
define devclass ecart_encrypt library=sl3000
devtype=ecartridge driveencryption=on
3. Define a storage pool:
define stgpool ecart_encrypt_pool ecart_encrypt
Related concepts:
“Choosing an encryption method” on page 537

196 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 Disabling ECARTRIDGE drive encryption
To disable encryption on new volumes, set the DRIVEENCRYPTION parameter to
OFF. The default value is ALLOW. You can use drive encryption for empty
volumes if another method of encryption is enabled.

Defining device classes for CENTERA devices

To use a Centera device, you must define a device class whose device type is
CENTERA.

Concurrent access to Centera volumes

Concurrent access improves performance while restoring or retrieving data because
two or more clients can access the same volume at the same time.

Multiple client retrieve sessions, restore sessions, or server processes can read a
volume concurrently in a storage pool that is associated with the CENTERA device
type. In addition, one client session or one server process can write to the volume
while it is being read.

The following server processes can share read access to Centera volumes:
v EXPORT NODE
v EXPORT SERVER
v GENERATE BACKUPSET

The following server processes cannot share read access to Centera volumes:
v AUDIT VOLUME
v DELETE VOLUME

Server operations not supported by Centera

Centera storage devices do not support some Tivoli Storage Manager server
operations.

The following server operations are not supported:

v Data-movement operations:
– Moving node data into or out of a Centera storage pool.
– Migrating data into or out of a Centera storage pool.
– Reclaiming a Centera storage pool.
– LAN-free data movement or Network Data Management Protocol (NDMP)
operations. Centera storage pools cannot be the target or source of data for
either of these operations.
v Backup operations:
– Backing up a Centera storage pool.
– Using a Centera device class to back up a database.
– Backing up a storage pool to a Centera storage pool.
v Restore operations:
– Restoring data from a copy storage pool or an active-data pool to a Centera
storage pool.
– Restoring volumes in a Centera storage pool.
v Other:
– Exporting data to a Centera device class or importing data from a Centera
device class. However, files stored in Centera storage pools can be exported
and files being imported can be stored on Centera.

Chapter 8. Tips for defining devices 197

 – Using a Centera device class for creating backup sets; however, files stored in
Centera storage pools can be sent to backup sets.
– Defining Centera volumes.
– Using a Centera device class as the target of volume history, device
configuration, trace logs, error logs, or query output files.
– Using a Centera device class as the target for a virtual volume operation.
– Data deduplication.
– Copying active versions of backup data either to or from a Centera storage
pool.

Controlling the number of concurrently open mount points for

Centera devices
You can control the number of mount points that can be opened concurrently for
access by server storage and retrieval operations. Any attempts to access more
mount points than indicated by the mount limit causes the requester to wait.

About this task

When you select a mount limit for this device class, consider how many Tivoli
Storage Manager processes you want to run at the same time.

Tivoli Storage Manager automatically cancels some processes to run other, higher
priority processes. If the server is using all available mount points in a device class
to complete higher priority processes, lower priority processes must wait until a
mount point becomes available. For example, the Tivoli Storage Manager server is
currently performing a client backup request to an output volume and another
request from another client to restore data from the same volume. The backup
request is preempted and the volume is released for use by the restore request. For
more information, see “Preempting operations” on page 611.

To control the number of mount points concurrently open for Centera devices, use
the MOUNTLIMIT parameter when you define the device class or update its
definition.

Obtaining information about device classes

You can choose to view a standard or detailed report for a device class.

About this task

Task Required Privilege Class
Request information about device classes Any administrator

To display a standard report on device classes, enter:

query devclass

Figure 11 on page 199 provides an example of command output.

198 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 Device Device Storage Device Format Est/Max Mount
Class Access Pool Type Capacity Limit
Name Strategy Count (MB)
--------- ---------- ------- ------- ------ -------- -----
DISK Random 9
TAPE8MM Sequential 1 8MM 8200 2
FILE Sequential 1 FILE DRIVE 5,000.0 1
GEN1 Sequential 2 LTO ULTRIUM DRIVES

Figure 11. Example of a standard device class report

To display a detailed report on the GEN1 device class, enter:

query devclass gen1 format=detailed

Figure 12 provides an example of command output.

Device Class Name: GEN1

Device Access Strategy: Sequential
Storage Pool Count: 2
Device Type: LTO
Format: ULTRIUM
Est/Max Capacity (MB):
Mount Limit: DRIVES
Mount Wait (min): 60
Mount Retention (min): 60
Label Prefix: ADSM
Drive Letter:
Library: GEN2LIB
Directory:
Server Name:
Retry Period:
Retry Interval:
TwoSided:
Shared:
High-level Address:
Minimum Capacity:
WORM:
Scaled Capacity:
Last Update by (administrator): ADMIN
Last Update Date/Time: 01/23/03 [Link]

Figure 12. Example of a detailed device class report

How Tivoli Storage Manager fills volumes

The DEFINE DEVCLASS command has an optional ESTCAPACITY parameter that
indicates the estimated capacity for sequential volumes associated with the device
class. Tivoli Storage Manager uses the estimated capacity of volumes to determine
the estimated capacity of a storage pool, and the estimated percent utilized.

If the ESTCAPACITY parameter is not specified, Tivoli Storage Manager uses a

default value based on the recording format specified for the device class
(FORMAT=).

If you specify an estimated capacity that exceeds the actual capacity of the volume
in the device class, Tivoli Storage Manager updates the estimated capacity of the
volume when the volume becomes full. When Tivoli Storage Manager reaches the
end of the volume, it updates the capacity for the amount that is written to the
volume.

You can either accept the default estimated capacity for a given device class, or
explicitly specify an estimated capacity. An accurate estimated capacity value is not
required, but is useful. Tivoli Storage Manager uses the estimated capacity of

Chapter 8. Tips for defining devices 199

 volumes to determine the estimated capacity of a storage pool, and the estimated
percent utilized. You may want to change the estimated capacity if:
v The default estimated capacity is inaccurate because data compression is being
performed by the drives.
v You have volumes of nonstandard size.

Data compression
Client files can be compressed to decrease the amount of data sent over networks
and the space occupied by the data in Tivoli Storage Manager storage. With Tivoli
Storage Manager, files can be compressed by the Tivoli Storage Manager client
before the data is sent to the Tivoli Storage Manager server, or by the device where
the file is finally stored.

Use either client compression or device compression, but not both. The following
table summarizes the advantages and disadvantages of each type of compression.

Type of Compression Advantages Disadvantages

Tivoli Storage Manager client Reduced load on the network Higher CPU usage by the
compression client

Longer elapsed time for client

operations such as backup
Drive compression Amount of compression can Using drive compression on
be better than Tivoli Storage files that have already been
Manager client compression compressed by the Tivoli
on some drives Storage Manager client can
increase file size

Either type of compression can affect tape drive performance, because compression
affects data rate. When the rate of data going to a tape drive is slower than the
drive can write, the drive starts and stops while data is written, meaning relatively
poorer performance. When the rate of data is fast enough, the tape drive can reach
streaming mode, meaning better performance. If tape drive performance is more
important than the space savings that compression can mean, you may want to
perform timed test backups using different approaches to determine what is best
for your system.

Drive compression is specified with the FORMAT parameter for the drive's device
class, and the hardware device must be able to support the compression format.
For information about how to set up compression on the client, see “Node
compression considerations” on page 422 and “Registering nodes with the server”
on page 420.

Tape volume capacity and data compression

How Tivoli Storage Manager views the capacity of the volume where the data is
stored depends on whether files are compressed by the Tivoli Storage Manager
client or by the storage device.

It may wrongly appear that you are not getting the full use of the capacity of your
tapes, for the following reasons:
v A tape device manufacturer often reports the capacity of a tape based on an
assumption of compression by the device. If a client compresses a file before it is
sent, the device may not be able to compress it any further before storing it.

200 IBM Tivoli Storage Manager for Windows: Administrator's Guide

v Tivoli Storage Manager records the size of a file as it goes to a storage pool. If
the client compresses the file, Tivoli Storage Manager records this smaller size in
the database. If the drive compresses the file, Tivoli Storage Manager is not
aware of this compression.

Figure 13 compares what Tivoli Storage Manager sees as the amount of data stored
on tape when compression is done by the device and by the client. For this
example, the tape has a physical capacity of 1.2 GB. However, the manufacturer
reports the capacity of the tape as 2.4 GB by assuming the device compresses the
data by a factor of two.

Suppose a client backs up a 2.4 GB file:

v When the client does not compress the file, the server records the file size as 2.4
GB, the file is compressed by the drive to 1.2 GB, and the file fills up one tape.
v When the client compresses the file, the server records the file size as 1.2 GB, the
file cannot be compressed any further by the drive, and the file still fills one
tape.

In both cases, Tivoli Storage Manager considers the volume to be full. However,
Tivoli Storage Manager considers the capacity of the volume in the two cases to be
different: 2.4 GB when the drive compresses the file, and 1.2 GB when the client
compresses the file. Use the QUERY VOLUME command to see the capacity of
volumes from Tivoli Storage Manager's viewpoint. See “Monitoring the use of
storage pool volumes” on page 384.

Client Server Drive

Drive
compression
only

2.4 GB 2.4 GB
1.2 GB

Client Server Drive

Client
compression
only

2.4 GB
1.2 GB 1.2 GB

Figure 13. Comparing compression at the client and compression at the device

For how to set up compression on the client, see “Node compression

considerations” on page 422 and “Registering nodes with the server” on page 420.

Chapter 8. Tips for defining devices 201

202 IBM Tivoli Storage Manager for Windows: Administrator's Guide
Chapter 9. Using NDMP for operations with NAS file servers
You can plan, configure, and manage a backup environment that protects your
network-attached storage (NAS) file server by using NDMP (network data
management protocol). Tivoli Storage Manager Extended Edition includes support
for the use of NDMP to back up and recover NAS file servers.

Tasks:
“Configuring Tivoli Storage Manager for NDMP operations” on page 211
“Determining the location of NAS backup” on page 213
“Configuring a tape library for NDMP operations” on page 217
“Configuring Tivoli Storage Manager policy for NDMP operations” on page 212
“Registering NAS nodes with the Tivoli Storage Manager server” on page 223
“Defining a data mover for the NAS file server” on page 224
“Defining paths to libraries for NDMP operations” on page 227
“Defining paths to libraries for NDMP operations” on page 227
“Defining paths for NDMP operations” on page 224
“Labeling and checking tapes into the library” on page 228
“Scheduling NDMP operations” on page 228
“Defining virtual file spaces” on page 228
“Tape-to-tape copy to back up data” on page 229
“Tape-to-tape copy to move data” on page 229
“Backing up and restoring NAS file servers using NDMP” on page 230
“Backing up NDMP file server to Tivoli Storage Manager server backups” on page 232
“Managing tables of contents” on page 209
“NDMP operations management” on page 206
“Managing NAS file server nodes” on page 207
“Managing data movers used in NDMP operations” on page 208
“Storage pool management for NDMP operations” on page 209

NDMP requirements
You must meet certain requirements when you use NDMP (network data
management protocol) for operations with network-attached storage (NAS) file
servers.
Tivoli Storage Manager Extended Edition
Licensed program product that includes support for the use of NDMP.
NAS File Server
A NAS file server. The operating system on the file server must be
supported by Tivoli Storage Manager. Visit [Link]
support/entry/portal/Overview/Software/Tivoli/Tivoli_Storage_Manager
for a list of NAS file servers that are certified through the “Ready for IBM
Tivoli software.”

© Copyright IBM Corp. 1993, 2013 203

 Note: Vendors on the“Ready for IBM Tivoli software” list follow
guidelines to implement NDMP as specified by Tivoli Storage Manager. If
a file server is on the list, it has run tests to ensure that it is compatible
with Tivoli Storage Manager.
The combination of file server model and operating system must be
supported by the NAS file server. For more specifics, consult the product
information for the NAS file server.
Tape Libraries
This requirement is only necessary for a backup to a locally attached NAS
device. The Tivoli Storage Manager server supports three types of libraries
for operations that use NDMP. The libraries that are supported are SCSI,
ACSLS (automated cartridge system library software), and VTL (virtual
tape library). 349X tape libraries can also be used with certain NAS file
servers.
SCSI library
A SCSI library that is supported by the Tivoli Storage Manager
server. This type of library can be attached directly either to the
Tivoli Storage Manager server or to the NAS file server. When the
library is attached directly to the Tivoli Storage Manager server, the
Tivoli Storage Manager server controls the library operations by
passing the SCSI commands directly to the library. When the
library is attached directly to the NAS file server, the Tivoli Storage
Manager server controls the library by passing SCSI commands to
the library through the NAS file server.
ACSLS library
An ACSLS library can be directly connected only to the Tivoli
Storage Manager server. The Tivoli Storage Manager server
controls the library by passing the library request through TCP/IP
to the library control server.

Note: The Tivoli Storage Manager server does not include External
Library support for the ACSLS library when the library is used for
NDMP operations.
VTL library
A virtual tape library that is supported by the Tivoli Storage
Manager server. This type of library can be attached directly either
to the Tivoli Storage Manager server or to the NAS file server. A
virtual tape library is essentially the same as a SCSI library but is
enhanced for virtual tape library characteristics and allows for
better mount performance.

Note: If you are defining a VTL library, your environment must

not include any mixed-media. Paths must be defined between all
drives in the library and all defined servers, including storage
agents, that use the library. If either of these characteristics are not
true, the overall performance can degrade to the same levels as the
SCSI library type; especially during times of high stress.
349X library
A 349X library can be directly connected only to the Tivoli Storage
Manager server. The Tivoli Storage Manager server controls the
library by passing the library request through TCP/IP to the
library manager.

204 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 Library Sharing: The Tivoli Storage Manager server running NDMP
operations can be a library manager for either an ACSLS, SCSI, VTL, or
349X library, but cannot be a library client. The Tivoli Storage Manager
server can also be a library client, in a configuration where the NAS filer
sends data to a Tivoli Storage Manager server using TCP/IP rather than to
a tape library attached to the NAS filer. If the Tivoli Storage Manager
server that runs NDMP operations is a library manager, that server must
control the library directly and not by passing commands through the NAS
file server.
Tape Drives
One or more tape drives in the tape library. A tape drive is only necessary
for backup to a locally attached NAS device. The NAS file server must be
able to access the drives. A NAS device is not supported in a mixed device
library. The drives must be supported for tape backup operations by the
NAS file server and its operating system. For complete NDMP device
support, refer to the NAS file server product documentation.

Drive Sharing: The tape drives can be shared by the Tivoli Storage
Manager server and one or more NAS file servers. Also, when a SCSI,
VTL, or a 349X library is connected to the Tivoli Storage Manager server
and not to the NAS file server, the drives can be shared by one or more
NAS file servers and one or more Tivoli Storage Manager:
v Library clients
v Storage agents

Verify the compatibility of specific combinations of a NAS file server, tape devices,
and SAN-attached devices with the hardware manufacturers.

Attention: Tivoli Storage Manager supports NDMP Version 4 for all NDMP
operations. Tivoli Storage Manager continues to support all NDMP backup and
restore operations with a NAS device that runs NDMP version 3. The Tivoli
Storage Manager server negotiates the highest protocol level (either Version 3 or
Version 4) with the NDMP server when it establishes an NDMP connection. If you
experience any issues with Version 4, you might want to try Version 3.

Interfaces for NDMP operations

You can use several interfaces to perform NDMP (network data management
protocol) operations. You can schedule an NDMP operation using the BACKUP NODE
and RESTORE NODE commands, and scheduling the operation as an administrative
schedule.

Client Interfaces:
v Backup-archive command-line client (on a Windows, 64 bit AIX, or 64 bit Oracle
Solaris system)
v web client

Server Interfaces:
v Server console
v Command line on the administrative client

Tip: All examples in this chapter use server commands.

Chapter 9. Using NDMP for operations with NAS file servers 205
 The Tivoli Storage Manager web client interface, available with the backup-archive
client, displays the file systems of the network-attached storage (NAS) file server in
a graphical view. The client function is not required, but you can use the client
interfaces for NDMP operations. The client function is recommended for file-level
restore operations. See “File-level backup and restore for NDMP operations” on
page 233 for more information about file-level restore.

Tivoli Storage Manager prompts you for an administrator ID and password when
you perform NDMP functions using either of the client interfaces. See the
Backup-Archive Clients Installation and User's Guide for more information about
installing and activating client interfaces.

Attention: In order to use the Tivoli Storage Manager backup-archive client or

web client to perform NAS operations, the file system names on the NAS device
must have a forward slash (“/”) as the first character. This restriction does not
affect NAS operations initiated from the Tivoli Storage Manager server command
line.

Data formats for NDMP backup operations

During filer-to-filer backup operations that use NDMP (network data management
protocol) and are not stored in the Tivoli Storage Manager server storage hierarchy,
the network-attached storage (NAS) file server controls the format of the data
written to the tape library.

The NDMP format is not the same as the data format used for traditional Tivoli
Storage Manager backups. When you define a NAS file server as a data mover and
define a storage pool for NDMP operations, you specify the data format. For
example, you would specify NETAPPDUMP if the NAS file server is a NetApp or
an IBM System Storage N Series device. You would specify CELERRADUMP if the
NAS file server is an EMC Celerra device. For all other devices, you would specify
NDMPDUMP.

NDMP operations management

There are several administrator activities for NDMP operations.

These include:
v NAS nodes
v Data movers
v Tape libraries and drives
v Paths
v Device classes
v Storage pools
v Table of contents

206 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Managing NAS file server nodes
You can update, query, rename, and remove NAS (network attached storage)
nodes.

About this task

For example, assume you have created a new policy domain named NASDOMAIN
for NAS nodes and you want to update a NAS node named NASNODE1 to
include it in the new domain.

Procedure
1. Query the node.
query node nasnode1 type=nas
2. Change the domain of the node by issuing the following command:
update node nasnode1 domain=nasdomain

Renaming a NAS node

To rename a NAS (network attached storage) node, you must also rename the
corresponding NAS data mover; both must have the same name.

About this task

For example, to rename NASNODE1 to NAS1 you must perform the following
steps:

Procedure
1. Delete all paths between data mover NASNODE1 and libraries and between
data mover NASNODE1 and drives.
2. Delete the data mover defined for the NAS node.
3. To rename NASNODE1 to NAS1, issue the following command:
rename node nasnode1 nas1
4. Define the data mover using the new node name. In this example, you must
define a new data mover named NAS1 with the same parameters used to
define NASNODE1.
Attention: When defining a new data mover for a node that you have
renamed, ensure that the data mover name matches the new node name and
that the new data mover parameters are duplicates of the original data mover
parameters. Any mismatch between a node name and a data mover name or
between new data mover parameters and original data mover parameters can
prevent you from establishing a session with the NAS file server.
5. For SCSI or 349X libraries, define a path between the NAS data mover and a
library only if the tape library is physically connected directly to the NAS file
server.
6. Define paths between the NAS data mover and any drives used for NDMP
(network data management protocol) operations.

Chapter 9. Using NDMP for operations with NAS file servers 207
 Deleting a NAS node
To delete a NAS (network attached storage) node, first delete any file spaces for
the node. Then delete any paths from the data mover before deleting the data
mover.

Procedure
1. Delete any virtual file space definitions for the node.
2. Enter the following command:
remove node nas1

Managing data movers used in NDMP operations

You can update, query, and delete the data movers that you define for NAS
(network attached storage) file servers.

About this task

For example, if you shut down a NAS file server for maintenance, you might want
to take the data mover offline.

Procedure
1. Query your data movers to identify the data mover for the NAS file server that
you want to maintain.
query datamover nasnode1
2. Issue the following command to make the data mover offline:
update datamover nasnode1 online=no

To delete the data mover, you must first delete any path definitions in which
the data mover has been used as the source.
3. Issue the following command to delete the data mover:
delete datamover nasnode1

Attention: If the data mover has a path to the library, and you delete the data
mover or make the data mover offline, you disable access to the library.

Dedicating a Tivoli Storage Manager drive to NDMP operations

If you are already using a drive for Tivoli Storage Manager operations, you can
dedicate that drive to NDMP (network data management protocol) operations.

Procedure

Remove Tivoli Storage Manager server access by deleting the path definition with
the following command:
delete path server1 nasdrive1 srctype=server desttype=drive library=naslib

208 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Storage pool management for NDMP operations
When NETAPPDUMP, CELERRADUMP, or NDMPDUMP are designated as the
type of storage pool, managing the storage pools produced by NDMP (network
data management protocol) operations is different from managing storage pools
containing media for traditional Tivoli Storage Manager backups.

You can query and update storage pools. You cannot update the DATAFORMAT
parameter.

You cannot designate a Centera storage pool as a target pool of NDMP operations.

Maintaining separate storage pools for data from different NAS vendors is
suggested even though the data format for both is NDMPDUMP.

The following DEFINE STGPOOL and UPDATE STGPOOL parameters are ignored
because storage pool hierarchies, reclamation, and migration are not supported for
these storage pools:
MAXSIZE
NEXTSTGPOOL
LOWMIG
HIGHMIG
MIGDELAY
MIGCONTINUE
RECLAIMSTGPOOL
OVFLOLOCATION

Attention: Ensure that you do not accidentally use storage pools that were
defined for NDMP operations in traditional Tivoli Storage Manager operations. Be
especially careful when you assign the storage pool name as the value for the
DESTINATION parameter of the DEFINE COPYGROUP command. Unless the destination
is a storage pool with the appropriate data format, the backup fails.

Managing tables of contents

You can use several commands to manage different aspects of your data contents.

About this task

The SET TOCLOADRETENTION command can be used to specify the approximate

number of minutes that an unreferenced table of contents (TOC) remains loaded in
the Tivoli Storage Manager database. The Tivoli Storage Manager server-wide table
of contents retention value will determine how long a loaded TOC is retained in
the database after the latest access to information in the TOC.

Because TOC information is loaded into temporary database tables, this

information is lost if the server is halted, even if the TOC retention period has not
elapsed. At installation, the retention time is set to 120 minutes. Use the QUERY
STATUS command to see the TOC retention time.

Issue the QUERY NASBACKUP command to display information about the file system
image objects that have been backed up for a specific NAS (network attached
storage) node and file space. By issuing the command, you can see a display of all
backup images generated by NDMP (network data management protocol) and
whether each image has a corresponding table of contents.

Chapter 9. Using NDMP for operations with NAS file servers 209
 Note: The Tivoli Storage Manager server can store a full backup in excess of the
number of versions you specified, if that full backup has dependent differential
backups. Full NAS backups with dependent differential backups behave like other
base files with dependent subfiles. Due to retention time specified in the RETAIN
EXTRA setting, the full NAS backup will not be expired, and the version will be
displayed in the output of a QUERY NASBACKUP command. See “File expiration and
expiration processing” on page 479 for details.

Use the QUERY TOC command to display files and directories in a backup image
generated by NDMP. By issuing the QUERY TOC server command, you can
display all directories and files within a single specified TOC. The specified TOC
will be accessed in a storage pool each time the QUERY TOC command is issued
because this command does not load TOC information into the Tivoli Storage
Manager database. Then, use the RESTORE NODE command with the FILELIST
parameter to restore individual files.

Preventing long-running, inactive NDMP connections from

closing
To prevent firewalls from closing network data-management protocol (NDMP)
connections that are long-running but inactive, you can enable Transmission
Control Protocol (TCP) keepalive on the NDMP control connections.

About this task

The Tivoli Storage Manager server initiates control connections to network-attached

storage (NAS) devices during NDMP backup or restore operations. These control
connections might remain open and inactive for an extended amount of time. For
example, suppose that two NDMP operations are started for the same NAS device.
The control connection for one NDMP operation might remain open but inactive if
the operation requires a resource, for example, a tape drive or sequential volume,
that is being used by the other NDMP operation.

Some firewall software is configured to automatically close network connections

that are inactive for a specified length of time. If a firewall exists between a Tivoli
Storage Manager server and a NAS device, it is possible that the firewall can close
NDMP control connections unexpectedly and cause the NDMP operation to fail.

The Tivoli Storage Manager server provides a mechanism, TCP keepalive, that you
can enable to prevent long-running, inactive connections from being closed. If TCP
keepalive is enabled, small packets are sent across the network at predefined
intervals to the connection partner.

Restriction: To prevent errors, do not enable TCP keepalive in certain types of

environments. One example is environments that do not have firewalls between
the Tivoli Storage Manager server and a NAS device. Another example is
environments with firewalls that tolerate long-running, inactive connections.
Enabling TCP keepalive in these types of environments can cause an idle
connection to be inadvertently closed if the connection partner temporarily fails to
respond to TCP keepalive packets.

210 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 Enabling TCP keepalive
To enable TCP keepalive, which keeps NDMP connections open, use the
NDMPENABLEKEEPALIVE server option.

Procedure

Add the option to the server options file [Link]:

ndmpenablekeepalive yes

Example

To update the server option, you can use the SETOPT command.

Specifying connection idle time for TCP keepalive

To specify the amount of connection idle time, in minutes, before the first TCP
keepalive packet is sent, use the NDMPKEEPIDLEMINUTES server option.

Procedure

Add the option to the server options file [Link]:

ndmpkeepidleminutes minutes

Example

To update the server option, you can use the SETOPT command.

Configuring Tivoli Storage Manager for NDMP operations

Before beginning the configuration of Tivoli Storage Manager for NDMP (network
data management protocol) operations, ensure that you register the required
license.

Procedure
1. Set up the tape library and media. See “Configuring a tape library for NDMP
operations” on page 217, where the following steps are described in more
detail.
a. Attach the SCSI or virtual tape library (VTL) library to the NAS file server
or to the Tivoli Storage Manager server, or attach the ACSLS library or
349X library to the Tivoli Storage Manager server.
b. Define the library with a library type of SCSI, VTL, ACSLS, or 349X.
c. Define a device class for the tape drives.
d. Define a storage pool for NAS backup media.
e. Define a storage pool for storing a table of contents. This step is optional.
2. Configure Tivoli Storage Manager policy for managing NAS image backups.
See “Configuring Tivoli Storage Manager policy for NDMP operations” on
page 212.
3. Register a NAS file server node with the Tivoli Storage Manager server. See
“Registering NAS nodes with the Tivoli Storage Manager server” on page 223.
4. Define a data mover for the NAS file server. See “Defining a data mover for
the NAS file server” on page 224.
5. Define a path from either the Tivoli Storage Manager server or the NAS file
server to the library. See “Defining paths to libraries for NDMP operations” on
page 227.

Chapter 9. Using NDMP for operations with NAS file servers 211
 6. Define the tape drives to Tivoli Storage Manager, and define the paths to
those drives from the NAS file server and optionally from the Tivoli Storage
Manager server. See “Defining paths for NDMP operations” on page 224.
7. Check tapes into the library and label them. See “Labeling and checking tapes
into the library” on page 228.
8. Set up scheduled backups for NAS file servers. This step is optional. See
“Scheduling NDMP operations” on page 228.
9. Define a virtual file space name. This step is optional. See “Defining virtual
file spaces” on page 228.
10. Configure for tape-to-tape copy to back up data. This step is optional. See
“Tape-to-tape copy to back up data” on page 229.
11. Configure for tape-to-tape copy to move data to a different tape technology.
This step is optional. See “Tape-to-tape copy to move data” on page 229.

Configuring Tivoli Storage Manager policy for NDMP

operations
Policy lets you manage the number and retention time of NDMP (network data
management protocol) image backup versions.

About this task

See “Configuring policy for NDMP operations” on page 524 for more information.
Complete the following steps to configure Tivoli Storage Manager policy for
NDMP operations:

Procedure
1. Create a policy domain for NAS (network attached storage) file servers. For
example, to define a policy domain that is named NASDOMAIN, enter the
following command:
define domain nasdomain description=’Policy domain for NAS file servers’
2. Create a policy set in that domain. For example, to define a policy set named
STANDARD in the policy domain named NASDOMAIN, issue the following
command:
define policyset nasdomain standard
3. Define a management class, and then assign the management class as the
default for the policy set. For example, to define a management class named
MC1 in the STANDARD policy set, and assign it as the default, issue the
following commands:
define mgmtclass nasdomain standard mc1
assign defmgmtclass nasdomain standard mc1
4. Define a backup copy group in the default management class. The destination
must be the storage pool you created for backup images produced by NDMP
operations. In addition, you can specify the number of backup versions to
retain. For example, to define a backup copy group for the MC1 management
class where up to four versions of each file system are retained in the storage
pool named NASPOOL, issue the following command:
define copygroup nasdomain standard mc1 destination=naspool verexists=4

If you also chose the option to create a table of contents, TOCDESTINATION

must be the storage pool you created for the table of contents.
define copygroup nasdomain standard mc1 destination=naspool
tocdestination=tocpool verexists=4

212 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 Attention: When defining a copy group for a management class to which a
file system image produced by NDMP will be bound, be sure that the
DESTINATION parameter specifies the name of a storage pool that is defined
for NDMP operations. If the DESTINATION parameter specifies an invalid
storage pool, backups via NDMP will fail.
5. Activate the policy set. For example, to activate the STANDARD policy set in
the NASDOMAIN policy domain, issue the following command:
activate policyset nasdomain standard
The policy is ready to be used. Nodes are associated with Tivoli Storage
Manager policy when they are registered. For more information, see
“Registering NAS nodes with the Tivoli Storage Manager server” on page 223.

Policy for backups initiated with the client interface

When a client node initiates a backup, the policy is affected by the option file for
that client node.

You can control the management classes that are applied to backup images
produced by NDMP (network data management protocol) operations regardless of
which node initiates the backup. You can do this by creating a set of options to be
used by the client nodes. The option set can include an [Link] statement
to specify the management class for NAS (network attached storage) file server
backups. See “Creating client option sets on the server” on page 465 for more
information.

Determining the location of NAS backup

When Tivoli Storage Manager uses NDMP (network data management protocol) to
protect NAS (network attached storage) file servers, the Tivoli Storage Manager
server controls operations while the NAS file server transfers the data, either to an
attached library or directly to the Tivoli Storage Manager server.

You can also use a backup-archive client to back up a NAS file server by mounting
the NAS file-server file system on the client machine (with either an NFS [network
file system] mount or a CIFS [common internet file system] map) and then backing
up as usual. Table 17 compares the three backup-and-restore methods.

Note: You can use a single method or a combination of methods in your

individual storage environment.
Table 17. Comparing methods for backing up NDMP data
NDMP: Filer to attached Backup-archive client to
Property NDMP: Filer to server library server
Network data traffic All backup data goes across The server controls All backup data goes across
the LAN from the NAS file operations remotely, but the the LAN from the NAS
server to the server. NAS device moves the data device to the client and
locally. then to the server.
File server processing Less file server processing Less file server processing More file server processing
during backup is required, compared to is required, compared to is required because file
the backup-archive client the backup-archive client backups require additional
method, because the method, because the overhead for file access
backup does not use file backup does not use file protocols such as NFS and
access protocols such as access protocols such as CIFS.
NFS and CIFS. NFS and CIFS.

Chapter 9. Using NDMP for operations with NAS file servers 213
Table 17. Comparing methods for backing up NDMP data (continued)
NDMP: Filer to attached Backup-archive client to
Property NDMP: Filer to server library server
Distance between devices The Tivoli Storage Manager The Tivoli Storage Manager The Tivoli Storage Manager
server must be within SCSI server can be distant from server must be within SCSI
or Fibre Channel range of the NAS file server and the or Fibre Channel range of
the tape library. tape library. the tape library.
Firewall considerations More stringent than Less stringent than Client passwords and data
filer-to-attached- library filer-to-server because are encrypted.
because communications communications can be
can be initiated by either initiated only by the Tivoli
the Tivoli Storage Manager Storage Manager server.
server or the NAS file
server.
Security considerations Data is sent unencrypted Method must be used in a Port number configuration
from NAS file server to the trusted environment allows for secure
Tivoli Storage Manager because port numbers are administrative sessions
server. not secure. within a private network.
Load on the Tivoli Storage Higher CPU workload is Lower CPU workload is Higher CPU workload is
Manager server required to manage all back required because migration required to manage all back
end data processes (for and reclamation are not end data processes.
example, migration). supported.
Backup of primary storage Data can be backed up only Data can be backed up only Data can be backed up only
pools to copy storage pools to copy storage pools that to copy storage pools that to copy storage pools that
have the NATIVE data have the same NDMP data have the NATIVE data
format. format (NETAPPDUMP, format.
CELERRADUMP, or
NDMPDUMP).
Restore of primary storage Data can be restored only Data can be restored only Data can be restored only
pools and volumes from to storage pools and to storage pools and to storage pools and
copy storage pools volumes that have the volumes that have the same volumes that have the
NATIVE data format. NDMP format. NATIVE data format.
Moving NDMP data from Data can be moved to Data can be moved to Data can be moved to
storage pool volumes another storage pool only if another storage pool only if another storage pool only if
it has a NATIVE data it has the same NDMP data it has a NATIVE data
format. format. format.
Migration from one Supported Not supported Supported
primary storage pool to
another
Reclamation of a storage Supported Not supported Supported
pool
Simultaneous-write Not supported Not supported Supported
operations during backups
Export and import Not supported Not supported Supported
operations
Backup set generation Not supported Not supported Supported
Cyclic Redundancy Supported Not supported Supported
Checking (CRC) when data
is moved using Tivoli
Storage Manager processes
Validation using Tivoli Supported Not supported Supported
Storage Manager audit
commands

214 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Table 17. Comparing methods for backing up NDMP data (continued)
NDMP: Filer to attached Backup-archive client to
Property NDMP: Filer to server library server
Disaster recovery manager Supported Supported Supported

Tape libraries and drives for NDMP operations

Most of the planning required to implement backup and recovery operations that
use NDMP (network data management protocol) is related to device configuration.
You have choices about how to connect and use the libraries and drives.

Many of the configuration choices you have for libraries and drives are determined
by the hardware features of your libraries. You can set up NDMP operations with
any supported library and drives. However, the more features your library has, the
more flexibility you can exercise in your implementation.

You might start by answering the following questions:

v What type of library (SCSI, ACSLS, or 349X) will you use?
v If you are using a SCSI library, do you want to attach tape library robotics to the
Tivoli Storage Manager server or to the network-attached storage (NAS) file
server?
v Will you want to move your NDMP data to tape?
v How do you want to use the tape drives in the library?
– Dedicate all tape drives to NDMP operations.
– Dedicate some tape drives to NDMP operations and others to traditional
Tivoli Storage Manager operations.
– Share tape drives between NDMP operations and traditional Tivoli Storage
Manager operations.
v Will you back up data tape-to-tape for disaster recovery functions?
v Will you send backup data to a single Tivoli Storage Manager server instead of
attaching a tape library to each NAS device?
v Do you want to keep all hardware on the Tivoli Storage Manager server and
send NDMP data over the LAN?

Determining library drive usage when backing up to

NAS-attached libraries
Drives can be used for multiple purposes because of the flexible configurations
allowed by Tivoli Storage Manager. For NDMP (network data management
protocol) operations, the NAS (network attached storage) file server must have
access to the drive. The Tivoli Storage Manager server can also have access to the
same drive, depending on your hardware connections and limitations.

About this task

All drives are defined to the Tivoli Storage Manager server. However, the same
drive may be defined for both traditional Tivoli Storage Manager operations and
NDMP operations. Figure 14 on page 216 illustrates one possible configuration. The
Tivoli Storage Manager server has access to drives 2 and 3, and each NAS file
server has access to drives 1 and 2.

Chapter 9. Using NDMP for operations with NAS file servers 215
 NAS File Server 1

Tape Library

NAS File Server 2

1
2
3

Tivoli Storage Manager

Server

Legend:
Drive access
Drives 1 2 3

Figure 14. Tivoli Storage Manager drive usage example

To create the configuration shown in Figure 14, perform the following steps:

Procedure
1. Define all three drives to Tivoli Storage Manager.
2. Define paths from the Tivoli Storage Manager server to drives 2 and 3. Because
drive 1 is not accessed by the server, no path is defined.
3. Define each NAS file server as a separate data mover.
4. Define paths from each data mover to drive 1 and to drive 2.

Results

To use the Tivoli Storage Manager back end data movement operations, the Tivoli
Storage Manager server requires two available drive paths from a single NAS data
mover. The drives can be in different libraries and can have different device types
that are supported by NDMP. You can make copies between two different tape
devices, for example, the source tape drive can be an DLT drive in a library and
the target drive can be an LTO drive in another library.

During Tivoli Storage Manager back end data movements, the Tivoli Storage
Manager server locates a NAS data mover that supports the same data format as
the data to be copied from and that has two available mount points and paths to
the drives. If the Tivoli Storage Manager server cannot locate such a data mover,
the requested data movement operation is not performed. The number of available
mount points and drives depends on the mount limits of the device classes for the
storage pools involved in the back end data movements.

If the back end data movement function supports multiprocessing, each concurrent
Tivoli Storage Manager back end data movement process requires two available
mount points and two available drives. To run two Tivoli Storage Manager
processes concurrently, at least four mount points and four drives must be
available.

216 IBM Tivoli Storage Manager for Windows: Administrator's Guide

See “Defining paths for NDMP operations” on page 224 for more information.

Configuring a tape library for NDMP operations

You can configure a tape library to backup a network-attached storage (NAS)
device to tape.

About this task

Perform the following steps to set up tape libraries for NDMP operations:

Procedure
1. Connect the library and drives to be used for NDMP operations.
a. Connect the SCSI library. Before you set up a SCSI tape library for NDMP
operations, determine whether you want to attach your library robotics
control to the Tivoli Storage Manager server or to the NAS file server. See
“Tape libraries and drives for NDMP operations” on page 215. Connect the
SCSI tape library robotics to the Tivoli Storage Manager server or to the
NAS file server. Refer to your device manufacturer documentation for
instructions.
If the library is connected to Tivoli Storage Manager, make a SCSI or
Fibre Channel connection between the Tivoli Storage Manager server
and the library robotics control port. Then, connect the NAS file server
with the drives.
If the library is connected to NAS file server, make a SCSI or Fibre
Channel connection between the NAS file server and the library robotics
and drives.
b. Connect the ACSLS Library. Connect the ACSLS tape library to the Tivoli
Storage Manager server.
c. Connect the 349X Library. Connect the 349X tape library to the Tivoli
Storage Manager server.
2. Define the library for you library device by issuing the DEFINE LIBRARY
command. The library must be a single device type, not a mixed device. Issue
one of the following commands to define the library depending on the type of
device that you are configuring:
SCSI Library
define library tsmlib libtype=scsi
ACSLS Library
define library acslib libtype=acsls acsid=1
349X Library
define library tsmlib libtype=349x
3. Define a device class for your NDMP device by issuing the DEFINE DEVCLASS
command.

Tip: A device class that is defined with a device type of NAS is not explicitly
associated with a specific drive type, for example, LTO. However, as best
practice, define a separate device classes for different drive types.
In the DEFINE DEVCLASS command, use the following parameters and values:
v Specify DEVTYPE=NAS.
v Specify MOUNTRETENTION=0 . It is required for NDMP operations.
v Specify a value for the ESTCAPACITY parameter.

Chapter 9. Using NDMP for operations with NAS file servers 217
 For example, to define a device class named NASCLASS for a library named
NASLIB with an estimated capacity is 40 GB for the media, issue the following
command:
define devclass nasclass devtype=nas library=naslib mountretention=0
estcapacity=40g
4. Define a storage pool for NDMP media by issuing the DEFINE STGPOOL
command. When NETAPPDUMP, CELERRADUMP, or NDMPDUMP is designated as the
type of storage pool, managing the storage pools that are produced by NDMP
operations is different from managing storage pools that contain media for
traditional Tivoli Storage Manager backups. Tivoli Storage Manager operations
use storage pools that are defined with a NATIVE or NONBLOCK data format.
If you select NETAPPDUMP, CELERRADUMP, or NDMPDUMP, NDMP operations require
storage pools with a data format that matches the NAS file server and the
selected backup method. Maintaining separate storage pools for data from
different NAS vendors is best practice, even though the data format for both is
NDMPDUMP.
For example, to define a storage pool named NDMPPOOL for a file server,
which is neither a NetApp or a Celerra file server, issue the following
command:
define stgpool ndmppool nasclass maxscratch=10 dataformat=ndmpdump
To define a storage pool named NASPOOL for a NetApp file server, issue the
following command:
define stgpool naspool nasclass maxscratch=10 dataformat=netappdump
To define a storage pool named CELERRAPOOL for an EMC Celerra file server,
issue the following command:
define stgpool celerrapool nasclass maxscratch=10 dataformat=celerradump

Attention: Ensure that you do not accidentally use storage pools that are
defined for NDMP operations in traditional Tivoli Storage Manager operations.
Be especially careful when you assign the storage pool name as the value for
the DESTINATION parameter of the DEFINE COPYGROUP command. Unless the
destination is a storage pool with the appropriate data format, the backup can
fail.
5. Optionally, define a storage pool for a table of contents. If you plan to create a
table of contents, you should also define a disk storage pool in which to store
the table of contents. You must set up policy so that the Tivoli Storage Manager
server stores the table of contents in a different storage pool from the one
where the backup image is stored. The table of contents is treated like any
other object in that storage pool.
For example, to define a storage pool named TOCPOOL for a DISK device
class, issue the following command:
define stgpool tocpool disk

Then, define volumes for the storage pool.

For more information about defining volumes, see “Configuring random access
volumes on disk devices” on page 69.

218 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 Attaching tape library robotics for NAS-attached libraries
If you have decided to back up your network-attached storage (NAS) data to a
library directly attached to the NAS device and are using a SCSI tape library, one
of the first steps in planning for NDMP (network data management protocol)
operations is to determine where to attach it.

About this task

You must determine whether to attach the library robotics to the Tivoli Storage
Manager server or to the NAS file server. Regardless of where you connect library
robotics, tape drives must always be connected to the NAS file server for NDMP
operations.

Distance and your available hardware connections are factors to consider for SCSI
libraries. If the library does not have separate ports for robotics control and drive
access, the library must be attached to the NAS file server because the NAS file
server must have access to the drives. If your SCSI library has separate ports for
robotics control and drive access, you can choose to attach the library robotics to
either the Tivoli Storage Manager server or the NAS file server. If the NAS file
server is at a different location from the Tivoli Storage Manager server, the distance
may mean that you must attach the library to the NAS file server.

Whether you are using a SCSI, ACSLS, or 349X library, you have the option of
dedicating the library to NDMP operations, or of using the library for NDMP
operations as well as most traditional Tivoli Storage Manager operations.
Table 18. Summary of configurations for NDMP operations
Distance between Drive sharing Drive sharing
Tivoli Storage between Tivoli Drive sharing between storage
Manager server and Storage Manager between NAS agent and NAS
Configuration library Library sharing and NAS file server file servers file server
Configuration 1
(SCSI library
Limited by SCSI or
connected to the Supported Supported Supported Supported
FC connection
Tivoli Storage
Manager server)
Configuration 2
(SCSI library
No limitation Not supported Supported Supported Not supported
connected to the
NAS file server)
Configuration 3 May be limited by
Supported Supported Supported Supported
(349X library) 349X connection
Configuration 4 May be limited by
Supported Supported Supported Supported
(ACSLS library) ACSLS connection

Chapter 9. Using NDMP for operations with NAS file servers 219
 Configuration 1: SCSI library connected to the Tivoli Storage
Manager server
In this configuration, the tape library must have separate ports for robotics control
and for drive access. In addition, the library must be within Fibre-Channel range
or SCSI bus range of both the Tivoli Storage Manager server and the
network-attached storage (NAS) file server.

In this configuration, the Tivoli Storage Manager server controls the SCSI library
through a direct, physical connection to the library robotics control port. For
NDMP (network data management protocol) operations, the drives in the library
are connected directly to the NAS file server, and a path must be defined from the
NAS data mover to each of the drives to be used. The NAS file server transfers
data to the tape drive at the request of the Tivoli Storage Manager server. To also
use the drives for Tivoli Storage Manager operations, connect the Tivoli Storage
Manager server to the tape drives and define paths from the Tivoli Storage
Manager server to the tape drives. This configuration also supports a Tivoli
Storage Manager storage agent having access to the drives for its LAN-free
operations, and the Tivoli Storage Manager server can be a library manager.

Tivoli Storage
Manager
server
Tape
Web client library
(optional) 1 2

NAS file
server

Legend:
SCSI or Fibre Channel connection
TCP/IP connection
Data flow
Robotics control 1
NAS file server
Drive access 2 file system disks

Figure 15. Configuration 1: SCSI library connected to Tivoli Storage Manager server

Configuration 2: SCSI library connected to the NAS file server

In this configuration, the library robotics and the drives must be physically
connected directly to the NAS (network attached storage) file server, and paths
must be defined from the NAS data mover to the library and drives. No physical
connection is required between the Tivoli Storage Manager server and the SCSI
library.

The Tivoli Storage Manager server controls library robotics by sending library
commands across the network to the NAS file server. The NAS file server passes
the commands to the tape library. Any responses generated by the library are sent
to the NAS file server, and passed back across the network to the Tivoli Storage
Manager server. This configuration supports a physically distant Tivoli Storage
Manager server and NAS file server. For example, the Tivoli Storage Manager
server could be in one city, while the NAS file server and tape library are in
220 IBM Tivoli Storage Manager for Windows: Administrator's Guide
another city.

Tivoli Storage
Manager
server
Tape
Web client library
(optional)

2
1

NAS file
server

Legend:
SCSI or Fibre Channel connection
TCP/IP connection
Data flow
Robotics control 1
NAS file server
Drive access 2 file system disks

Figure 16. Configuration 2: SCSI library connected to the NAS file server

Configuration 3: 349x library connected to the Tivoli Storage

Manager server
For this configuration, you connect the tape library to the system as for traditional
operations.

In this configuration, the 349X tape library is controlled by the Tivoli Storage
Manager server. The Tivoli Storage Manager server controls the library by passing
the request to the 349X library manager through TCP/IP.

In order to perform NAS (network attached storage) backup or restore operations,

the NAS file server must be able to access one or more tape drives in the 349X
library. Any tape drives used for NAS operations must be physically connected to
the NAS file server, and paths need to be defined from the NAS data mover to the
drives. The NAS file server transfers data to the tape drive at the request of the
Tivoli Storage Manager server. Follow the manufacturer's instructions to attach the
device to the server system.

This configuration supports a physically distant Tivoli Storage Manager server and
NAS file server. For example, the Tivoli Storage Manager server could be in one
city, while the NAS file server and tape library are in another city.

Chapter 9. Using NDMP for operations with NAS file servers 221
 Tivoli Storage
Manager server

349X tape
Web client 1 library
(optional)

NAS file server

Legend:
SCSI or Fibre Channel connection
Library connection
TCP/IP connection
Data flow
NAS file server
Drive access 1 2 file system disks
1 is optional

Figure 17. Configuration 3: 349x library connected to the Tivoli Storage Manager server

Configuration 4: ACSLS library connected to the Tivoli Storage

Manager server
For this configuration, connect the tape library to the system as you do for
traditional Tivoli Storage Manager operations.

The ACSLS (automated cartridge system library software) tape library is controlled
by the Tivoli Storage Manager server. The Tivoli Storage Manager server controls
the library by passing the request to the ACSLS library server through TCP/IP. The
ACSLS library supports library sharing and LAN-free operations.

Restriction: In order to utilize ACSLS functions, StorageTek Library Attach

software must be installed. See “ACSLS-managed libraries” on page 109 for more
information.

In order to perform NAS (network attached storage) backup or restore operations,

the NAS file server must be able to access one or more tape drives in the ACSLS
library. Any tape drives used for NAS operations must be physically connected to
the NAS file server, and any paths need to be defined from the NAS data mover to
the drives. The NAS file server transfers data to the tape drive at the request of the
Tivoli Storage Manager server. Follow the manufacturer's instructions to attach the
device to the server system.

This configuration supports a physically distant Tivoli Storage Manager server and
NAS file server. For example, the Tivoli Storage Manager server could be in one
city while the NAS file server and tape library are in another city.

To also use the drives for Tivoli Storage Manager operations, connect the Tivoli
Storage Manager server to the tape drives and define paths from the Tivoli Storage
Manager server to the tape drives.

222 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 Tivoli Storage
Manager server

ACSLS tape
Web client 1 library
(optional)

NAS file server

Legend:
SCSI or Fibre Channel connection
Library connection
TCP/IP connection
Data flow
NAS file server
Drive access 1 2 file system disks
1 is optional

Figure 18. Configuration 4: ACSLS library connected to the Tivoli Storage Manager server

Registering NAS nodes with the Tivoli Storage Manager server

Register the NAS (network attached storage) file server as a Tivoli Storage
Manager node, specifying TYPE=NAS. This node name is used to track the image
backups for the NAS file server.

Procedure

To register a NAS file server as a node named NASNODE1, with a password of

NASPWD1, in a policy domain named NASDOMAIN, issue the following example
command:
register node nasnode1 naspwd1 domain=nasdomain type=nas

If you are using a client option set, specify the option set when you register the
node.
You can verify that this node is registered by issuing the following command:
query node type=nas

Important: You must specify TYPE=NAS so that only NAS nodes are displayed.

Chapter 9. Using NDMP for operations with NAS file servers 223
 Defining a data mover for the NAS file server
Define a data mover for each NAS (network attached storage) file server, using
NDMP (network data management protocol) operations in your environment. The
data mover name must match the node name that you specified when you
registered the NAS node to the Tivoli Storage Manager server.

Procedure

To define a data mover for a NAS node named NASNODE1, enter the following
example command:
define datamover nasnode1 type=nas hladdress=netapp2 lladdress=10000 userid=root
password=admin dataformat=netappdump

In this command:
v The high-level address is an IP address for the NAS file server, either a
numerical address or a host name.
v The low-level address is the IP port for NDMP sessions with the NAS file server.
The default is port number 10000.
v The user ID is the ID defined to the NAS file server that authorizes an NDMP
session with the NAS file server (for this example, the user ID is the
administrative ID for the NetApp file server).
v The password parameter is a valid password for authentication to an NDMP
session with the NAS file server.
v The data format is NETAPPDUMP. This is the data format that the NetApp file
server uses for tape backup. This data format must match the data format of the
target storage pool.

Defining paths for NDMP operations

For network data management protocol (NDMP) operations, you create paths to
drives and to libraries.

Defining paths to drives for NDMP operations

The method that you choose for creating paths to drives depends on whether the
drives are accessed by a network-attached storage (NAS) file server and the Tivoli
Storage Manager server or only by a NAS file server.

Defining paths for drives attached only to a NAS file server and to the Tivoli
Storage Manager server:

If a tape drive is to be accessed by a network-attached storage (NAS) file server

and the Tivoli Storage Manager server, you must create two paths. One path exists
between the tape drive and the NAS file server. The other path exists between the
tape drive and the Tivoli Storage Manager server.

About this task

Complete the following steps:

Procedure
1. If the drive is not defined for the Tivoli Storage Manager server, create the
drive definition. For example, to define a drive NASDRIVE1 for a library
NASLIB, issue the following command:
define drive naslib nasdrive1 element=autodetect

224 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 Remember: If the drive is attached to the Tivoli Storage Manager server, the
element address is automatically detected .
2. Map the NAS drive name to the corresponding drive definition on the Tivoli
Storage Manager server:
v On the Tivoli Storage Manager server, issue the QUERY DRIVE FORMAT=DETAILED
command to obtain the worldwide name (WWN) and serial number for the
drive that is to be connected to the NAS file server.
v On the NAS device, obtain the tape device name, serial number, and WWN
for the drive.
If the WWN or serial number matches, a drive on a NAS file server is the same
as the drive on the Tivoli Storage Manager server .
3. Using the drive name, define a path to the drive from the NAS file server and
a path to the drive from the Tivoli Storage Manager server.
v For example, to define a path between a tape drive with a device name of
rst01 and a NetApp file server, issue the following command:
define path nasnode1 nasdrive1 srctype=datamover desttype=drive
library=naslib device=rst01
v To define a path between the tape drive and the Tivoli Storage Manager
server, issue the following command:
define path server1 nasdrive1 srctype=server desttype=drive
library=naslib device=mt3.0.0.2
Related information:
Obtaining device names for devices attached to NAS file servers

Defining paths for drives attached only to NAS file servers:

If a tape drive is to be accessed only by a network-attached storage (NAS) file

server and not by the Tivoli Storage Manager server, only a single path between
the tape drive and the NAS file server is required.

About this task

Complete the following steps:

Procedure
1. Obtain the SCSI element addresses, worldwide name (WWN), and serial
numbers for the drive to be connected to NAS file server.

Restriction: If the SCSI drive is connected only to a NAS file server, the
element address is not automatically detected, and you must supply it. If a
library has more than one drive, you must specify an element address for each
drive.
To obtain a SCSI element address, go to one of the following Tivoli
device-support websites:
v AIX, HP-UX, Solaris, and Windows: [Link]
sysmgmt/products/support/
IBM_TSM_Supported_Devices_for_AIXHPSUNWIN.html
v Linux: [Link]
IBM_TSM_Supported_Devices_for_Linux.html
Element number assignment and device WWN assignments are also available
from tape-library device manufacturers.

Chapter 9. Using NDMP for operations with NAS file servers 225
 2. Create drive definitions by specifying the element addresses identified in the
preceding step. Specify the element address in the ELEMENT parameter of the
DEFINE DRIVE command. For example, to define a drive NASDRIVE1 with the
element address 82 for the library NASLIB, issue the following command:
define drive naslib nasdrive1 element=82

Attention: For a drive connected only to the NAS file server, do not specify
ASNEEDED as the value for the CLEANFREQUENCY parameter of the DEFINE DRIVE
command.
3. Obtain the device name, serial number, and WWN for the drive on the NAS
device.
4. Using the information obtained in steps 1 and 3, map the NAS device name to
the element address in the drive definition in the Tivoli Storage Manager
server.
5. Define a path between the tape drive and the NAS file server. For example, to
define a path between a NetApp file server and a tape drive with a device
name of rst01, issue the following command:
define path nasnode1 nasdrive1 srctype=datamover desttype=drive
library=naslib device=rst0l
Related information:
Obtaining device names for devices attached to NAS file servers

Obtaining names for devices attached to NAS file servers:

For paths from a network-attached storage (NAS) data mover, the value of the
DEVICE parameter in the DEFINE PATH command is the name by which the NAS file
server knows a library or drive.

About this task

You can obtain these device names, also known as special file names, by querying
the NAS file server. For information about how to obtain names for devices that
are connected to a NAS file server, consult the product information for the file
server.

Procedure
v To obtain the device names for tape libraries on a Netapp Release ONTAP 10.0
GX, or later, file server, connect to the file server using telnet and issue the
SYSTEM HARDWARE TAPE LIBRARY SHOW command. To obtain the device names for
tape drives on a Netapp Release ONTAP 10.0 GX, or later, file server, connect to
the file server using telnet and issue the SYSTEM HARDWARE TAPE DRIVE SHOW
command. For details about these commands, see the Netapp ONTAP GX file
server product documentation.
v For releases earlier than Netapp Release ONTAP 10.0 GX, continue to use the
SYSCONFIG command. For example, to display the device names for tape libraries,
connect to the file server using telnet and issue the following command:
sysconfig -m

To display the device names for tape drives, issue the following command:
sysconfig -t
v For fibre-channel-attached drives and the Celerra data mover, complete the
following steps:

226 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 1. Log on to the EMC Celerra control workstation using an administrative ID.
Issue the following command:
server_devconfig server_1 -l -s -n

Tip: The -l option for this command lists only the device information that
was saved in the database of the data mover. The command and option do
not display changes to the device configuration that occurred after the last
database refresh on the data mover. For details about how to obtain the most
recent device configuration for your data mover, see the EMC Celerra
documentation.
The output for the server_devconfig command includes the device names
for the devices attached to the data mover. The device names are listed in the
addr column, for example:
server_1:
Scsi Device Table
name addr type info
tape1 c64t0l0 tape IBM ULT3580-TD2 53Y2
ttape1 c96t0l0 tape IBM ULT3580-TD2 53Y2
2. Map the Celerra device name to the device worldwide name (WWN):
a. To list the WWN, log on to the EMC Celerra control workstation and
issue the following command. Remember to enter a period ( . ) as the
first character in this command.
.server_config server_# -v "fcp bind show"

The output for this command includes the WWN, for example:
Chain 0064: WWN 500507630f418e29 HBA 2 N_PORT Bound
Chain 0096: WWN 500507630f418e18 HBA 2 N_PORT Bound

Note: The .server_config command is an undocumented EMC Celerra

command. For more information about how to use it, contact EMC.
b. Use the chain number to identify the tape device that was listed in the
output of the server_devconfig command and that has the same WWN,
for example:

Tape device name Chain number WWN

c64t0l0 0064 500507630f418e29
c96t0l0 0096 500507630f418e18

Celerra commands might behave differently on different EMC Celerra systems

and operating system levels. For details, see the EMC Celerra documentation or
contact EMC.

Defining paths to libraries for NDMP operations

Define a path to the SCSI library from either the Tivoli Storage Manager server or
the network-attached storage (NAS) file server.

Procedure
1. For a SCSI Library connected to Tivoli Storage Manager, issue the following
example command to define a path from the server, named SERVER1, to the
SCSI library named TSMLIB:
define path server1 tsmlib srctype=server desttype=library
device=lb0.0.0.2

Chapter 9. Using NDMP for operations with NAS file servers 227
 2. For a SCSI library connected to a NAS file server, issue the following example
command to define a path between a NetApp NAS data mover named
NASNODE1 and a library named NASLIB:
define path nasnode1 naslib srctype=datamover desttype=library device=mc0
3. For a 349X library, define a path to the library from the Tivoli Storage Manager
server. For example, issue the following command to define a path from the
server, named SERVER1, to the 349X library named TSMLIB:
define path server1 tsmlib srctype=server desttype=library
device=library1

Attention: The DEFINE PATH command is not needed for an automated

cartridge system library software (ACSLS) library.

Labeling and checking tapes into the library

You must label the tapes and check them into the tape library.

About this task

These tasks are the same as for other libraries. For more information, see:
“Labeling media” on page 129

Scheduling NDMP operations

You can schedule the backup or restore of images produced by NDMP (network
data management protocol) operations by using administrative schedules that
process the BACKUP NODE or RESTORE NODE administrative commands.

About this task

The BACKUP NODE and RESTORE NODE commands can be used only for nodes of
TYPE=NAS. See “Backing up and restoring NAS file servers using NDMP” on
page 230 for information about the commands.

Procedure

For example, to create an administrative schedule called NASSCHED to back up

all file systems for a node named NASNODE1, enter the following:
define schedule nassched type=administrative cmd=’backup node nasnode1’ active=yes
starttime=20:00 period=1 perunits=days

The schedule is active, and is set to run at 8:00 p.m. every day. See Chapter 19,
“Automating server operations,” on page 619 for more information.

Defining virtual file spaces

Use a virtual file space definition to perform NAS (network attached storage)
directory level backups. In order to reduce backup and restore times for large file
systems, map a directory path from a NAS file server to a virtual file space name
on the Tivoli Storage Manager server.

About this task

Procedure

To create a virtual file space name for the directory path on the NAS device, issue
the DEFINE VIRTUALFSMAPPING command:

228 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 define virtualfsmapping nas1 /mikesdir /vol/vol1 /mikes

This command defines a virtual file space name of /MIKESDIR on the server which
represents the directory path of /VOL/VOL1/MIKES on the NAS file server
represented by node NAS1. See “Directory-level backup and restore for NDMP
operations” on page 236 for more information.

Tape-to-tape copy to back up data

When using NDMP (network data management protocol) tape-to-tape function to
back up data, the library type can be SCSI, 349X, or ACSLS (automated cartridge
system library software). Drives can be shared between the NAS (network attached
storage) devices and the Tivoli Storage Manager server.

About this task

Note: When using the NDMP tape-to-tape copy function, your configuration setup
could affect the performance of the Tivoli Storage Manager back end data
movement.

Procedure

To have one NAS device with paths to four drives in a library, use the MOVE DATA
command after you are done with your configuration setup. This moves data on
the volume VOL1 to any available volumes in the same storage pool as VOL1:
move data vol1

Tape-to-tape copy to move data

In order to move data from an old tape technology to a new tape technology, using
NDMP (network data management protocol) tape-to-tape copy operation, perform
the steps below as well as the regular steps in your configuration setup.

About this task

Note: When using the NDMP tape-to-tape copy function, your configuration setup
could affect the performance of the Tivoli Storage Manager back end data
movement.

Procedure
1. Define one drive in the library, lib1, that has old tape technology:
define drive lib1 drv1 element=1035
2. Define one drive in the library, lib2, that has new tape technology:
define drive lib2 drv1 element=1036
3. Move data on volume vol1 in the primary storage pool to the volumes in
another primary storage pool, nasprimpool2:
move data vol1 stgpool=nasprimpool2

Chapter 9. Using NDMP for operations with NAS file servers 229
Backing up and restoring NAS file servers using NDMP
After you have completed the steps to configure Tivoli Storage Manager for NDMP
(network data management protocol) operations, you are ready to begin using
NDMP.

Procedure

Use either a client interface or an administrative interface to perform a file system

image backup. For example, to use the Windows backup-archive client interface to
back up a file system named /vol/vol1 on a NAS (network attached storage) file
server named NAS1, issue the following command:
dsmc backup nas -nasnodename=nas1 {/vol/vol1}

For more information on the command, see the Tivoli Storage Manager
Backup-Archive Clients Installation and User's Guide.

Tip: Whenever you use the client interface, you are asked to authenticate yourself
as a Tivoli Storage Manager administrator before the operation can begin. The
administrator ID must have at least client owner authority for the NAS node.
You can perform the same backup operation with a server interface. For example,
from the administrative command-line client, back up the file system named
/vol/vol1 on a NAS file server named NAS1, by issuing the following command:
backup node nas1 /vol/vol1

Note: The BACKUP NAS and BACKUP NODE commands do not include snapshots. To
back up snapshots see “Backing up and restoring with snapshots” on page 236.
You can restore the image using either interface. Backups are identical whether
they are backed up using a client interface or a server interface. For example,
suppose you want to restore the image backed up in the previous examples. For
this example the file system named /vol/vol1 is being restored to /vol/vol2.
Restore the file system with the following command, issued from a Windows
backup-archive client interface:
dsmc restore nas -nasnodename=nas1 {/vol/vol1} {/vol/vol2}

You can choose to restore the file system, using a server interface. For example, to
restore the file system name /vol/vol1 to file system /vol/vol2, for a NAS file
server named NAS1, enter the following command:
restore node nas1 /vol/vol1 /vol/vol2

You can restore data from one NAS vendor system to another NAS vendor system
when you use the NDMPDUMP data format, but you should either verify
compatibility between systems or maintain a separate storage pool for each NAS
vendor.

230 IBM Tivoli Storage Manager for Windows: Administrator's Guide

NAS file servers: backups to a single Tivoli Storage Manager
server
If you have several NAS (network attached storage) file servers located in different
locations, you might prefer to send the backup data to a single Tivoli Storage
Manager server rather than attaching a tape library to each NAS device.

When you store NAS backup data in the Tivoli Storage Manager server's storage
hierarchy, you can apply Tivoli Storage Manager back end data management
functions. Migration, reclamation, and disaster recovery are among the supported
features when using the NDMP file server to Tivoli Storage Manager server option.

In order to back up a NAS device to a Tivoli Storage Manager native storage pool,
set the destination storage pool in the copy group to point to the desired native
storage pool. The destination storage pool provides the information about the
library and drives used for backup and restore. You should ensure that there is
sufficient space in your target storage pool to contain the NAS data, which can be
backed up to sequential, disk, or file type devices. Defining a separate device class
is not necessary.

If you are creating a table of contents, a management class should be specified

with the TOCDESTINATION parameter in the DEFINE and UPDATE COPYGROUP
commands. When backing up a NAS file server to Tivoli Storage Manager native
pools, the TOCDESTINATION may be the same as the destination of the NDMP
(network data management protocol) data.

Firewall considerations are more stringent than they are for filer-to-attached-library
because communications can be initiated by either the Tivoli Storage Manager
server or the NAS file server. NDMP tape servers run as threads within the Tivoli
Storage Manager server and the tape server accepts connections on port of 10001.
This port number can be changed through the following option in the Tivoli
Storage Manager server options file: NDMPPORTRANGE port-number-low,
port-number-high.

During NDMP filer-to-server backup operations, you can use the

NDMPPREFDATAINTERFACE option to specify which network interface the Tivoli
Storage Manager server uses to receive NDMP backup data. The value for this
option is a hostname or IPV4 address that is associated with one of the active
network interfaces of the system on which the Tivoli Storage Manager server is
running. This interface must be IPV4 enabled.

Before using this option, verify that your NAS device supports NDMP operations
that use a different network interface for NDMP control and NDMP data
connections. NDMP control connections are used by Tivoli Storage Manager to
authenticate with an NDMP server and monitor an NDMP operation while NDMP
data connections are used to transmit and receive backup data during NDMP
operations. You must still configure your NAS device to route NDMP backup and
restore data to the appropriate network interface.

When enabled, the NDMPPREFDATAINTERFACE option affects all subsequent

NDMP filer-to-server operations. It does not affect NDMP control connections
because they use the system's default network interface. You can update this server
option without stopping and restarting the server by using the SETOPT command
(Set a server option for dynamic update).

Chapter 9. Using NDMP for operations with NAS file servers 231
 NetApp file servers provide an NDMP option (ndmpd.preferred_interface) to
change the interface used for NDMP data connections. Refer to the documentation
that came with your NAS device for more information.

See “Backing up NDMP file server to Tivoli Storage Manager server backups” for
steps on how to perform NDMP filer-to-server backups.

See the Administrator's Reference for server option information.

Backing up NDMP file server to Tivoli Storage Manager server

backups
You can back up data to a single Tivoli Storage Manager server rather than
attaching a tape library to each NAS device.

About this task

To back up a server on a NAS file system, complete the following steps:

Procedure
1. Select an existing storage pool or set up a storage pool for the NAS data by
issuing the following command:
define stgpool naspool disk
2. Define volumes to add to the storage pool. For example, define a volume
named naspool_volAB:
define volume naspool /usr/storage/naspool_volAB formatsize=100
3. Set the copy destination to the storage pool defined previously and activate the
associated policy set.
update copygroup standard standard standard destination=naspool
tocdestination=naspool
activate policyset standard standard

The destination for NAS data is determined by the destination in the copy
group. The storage size estimate for NAS differential backups uses the
occupancy of the file space, the same value that is used for a full backup. You
can use this size estimate as one of the considerations in choosing a storage
pool. One of the attributes of a storage pool is the MAXSIZE value, which
indicates that data be sent to the NEXT storage pool if the MAXSIZE value is
exceeded by the estimated size. Because NAS differential backups to Tivoli
Storage Manager native storage pools use the base file space occupancy size as
a storage size estimate, differential backups end up in the same storage pool as
the full backup. Depending on collocation settings, differential backups may
end up on the same media as the full backup.
4. Set up a node and data mover for the NAS device. The data format signifies
that the backup images created by this NAS device are a dump type of backup
image in a NetApp specific format.
register node nas1 nas1 type=nas domain=standard
define datamover nas1 type=nas hla=nas1 user=root
password=***** dataformat=netappdump

The NAS device is now ready to be backed up to a Tivoli Storage Manager

server storage pool. Paths may be defined to local drives, but the destination
specified by the management class determines the target location for this
backup operation.

232 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 5. Back up the NAS device to the Tivoli Storage Manager storage pool by issuing
the following command:
backup node nas1 /vol/vol0
6. Restore a NAS device from the Tivoli Storage Manager storage pool by issuing
the following command:
restore node nas1 /vol/vol0

File-level backup and restore for NDMP operations

When you do a backup via NDMP (network data management protocol), you can
specify that the Tivoli Storage Manager server collect and store file-level
information in a table of contents (TOC).

If you specify this option at the time of backup, you can later display the table of
contents of the backup image. Through the backup-archive Web client, you can
select individual files or directories to restore directly from the backup images
generated.

Collecting file-level information requires additional processing time, network

resources, storage pool space, temporary database space, and possibly a mount
point during the backup. You should consider dedicating more space in the Tivoli
Storage Manager server database. You must set up policy so that the Tivoli Storage
Manager server stores the table of contents in a different storage pool from the one
where the backup image is stored. The table of contents is treated like any other
object in that storage pool.

You also have the option to do a backup via NDMP without collecting file-level
restore information.

To allow creation of a table of contents for a backup via NDMP, you must define
the TOCDESTINATION attribute in the backup copy group for the management
class to which this backup image is bound. You cannot specify a copy storage pool
or an active-data pool as the destination. The storage pool you specify for the TOC
destination must have a data format of either NATIVE or NONBLOCK, so it
cannot be the tape storage pool used for the backup image.

If you choose to collect file-level information, specify the TOC parameter in the
BACKUP NODE server command. Or, if you initiate your backup using the client, you
can specify the TOC option in the client options file, client option set, or client
command line. You can specify NO, PREFERRED, or YES. When you specify
PREFERRED or YES, the Tivoli Storage Manager server stores file information for a
single NDMP-controlled backup in a table of contents (TOC). The table of contents
is placed into a storage pool. After that, the Tivoli Storage Manager server can
access the table of contents so that file and directory information can be queried by
the server or client. Use of the TOC parameter allows a table of contents to be
generated for some images and not others, without requiring different
management classes for the images.

See the Administrator's Reference for more information about the BACKUP NODE
command.

To avoid mount delays and ensure sufficient space, use random access storage
pools (DISK device class) as the destination for the table of contents. For sequential
access storage pools, no labeling or other preparation of volumes is necessary if
scratch volumes are allowed.

Chapter 9. Using NDMP for operations with NAS file servers 233
 See “Managing tables of contents” on page 209 for more information.

Interfaces for file-level restore

When you restore individual files and directories, you have the choice of using one
of two interfaces to initiate the restore: the backup-archive Web client or the server
interface.
Restore Using Backup-Archive Web Client
The backup-archive Web client requires that a table of contents exist in
order to restore files and directories. The Web client must be on a Windows
system. The Tivoli Storage Manager server accesses the table of contents
from the storage pool and loads TOC information into a temporary
database table. Then, you can use the backup-archive Web client to
examine directories and files contained in one or more file system images,
and select individual files or directories to restore directly from the backup
images generated.
Restore Using Server Interface
v If you have a table of contents, use the QUERY NASBACKUP command to
display information about backup images generated by NDMP (network
data management protocol), and to see which images have a
corresponding table of contents. Then, use the RESTORE NODE command
with the FILELIST parameter.
v If you did not create a table of contents, the contents of the backup
image cannot be displayed. You can restore individual files, directories,
or both if you know the name of the file or directory, and in which
image the backup is located. Use the RESTORE NODE command with the
FILELIST parameter.

International characters for NetApp file servers

All systems that create or access data on a particular NAS (network attached
storage) file server volume must do so in a manner compatible with the volume
language setting.

You should install Data ONTAP 6.4.1 or later, if it is available, on your NetApp
NAS file server in order to garner full support of international characters in the
names of files and directories.

If your level of Data ONTAP is earlier than 6.4.1, you must have one of the
following two configurations in order to collect and restore file-level information.
Results with configurations other than these two are unpredictable. The Tivoli
Storage Manager server will print a warning message (ANR4946W) during backup
operations. The message indicates that the character encoding of NDMP file history
messages is unknown, and UTF-8 will be assumed in order to build a table of
contents. It is safe to ignore this message only for the following two configurations.
v Your data has directory and file names that contain only English (7-bit ASCII)
characters.
v Your data has directory and file names that contain non-English characters and
the volume language is set to the UTF-8 version of the proper locale (for
example, [Link]-8 for German).

If your level of Data ONTAP is 6.4.1 or later, you must have one of the following
three configurations in order to collect and restore file-level information. Results
with configurations other than these three are unpredictable.

234 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 v Your data has directory and file names that contain only English (7-bit ASCII)
characters and the volume language is either not set or is set to one of these:
– C (POSIX)
– en
– en_US
– [Link]-8
– en_US.UTF-8
v Your data has directory and file names that contain non-English characters, and
the volume language is set to the proper locale (for example, [Link]-8 or de for
German).

Tip: Using the UTF-8 version of the volume language setting is more efficient in
terms of Tivoli Storage Manager server processing and table of contents storage
space.
v You only use CIFS to create and access your data.

File level restore from a directory-level backup image

File-level restore is supported for directory-level backup images.

As with a NAS (network attached storage) file system backup, a table of contents
(TOC) is created during a directory-level backup and you are able to browse the
files in the image, using the Web client. The default is that the files are restored to
the original location. During a file-level restore from a directory-level backup,
however, you can either select a different file system or another virtual file space
name as a destination.

For a TOC of a directory level backup image, the path names for all files are
relative to the directory specified in the virtual file space definition, not the root of
the file system.

Directory-level backup and restore

If you have a large NAS (network attached storage) file system, initiating a backup
at a directory level will reduce backup and restore times and provide more
flexibility in configuring your NAS backups. By defining virtual file spaces, a file
system backup can be partitioned among several NDMP backup operations and
multiple tape drives. You can also use different backup schedules to back up
sub-trees of a file system.

The virtual file space name cannot be identical to any file system on the NAS
node. If a file system is created on the NAS device with the same name as a virtual
file system, a name conflict will occur on the Tivoli Storage Manager server when
the new file space is backed up. See the Administrator's Reference for more
information about virtual file space mapping commands.

Note: Virtual file space mappings are only supported for NAS nodes.

Chapter 9. Using NDMP for operations with NAS file servers 235
 Directory-level backup and restore for NDMP operations
The DEFINE VIRTUALFSMAPPING command maps a directory path of a NAS (network
attached storage) file server to a virtual file space name on the Tivoli Storage
Manager server. After a mapping is defined, you can conduct NAS operations such
as BACKUP NODE and RESTORE NODE, using the virtual file space names as if they
were actual NAS file spaces.

To start a backup of the directory, issue the BACKUP NODE command specifying the
virtual file space name instead of a file space name. To restore the directory subtree
to the original location, run the RESTORE NODE command and specify the virtual file
space name.

Virtual file space definitions can also be specified as the destination in a RESTORE
NODE command. This allows you restore backup images (either file system or
directory) to a directory on any file system of the NAS device.

You can use the Web client to select files for restore from a directory-level backup
image because the Tivoli Storage Manager client treats the virtual file space names
as NAS file spaces.

Backing up and restoring with snapshots

NDMP directory level backup gives you the ability to back up user created
snapshots of a NAS file system; those are then stored as subdirectories. The
snapshots can be taken at any time, and the backup to tape can be deferred to a
more convenient time.

About this task

For example, to backup a snapshot created for a NetApp file system, perform the
following:

Procedure
1. On the console for the NAS device, issue the command to create the snapshot.
SNAP CREATE is the command for a NetApp device.
snap create vol2 february17

This command creates a snapshot named FEBRUARY 17 of the /vol/vol2 file

system. The physical location for the snapshot data is in the directory
/vol/vol2/.snapshot/february17. The stored location for snapshot data is dependent
on the NAS vendor implementation. For NetApp, the SNAP LIST command
can be used to display all snapshots for a given file system.
2. Define a virtual file space mapping definition on the Tivoli Storage Manager
server for the snapshot data created in the previous step.
define virtualfsmapping nas1 /feb17snapshot /vol/vol2 /.snapshot/february17

This creates a virtual file space mapping definition named /feb17snapshot.

3. Back up the virtual file space mapping.
backup node nas1 /feb17snapshot mode=full toc=yes
4. After the backup is created, you can either restore the entire snapshot image or
restore an individual file. Before restoring the data you can create a virtual file
space mapping name for the target directory. You can select any file system
name as a target. The target location in this example is the directory
/feb17snaprestore on the file system /vol/vol1.

236 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 define virtualfsmapping nas1 /feb17snaprestore /vol/vol1 /feb17snaprestore
5. Issue the restore of the snapshot backup image.
restore node nas1 /feb17snapshot /feb17snaprestore

This restores a copy of the /vol/vol2 file system to the directory

/vol/vol1/feb17snaprestore in the same state as when the snapshot was
created in the first step.

Backup and restore using NetApp SnapMirror to Tape feature

You can back up large NetApp file systems by using the NetApp SnapMirror to
Tape feature. Using a block-level copy of data for backup, the SnapMirror to Tape
method is faster than a traditional Network Data Management Protocol (NDMP)
full backup and can be used when NDMP full backups are impractical.

Use the NDMP SnapMirror to Tape feature as a disaster recovery option for
copying large NetAppfile systems to auxiliary storage. For most NetAppfile
systems, use the standard NDMP full or differential backup method.

Using a parameter option on the BACKUP NODE and RESTORE NODE commands, you
can back up and restore file systems by using SnapMirror to Tape. There are
several limitations and restrictions on how SnapMirror images can be used.
Consider the following guidelines before you use it as a backup method:
v You cannot initiate a SnapMirror to Tape backup or restore operation from the
Tivoli Storage Manager Operations Center, web client, or command-line client.
v You cannot perform differential backups of SnapMirror images.
v You cannot perform a directory-level backup using SnapMirror to Tape, thus
Tivoli Storage Manager does not permit an SnapMirror to Tape backup operation
on a server virtual file space.
v You cannot perform an NDMP file-level restore operation from SnapMirror to
Tape images. Therefore, a table of contents is never created during SnapMirror
to Tape image backups.
v At the start of a SnapMirror to Tape copy operation, the file server generates a
snapshot of the file system. NetAppprovides an NDMP environment variable to
control whether this snapshot should be removed at the end of the SnapMirror
to Tape operation. Tivoli Storage Manager always sets this variable to remove
the snapshot.
v After a SnapMirror to Tape image is retrieved and copied to a NetAppfile
system, the target file system is left configured as a SnapMirror partner.
NetAppprovides an NDMP environment variable to control whether this
SnapMirror relationship should be broken. Tivoli Storage Manager always
"breaks" the SnapMirror relationship during the retrieval. After the restore
operation is complete, the target file system is in the same state as that of the
original file system at the point-in-time of backup.

See the BACKUP NODE and RESTORE NODE commands in the Administrator's Reference
for more information about SnapMirror to Tape feature.

Chapter 9. Using NDMP for operations with NAS file servers 237
NDMP backup operations using Celerra file server integrated
checkpoints
When the Tivoli Storage Manager server initiates an NDMP backup operation on a
Celerra data mover, the backup of a large file system might take several hours to
complete. Without Celerra integrated checkpoints enabled, any changes occurring
on the file system are written to the backup image.

As a result, the backup image includes changes made to the file system during the
entire backup operation and is not a true point-in-time image of the file system.

If you are performing NDMP backups of Celerra file servers, you should upgrade
the operating system of your data mover to Celerra file server version T5.5.25.1 or
later. This version of the operating system allows enablement of integrated
checkpoints for all NDMP backup operations from the Celerra Control
Workstation. Enabling this feature ensures that NDMP backups represent true
point-in-time images of the file system that is being backed up.

Refer to the Celerra file server documentation for instructions on enabling

integrated checkpoints during all NDMP backup operations.

If your version of the Celerra file server operating system is earlier than version
T5.5.25.1 and if you use NDMP to back up Celerra data movers, you should
manually generate a snapshot of the file system using Celerra's command line
checkpoint feature and then initiate an NDMP backup of the checkpoint file system
rather than the original file system.

Refer to the Celerra file server documentation for instructions on creating and
scheduling checkpoints from the Celerra control workstation.

Replicating NAS nodes with NDMP backup data

You can replicate a network-attached storage (NAS) node that contains NDMP
(network data management protocol) backup data.

Only NDMP backup data in NATIVE data format storage pools can be replicated.
You cannot replicate NDMP images that are stored in storage pools that has the
following data formats:
v NETAPPDUMP
v CELERRADUMP
v NDMPDUMP

A differential NDMP backup image requires the associated full backup to be

restored. Therefore, to replicate a differential NDMP backup image, both full and
differential images must be in NATIVE data format storage pools. A full NDMP
image in a NATIVE data format storage pool is always eligible for replication. If an
NDMP image is replicated, its associated Table of Contents data is also replicated.

When an NDMP image is replicated to a replication target server, a data mover

definition must be created on the target replication server. To create a data mover
definition, issue the DEFINE DATAMOVER command to restore the data.

238 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Chapter 10. Managing storage pools and volumes
Logical storage pools and storage volumes are the principal components in the
Tivoli Storage Manager model of data storage. By manipulating the properties of
these objects, you can optimize the usage of storage devices.

About this task

When you configure devices so that the server can use them to store client data,
you create storage pools and storage volumes. The procedures for configuring
devices use the set of defaults that provides for storage pools and volumes. The
defaults can work well. However, you might have specific requirements not met by
the defaults. There are three common reasons to change the defaults:
v Optimize and control storage device usage by arranging the storage hierarchy
and tuning migration through the hierarchy (next storage pool, migration
thresholds).
v Reuse tape volumes through reclamation. Reuse is also related to policy and
expiration.
v Keep a client's files on a minimal number of volumes (collocation).

You can also make other adjustments to tune the server for your systems. See the
following sections to learn more. For some quick tips, see “Task tips for storage
pools” on page 251.

Concepts
“Storage pools” on page 240
“Storage pool volumes” on page 252
“Access modes for storage pool volumes” on page 259
“Storage pool hierarchies” on page 260
“Migrating files in a storage pool hierarchy” on page 271
“Caching in disk storage pools” on page 284
“Writing data simultaneously to primary, copy, and active-data pools” on page 331
“Keeping client files together using collocation” on page 357
“Reclaiming space in sequential-access storage pools” on page 367
“Estimating space needs for storage pools” on page 379

Tasks
“Defining storage pools” on page 245
“Preparing volumes for random-access storage pools” on page 255
“Preparing volumes for sequential-access storage pools” on page 255
“Defining storage pool volumes” on page 256
“Updating storage pool volumes” on page 257
“Setting up a storage pool hierarchy” on page 261
“Monitoring storage-pool and volume usage” on page 382
“Monitoring the use of storage pool volumes” on page 384

© Copyright IBM Corp. 1993, 2013 239

 Tasks
“Moving data from one volume to another volume” on page 400
“Moving data belonging to a client node” on page 406
“Renaming storage pools” on page 409
“Defining copy storage pools and active-data pools” on page 409
“Deleting storage pools” on page 413
“Deleting storage pool volumes” on page 414

For details about devices, see:

Chapter 4, “Magnetic disk devices,” on page 63
Chapter 6, “Configuring storage devices,” on page 85

The examples in topics show how to perform tasks using the Tivoli Storage
Manager command-line interface. For information about the commands, see the
Administrator's Reference, or issue the HELP command from the command line of a
Tivoli Storage Manager administrative client.

Storage pools
A storage pool is a collection of storage volumes. A storage volume is the basic
unit of storage, such as allocated space on a disk or a single tape cartridge. The
server uses the storage volumes to store backed-up, archived, or space-managed
files.

The server provides three types of storage pools that serve different purposes:
primary storage pools, copy storage pools, and active-data pools. You can arrange
primary storage pools in a storage hierarchy. The group of storage pools that you set
up for the Tivoli Storage Manager server to use is called server storage.

Primary storage pools

When a user tries to restore, retrieve, recall, or export file data, the requested file is
obtained from a primary storage pool, if possible. Primary storage pool volumes
are always located onsite.

To prevent a single point of failure, create separate storage pools for backed-up
and space-managed files. This also includes not sharing a storage pool in either
storage pool hierarchy. Consider setting up a separate, random-access disk storage
pool to give clients fast access to their space-managed files.

Restriction: Backing up a migrated, space-managed file might result in an error if

the destination for the backup is the same storage pool as the storage pool where
the space-managed file currently exists.

A primary storage pool can use random-access storage (DISK device class) or
sequential-access storage (for example, tape or FILE device classes).

240 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Copy storage pools
Copy storage pools contain active and inactive versions of data that is backed up
from primary storage pools. Copy storage pools provide a means of recovering
from disasters or media failures.

For example, when a client attempts to retrieve a file and the server detects an
error in the file copy in the primary storage pool, the server marks the file as
damaged. At the next attempt to access the file, the server can obtain the file from
a copy storage pool.

You can move copy storage pool volumes offsite and still have the server track the
volumes. Moving copy storage pool volumes offsite provides a means of
recovering from an onsite disaster.

A copy storage pool can use only sequential-access storage (for example, a tape
device class or FILE device class).

Remember:
v You can back up data from a primary storage pool defined with the NATIVE,
NONBLOCK, or any of the NDMP formats (NETAPPDUMP, CELERRADUMP,
or NDMPDUMP). The target copy storage pool must have the same data format
as the primary storage pool.
v You cannot back up data from a primary storage pool that is defined with a
CENTERA device class.

For details about copy storage pools, see:

v “Restoring storage pools and storage pool volumes” on page 907
v “Backing up primary storage pools” on page 883
v “Scenario: Recovering a lost or damaged storage pool volume” on page 910
v “Ensuring the integrity of files” on page 897
v “Backing up the data in a storage hierarchy” on page 265
v “Setting up copy storage pools and active-data pools” on page 266
v “Backing up primary storage pools” on page 883

Active-data pools
An active-data pool contains only active versions of client backup data. Active-data
pools are useful for fast client restores, reducing the number of onsite or offsite
storage volumes, or reducing bandwidth when copying or restoring files that are
vaulted electronically in a remote location.

Data that is migrated by hierarchical storage management (HSM) clients and

archive data are not permitted in active-data pools. As updated versions of backup
data continue to be stored in active-data pools, older versions are deactivated and
removed during reclamation processing.

Restoring a primary storage pool from an active-data pool might cause some or all
inactive files to be deleted from the database if the server determines that an
inactive file must be replaced but cannot find it in the active-data pool. As a best
practice and to protect your inactive data, create a minimum of two storage pools:
one active-data pool, which contains only active data, and one copy storage pool,
which contains both active and inactive data. You can use the active-data pool
volumes to restore critical client node data, and afterward you can restore the

Chapter 10. Managing storage pools and volumes 241

 primary storage pools from the copy storage pool volumes. Active-data pools must
not be considered for recovery of a primary pool or volume unless the loss of
inactive data is acceptable.

Active-data pools can use any type of sequential-access storage (for example, a
tape device class or FILE device class). However, the precise benefits of an
active-data pool depend on the specific device type that is associated with the
pool. For example, active-data pools associated with a FILE device class are ideal
for fast client restores because FILE volumes do not have to be physically mounted
and because the server does not have to position past inactive files that do not
have to be restored. In addition, client sessions restoring from FILE volumes in an
active-data pool can access the volumes concurrently, which also improves restore
performance.

Active-data pools that use removable media, such as tape, offer similar benefits.
Although tapes must be mounted, the server does not have to position past
inactive files. However, the primary benefit of using removable media in
active-data pools is the reduction of the number of volumes that are used for
onsite and offsite storage. If you vault data electronically to a remote location, an
active-data pool that is associated with a SERVER device class can save bandwidth
by copying and restoring only active data.

Remember:
v The server does not attempt to retrieve client files from an active-data pool
during a point-in-time restore. Point-in-time restores require both active and
inactive file versions. Active-data pools contain only active file versions. For
optimal efficiency during point-in-time restores and to avoid switching between
active-data pools and primary or copy storage pools, the server retrieves both
active and inactive versions from the same storage pool and volumes.
v You cannot copy active data to an active-data pool from a primary storage pool
that is defined with the NETAPPDUMP, the CELERRADUMP, or the
NDMPDUMP data format.
v You cannot copy active data from a primary storage pool that is defined with a
CENTERA device class.

For details about active-data pools, see:

v “Backing up the data in a storage hierarchy” on page 265
v “Setting up copy storage pools and active-data pools” on page 266
v “Copying active versions of client backup data to active-data pools”
v “Selection for active file versions” on page 243

Copying active versions of client backup data to active-data

pools
To copy active versions of client backup files from primary storage pools to
active-data pools, you can issue the COPY ACTIVEDATA command or you can use
the simultaneous-write function. The simultaneous-write function automatically
writes active backup data to active-data pools at the same time that the backup
data is written to a primary storage pool.

About this task

You can issue the COPY ACTIVEDATA command either manually or in an

administrative schedule or maintenance script.

242 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 Regardless whether you use the COPY ACTIVEDATA command or the
simultaneous-write function, the Tivoli Storage Manager server writes data to an
active-data pool only if the data belongs to a node that is a member of a policy
domain that specifies the active-data pool as the destination for active data.

Restriction: You cannot use the BACKUP STGPOOL command for active-data pools.

Selection for active file versions

The Tivoli Storage Manager uses a search order to locate active file versions.

During client sessions and processes that require active file versions, the Tivoli
Storage Manager server searches certain types of storage pools, if they exist.
1. An active-data pool that is associated with a FILE device class
2. A random-access disk (DISK) storage pool
3. A primary or copy storage pool that is associated with a FILE device class
4. A primary, copy, or active-data pool that is associated with onsite or offsite
removable tape media

Even though the list implies a selection order, the server might select a volume
with an active file version from a storage pool lower in the order if a volume
higher in the order cannot be accessed because of the requirements of the session
or process, volume availability, or contention for resources such as mount points,
drives, and data.

Example: Setting up server storage

All the data in four primary storage pools is backed up to one copy storage pool.
Active versions of data are stored in an active-data pool.

About this task

Figure 19 on page 244 shows one way to set up server storage. In this example, the
storage that is defined for the server includes:
v Three disk storage pools, which are primary storage pools: ARCHIVE, BACKUP,
and HSM
v One primary storage pool that consists of tape cartridges
v One copy storage pool that consists of tape cartridges
v One active-data pool that consists of FILE volumes for fast client restore

Policies that are defined in management classes direct the server to store files from
clients in the ARCHIVE, BACKUP, or HSM disk storage pools. An extra policy
specifies the following:
v A select group of client nodes that requires fast restore of active backup data
v The active-data pool as the destination for the active-data belonging to these
nodes
v The ARCHIVE, BACKUP, or HSM disk storage pools as destinations for archive,
backup (active and inactive versions), and space-managed data

For each of the three disk storage pools, the tape primary storage pool is next in
the hierarchy. As the disk storage pools fill, the server migrates files to tape to
make room for new files. Large files can go directly to tape. For more information
about setting up a storage hierarchy, see “Storage pool hierarchies” on page 260.

Chapter 10. Managing storage pools and volumes 243

 For more information about backing up primary storage pools, see “Backing up
primary storage pools” on page 883.

Server Server storage Onsite storage

primary storage pools active-data pools

Active
HSM backup
data only
Backup
Disk storage
Archive pool (FILE)

Disk storage pools

Archive, Offsite storage
backup copy storage pools
(active and
inactive data)

Database and Tape storage pool

Tape storage pool
recovery log

Figure 19. Example of server storage

To set up this server storage hierarchy, do the following:

Procedure
1. Define the three disk storage pools, or use the three default storage pools that
are defined when you install the server. Add volumes to the disk storage pools
if you did not already do so.
For more information, see
“Configuring random access volumes on disk devices” on page 69
2. Define policies that direct the server to initially store files from clients in the
disk storage pools. To do this, you define or change management classes and
copy groups so that they point to the storage pools as destinations. Then,
activate the changed policy. See “Changing policy” on page 478 for details.
Define another policy that specifies the active-data pool that you create as the
destination for active data.
3. Assign nodes to the domains. Nodes whose active data you want to restore
quickly must be assigned to the domain that specifies the active-data pool.
4. Attach one or more tape devices, or a tape library, to your server system.
For more information, see:
“Defining storage pools” on page 245
Chapter 6, “Configuring storage devices,” on page 85
5. Update the disk storage pools so that they point to the tape storage pool as the
next storage pool in the hierarchy. See “Example: Updating storage pools” on
page 250.
6. Define a copy storage pool and an active-data pool. The copy storage pool can
use the same tape device or a different tape device as the primary tape storage
pool. The active-data pool uses sequential-access disk storage (a FILE-type
device class) for fast client restores. See “Defining copy storage pools and
active-data pools” on page 409.
7. Set up administrative schedules or a script to back up the disk storage pools
and the tape storage pool to the copy storage pool. Use the same or different

244 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 schedules or scripts to copy active data to the active-data pool. Send the copy
storage pool volumes off-site for safekeeping. See “Backing up primary storage
pools” on page 883.

Defining storage pools

To optimize data storage, you can specify various properties when you define or
update a storage pool using the DEFINE STGPOOL and UPDATE STGPOOL commands.

About this task

Tip: When you define or update storage pools that use LTO Ultrium media,
special considerations might apply.

Task Required Privilege Class

Define storage pools System
Update storage pools System or unrestricted storage

Properties of storage pool definitions

You can define storage pools using a wide range of properties to control how data
is stored. Each storage pool represents one type of media as specified in the
device-class parameter.

When you define a primary storage pool, be prepared to specify some or all of the
information that is shown in Table 19. Most of the information is optional. Some
information applies only to random-access storage pools or only to
sequential-access storage pools. Required parameters are marked.
Table 19. Information for defining a storage pool
Type of
Information Explanation Storage Pool
Storage pool name The name of the storage pool. random,
sequential
(Required)
Device class The name of the device class assigned for the storage pool. random,
sequential
(Required)
Pool type The type of storage pool (primary or copy). The default is to define a random,
primary storage pool. A storage pool's type cannot be changed after it has sequential
been defined.
Maximum number of When you specify a value greater than zero, the server dynamically sequential
scratch volumes 2 acquires scratch volumes when needed, up to this maximum number.

(Required for sequential For automated libraries, set this value equal to the physical capacity of the
access) library. For details, see:
“Adding scratch volumes to automated library devices” on page 138

Chapter 10. Managing storage pools and volumes 245

Table 19. Information for defining a storage pool (continued)
Type of
Information Explanation Storage Pool
Access mode Defines access to volumes in the storage pool for user operations (such as random,
backup and restore) and system operations (such as reclamation and server sequential
migration). Possible values are:
Read/Write
User and system operations can read from or write to the
volumes.
Read-Only
User operations can read from the volumes, but not write. Server
processes can move files within the volumes in the storage pool.
However, no new writes are permitted to volumes in the storage
pool from volumes outside the storage pool.
Unavailable
User operations cannot get access to volumes in the storage pool.
No new writes are permitted to volumes in the storage pool from
other volumes outside the storage pool. However, system
processes (like reclamation) are permitted to move files within the
volumes in the storage pool.
1 2
Maximum file size To exclude large files from a storage pool, set a maximum file size. The random,
maximum file size applies to the size of a physical file (a single client file sequential
or an aggregate of client files).

Do not set a maximum file size for the last storage pool in the hierarchy
unless you want to exclude very large files from being stored in server
storage.
Cyclic Redundancy Check Specifies whether the server uses CRC to validate storage pool data during random,
(CRC) 1 audit volume processing. For additional information see “Data validation sequential
during audit volume processing” on page 890.

Using the CRC option in conjunction with scheduling audit volume

processing continually ensures the integrity of data stored in your storage
hierarchy. If you always want your storage pool data validated, set your
primary storage pool crcdata definition to YES.
Name of the next storage Specifies the name of the next storage pool in the storage pool hierarchy, random,
pool 1 2 where files can be migrated or where files are stored that exceed the sequential
maximum size for this storage pool. See “Storage pool hierarchies” on
page 260.
1 2
Migration thresholds Specifies a percentage of storage pool occupancy at which the server random,
begins migrating files to the next storage pool (high threshold) and the sequential
percentage when migration stops (low threshold). See “Migrating files in a
storage pool hierarchy” on page 271.
1 2
Migration processes Specifies the number of concurrent processes to use for migrating files random,
from this storage pool. See “Migrating disk storage pools” on page 273 sequential
and “Specifying multiple concurrent migration processes” on page 282.
1 2
Migration delay Specifies the minimum number of days a file must remain in a storage random,
pool before it is eligible for migration. See “Keeping files in a storage sequential
pool” on page 278 and “How the server migrates files from
sequential-access storage pools” on page 280.

246 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Table 19. Information for defining a storage pool (continued)
Type of
Information Explanation Storage Pool
Continue migration Specifies whether migration of files should continue even if files do not random,
process 1 2 meet the requirement for migration delay. This setting is used only when sequential
the storage pool cannot go below the low migration threshold without
moving additional files. See “Keeping files in a storage pool” on page 278
and “How the server migrates files from sequential-access storage pools”
on page 280.
Cache Enables or disables cache. When cache is enabled, copies of files migrated random
by the server to the next storage pool are left on disk after the migration.
In this way, a retrieval request can be satisfied quickly. See “Caching in
disk storage pools” on page 284.
2
Collocation With collocation enabled, the server tries to keep all files belonging on a sequential
minimal number of sequential-access storage volumes. The files can belong
to a single client node, a group of client nodes, a client file space, or a
group of file spaces. See “Keeping client files together using collocation”
on page 357.
1 2
Reclamation threshold Specifies what percentage of reclaimable space can accumulate on a sequential
volume before the server initiates a space reclamation process for the
volume. See “Reclamation thresholds” on page 370.
1 2
Reclamation processes Specifies the number of concurrent processes to use for reclaiming the sequential
volumes in a storage pool. See “Optimizing drive usage using multiple
concurrent reclamation processes” on page 371.
Off-site reclaim limit Specifies the number of off-site volumes to have their space reclaimed sequential
during reclamation for a storage pool. See “Reclamation of offsite
volumes” on page 375.
Reclamation storage pool Specifies the name of the storage pool to be used for storing data from sequential
1 2
volumes being reclaimed in this storage pool. Use for storage pools whose
device class only has one drive or mount point. See “Reclaiming volumes
in a storage pool with one drive” on page 372.
2
Reuse delay period Specifies the number of days that must elapse after all of the files have sequential
been deleted from a volume, before the volume can be rewritten or
returned to the scratch pool. See “Delaying reuse of volumes for recovery
purposes” on page 886.
1 2
Overflow location Specifies the name of a location where volumes are stored when they are sequential
ejected from an automated library by the MOVE MEDIA command.

Use for a storage pool that is associated with an automated library or an

external library.

For details, see:

“Returning reclaimed volumes to a library” on page 137
2
Data Format The format in which data will be stored. NATIVE is the default data sequential
format. NETAPPDUMP and NONBLOCK are examples of other data
formats.

Chapter 10. Managing storage pools and volumes 247

Table 19. Information for defining a storage pool (continued)
Type of
Information Explanation Storage Pool
1 2
Copy Storage Pools Specifies the names of copy storage pools where the server simultaneously random,
writes data when a client backup, archive, import or migration operation sequential
stores data to the primary storage pool. The server writes the data
simultaneously to all listed copy storage pools. This option is restricted to
primary random-access storage pools or to primary sequential-access
storage pools that use the NATIVE or NONBLOCK data format. See the
Copy Continue entry and “Writing data simultaneously to primary, copy,
and active-data pools” on page 331 for related information.

Attention: The COPYSTGPOOLS parameter is not intended to replace the

BACKUP STGPOOL command. If you use the simultaneous-write function,
ensure that the copy of the primary storage pool is complete by regularly
issuing the BACKUP STGPOOL command. Failure to do so could result in
the inability to recover the primary storage pool data if the primary
storage pool becomes damaged or lost.
1 2
Copy Continue Specifies how the server should react to a copy storage pool write failure sequential
for any of the copy storage pools listed in the COPYSTGPOOLS parameter.
With a value of YES, during a write failure, the server will exclude the
failing copy storage pool from any further writes while that specific client
session is active. With a value of NO, during a write failure, the server will
fail the entire transaction including the write to the primary storage pool.

This option has no effect on active-data pools.

1 2
Active-data pools Specifies the names of active-data pools where the server simultaneously random,
writes active versions of client node data during backups. The server sequential
writes the data simultaneously to all listed active-data pools. This option is
restricted to primary random-access storage pools or to primary
sequential-access storage pools that use the NATIVE or NONBLOCK data
format. Nodes whose data is to be written to an active-data pool during a
simultaneous-write operation must be members of a policy domain that
specifies the active-data pool as the destination for active backup data.

Attention: The ACTIVEDATAPOOLS parameter is not intended to

replace the COPY ACTIVEDATA command. If you use the
simultaneous-write function, ensure that the copy of active backup data is
complete by regularly issuing the COPY ACTIVEDATA command. If you
do not issue the COPY ACTIVEDATA command regularly and you do not
have copy storage pools, you might not be able to recover any of the data
in a primary storage pool if the primary storage pool becomes damaged or
lost.
Shredding Specifies whether data is physically overwritten when it is deleted. After random
client data is deleted, it might still be possible to recover it. For sensitive
data, this condition is a potential security exposure. Shredding the deleted
data increases the difficulty of discovering and reconstructing the data
later. For more information, including how to set up shred pools and how
shredding interacts with other command parameters, see “Securing
sensitive client data” on page 539.

1
This information is not available for sequential-access storage pools that use the following data formats:
v CELERRADUMP
v NDMPDUMP
v NETAPPDUMP
2
This information is not available or is ignored for Centera sequential-access storage pools.

248 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Example: Defining storage pools
An engineering department requires a separate storage hierarchy. You want the
department's backed-up files to go to a disk storage pool. When that pool fills, you
want the files to migrate to a tape storage pool.

About this task

You want the storage pools to have the following characteristics:

v Disk primary storage pool
– The pool named ENGBACK1 is the storage pool for the engineering
department.
– The size of the largest file that can be stored is five MB. Files larger than five
MB are stored in the tape storage pool.
– Files migrate from the disk storage pool to the tape storage pool when the
disk storage pool is 85% full. File migration to the tape storage pool stops
when the disk storage pool is down to 40% full.
– The access mode is the default, read/write.
– Cache is used.
v Tape primary storage pool
– The name of the pool is BACKTAPE.
– The pool uses the device class TAPE, which has already been defined.
– No limit is set for the maximum file size, because this is the last storage pool
in the hierarchy.
– To group files from the same client on a small number of volumes, use
collocation at the client node level.
– Use scratch volumes for this pool, with a maximum number of 100 volumes.
– The access mode is the default, read/write.
– Use the default for reclamation: Reclaim a partially full volume (to allow tape
reuse) when 60% of the volume's space can be reclaimed.

You can define the storage pools in a storage pool hierarchy from the top down or
from the bottom up. Defining the hierarchy from the bottom up requires fewer
steps. To define the hierarchy from the bottom up, complete the following steps:

Procedure
1. Define the storage pool named BACKTAPE with the following command:
define stgpool backtape tape
description=’tape storage pool for engineering backups’
maxsize=nolimit collocate=node maxscratch=100
2. Define the storage pool named ENGBACK1 with the following command:
define stgpool engback1 disk
description=’disk storage pool for engineering backups’
maxsize=5m nextstgpool=backtape highmig=85 lowmig=40

Results

Restrictions:
v You cannot establish a chain of storage pools that lead to an endless loop. For
example, you cannot define StorageB as the next storage pool for StorageA, and
then define StorageA as the next storage pool for StorageB.
v The storage pool hierarchy includes only primary storage pools, not copy
storage pools or active-data pools.

Chapter 10. Managing storage pools and volumes 249

 v If a storage pool uses the data format NETAPPDUMP, CELERRADUMP, or
NDMPDUMP, the server will not perform any of the following functions:
– Migration
– Reclamation
– Volume audits
– Data validation
– Simultaneous-write operations
For more information about data formats, see Chapter 9, “Using NDMP for
operations with NAS file servers,” on page 203.
v The Tivoli Storage Manager server does not support the following functions for
Centera storage pools:
– Data-movement operations:
- Moving node data into or out of a Centera storage pool.
- Migrating data into or out of a Centera storage pool.
- Reclaiming a Centera storage pool.
– Backup operations:
- Backing up a Centera storage pool.
- Using a Centera device class to back up a database.
- Backing up a storage pool to a Centera storage pool.
- Copying active data to an active-data pool.
– Restore operations:
- Restoring data from a copy storage pool or an active-data pool to a Centera
storage pool.
- Restoring volumes in a Centera storage pool.
– Other:
- Exporting data to a Centera device class or importing data from a Centera
device class; however, files stored in Centera storage pools can be exported
and files being imported can be stored on Centera.
- Using a Centera device class for creating backup sets; however, files stored
in Centera storage pools can be sent to backup sets.
- Defining Centera volumes.
- Using a Centera device class as the target of volume history, device
configuration, trace logs, error logs, or query output files.

Example: Updating storage pools

You decide to increase the maximum size of a physical file that can be stored in
the ENGBACK1 disk storage pool.

About this task

In this example, the ENGBACK1 disk storage pool is defined as shown in

“Example: Defining storage pools” on page 249. To increase the maximum size of a
physical file that can be stored in the storage pool, use the following command:
update stgpool engback1 maxsize=100m

Restrictions:
v You cannot use this command to change the data format for a storage pool.
v For storage pools that have the NETAPPDUMP, the CELERRADUMP, or the
NDMPDUMP data format, you can modify the following parameters only:

250 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 – ACCESS
– COLLOCATE
– DESCRIPTION
– MAXSCRATCH
– REUSEDELAY

Task tips for storage pools

Tivoli Storage Manager provides many functions, such as migration and
reclamation, for optimizing data-storage operations. To take advantage of these
functions, you can create specialized storage pools or specify certain properties in
your storage pool definitions.

Table 20 gives tips on how to accomplish some tasks that are related to storage
pools.
Table 20. Task tips for storage pools
For this Goal Do This For More Information
Keep the data for a group of client Enable collocation for the storage “Keeping client files together using
nodes, a single client node, or a client pool. collocation” on page 357
file space on as few volumes as
possible.
Reduce the number of volume Disable collocation for the storage “Keeping client files together using
mounts needed to back up multiple pool. collocation” on page 357
clients.
Write data simultaneously to a Provide a list of copy storage pools “Writing data simultaneously to
primary storage pool and to copy and active-data pools when defining primary, copy, and active-data pools”
storage pools and active-data pools. the primary storage pool. on page 331
Specify how the server reuses tapes. Set a reclamation threshold for the “Reclaiming space in
storage pool. sequential-access storage pools” on
page 367
Optional: Identify a reclamation
storage pool
Move data from disk to tape Set a migration threshold for the “Migrating disk storage pools” on
automatically as needed. storage pool. page 273

Identify the next storage pool.

Move data from disk to tape Set a migration threshold for the “Migrating disk storage pools” on
automatically based on how storage pool. page 273
frequently users access the data or
how long the data has been in the Identify the next storage pool.
storage pool.
Set the migration delay period.
Improve client restore performance Implement a storage pool associated
using concurrent access to FILE with the FILE device type. “Defining storage pools” on page 245
volumes.
“Setting up copy storage pools and
active-data pools” on page 266
Back up your storage pools. Implement a copy storage pool. “Setting up copy storage pools and
active-data pools” on page 266
Copy active data from a primary Implement an active-data pool. “Setting up copy storage pools and
storage pool. active-data pools” on page 266

Chapter 10. Managing storage pools and volumes 251

Table 20. Task tips for storage pools (continued)
For this Goal Do This For More Information
Have clients back up directly to a Define a sequential-access storage “Defining storage pools” on page 245
tape storage pool. pool that uses a tape device class.
“Changing policy” on page 478
Change the policy that the clients use,
so that the backup copy group points
to the tape storage pool as the
destination.
Make the best use of available tape Specify multiple concurrent processes.
drives and FILE volumes during “Optimizing drive usage using
reclamation and migration. multiple concurrent reclamation
processes” on page 371

“Specifying multiple concurrent

migration processes” on page 282
Ensure that reclamation completes Limit the number of off-site volumes
within the desired amount of time. to be reclaimed. “Reclamation of offsite volumes” on
page 375

“Starting reclamation manually or in

a schedule” on page 371
For storage pools associated with Use the DEFINE SPACETRIGGER
random-access and sequential-access and UPDATE SPACETRIGGER “Preparing volumes for
disk (DISK and FILE device classes), commands to specify the number and random-access storage pools” on page
automatically create private volumes size of volumes. 255
and preassign them to specified
storage pools when predetermined “Defining storage pool volumes” on
space utilization thresholds have been page 256
reached.
For storage pools associated with Use the DEFINE VOLUME command
random-access disk (DISK device to specify the size and number of “Preparing volumes for
class) and sequential-access disk volumes to be created. random-access storage pools” on page
(FILE device class), create and format 255
volumes using one command.
“Defining storage pool volumes” on
page 256

Storage pool volumes

Storage pool volumes are the physical media that are assigned to a storage pool.

The following list contains examples of a volume:

v Space that is allocated on a disk drive
v A tape cartridge
Storage pool volumes are either random access or sequential access, depending on
the device type of the device class to which the storage pool is assigned.

252 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Random-access storage pool volumes
Random-access storage pools consist of volumes on disk. Random-access storage
pools are always associated with the DISK device class. All volumes in this type of
storage pool have the same form.

A volume in a random-access storage pool is a fixed-size file that is created when

you define a volume for the storage pool or when you use space triggers to
automatically create volumes and assign them to specified storage pools.

For additional information, see:

“Preparing volumes for random-access storage pools” on page 255
“Requirements for disk systems” on page 63

Sequential-access storage pool volumes

Sequential-access volumes are volumes in data is accessed sequentially, one block
at a time, one after the other. Each volume defined in a sequential-access storage
pool must be the same type as the device class associated with the storage pool.

You can define volumes in a sequential-access storage pool or you can specify that
the server dynamically acquire scratch volumes. You can also use a combination of
defined and scratch volumes. What you choose depends on the amount of control
you want over individual volumes.

For information about preparing sequential-access volumes, see “Preparing

volumes for sequential-access storage pools” on page 255.

Types of sequential-access volumes

Each Tivoli Storage Manager sequential-access device type is associated with a
particular type of storage pool volume.

The following list contains some examples of sequential-access volumes:

v Tape cartridge
v File
Table 21 lists the types of volumes that are associated with each device type.
Table 21. Volume types
Device type Volume description Label
required
3590 IBM 3590 tape cartridge Yes
3592 IBM 3592 tape cartridge Yes
4MM 4 mm tape cartridge Yes
8MM 8 mm tape cartridge Yes
CENTERA A logical collection of files that are stored on the No
Centera storage device
DLT A digital linear tape Yes
ECARTRIDGE A cartridge tape that is used by a tape drive such as Yes
the StorageTek SD-3 or 9490 tape drive
FILE A file in the file system of an operating system No
GENERICTAPE A tape that is compatible with the drives that are Yes
defined to the device class

Chapter 10. Managing storage pools and volumes 253

 Table 21. Volume types (continued)
Device type Volume description Label
required
LTO IBM Ultrium tape cartridge Yes
NAS A tape drive that is used for NDMP backups by a Yes
network-attached storage (NAS) file server
REMOVABLEFILE A file on a removable medium. If the medium has Yes
two sides, each side is a separate volume.
SERVER One or more objects that are archived in the server No
storage of another server
VOLSAFE A StorageTek cartridge tape that is for write-once No
use on tape drives that are enabled for VolSafe
function.

Defined volumes
Use defined volumes when you want to control precisely which volumes are used
in the storage pool. Defined volumes can also be useful when you want to
establish a naming scheme for volumes.

You can also use defined volumes to reduce potential disk fragmentation and
maintenance overhead for storage pools associated with random-access and
sequential-access disk.

Scratch volumes
Use scratch volumes to enable the server to define a volume when needed and
delete the volume when it becomes empty. Using scratch volumes frees you from
the task of explicitly defining all of the volumes in a storage pool.

The server tracks whether a volume being used was originally a scratch volume.
Scratch volumes that the server acquired for a primary storage pool are deleted
from the server database when they become empty. The volumes are then available
for reuse by the server or other applications.

Scratch volumes in a copy storage pool or an active-data storage pool are handled
in the same way as scratch volumes in a primary storage pool, except for volumes
with the access value of off-site. If an off-site volume becomes empty, the server
does not immediately return the volume to the scratch pool. The delay prevents
the empty volumes from being deleted from the database, making it easier to
determine which volumes should be returned to the on-site location. The
administrator can query the server for empty off-site copy storage pool volumes or
active-data pool volumes, and return them to the on-site location. The volume is
returned to the scratch pool only when the access value is changed to
READWRITE, READONLY, or UNAVAILABLE.

For scratch volumes that were acquired in a FILE device class, the space that the
volumes occupied is freed by the server and returned to the file system.

254 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Preparing volumes for random-access storage pools
Volumes in random-access storage pools must be defined before the server can
access them.

About this task

Task Required Privilege Class
Define volumes in any storage pool System or unrestricted storage
Define volumes in specific storage pools System, unrestricted storage, or restricted
storage for those pools

Results

To prepare a volume for use in a random-access storage pool, define the volume.
For example, enter the following command to define a 21 MB volume named
stgvol.001 for the BACKUPPOOL storage pool that is located in the path
c:\program files\tivoli\tsm\server:
define volume backuppool ’c:\program files\tivoli\tsm\server\stgvol.001’
formatsize=21

If you do not specify a full path name for the volume name, the command uses the
path associated with the registry key of this server instance.

You can also define volumes in a single step using the DEFINE VOLUME
command. For example, to define ten, 5000 MB volumes in a random-access
storage pool that uses a DISK device class, you would enter the following
command:
define volume diskpool diskvol numberofvolumes=10 formatsize=5000

Tips:
1. For important disk-related information, see “Requirements for disk systems”
on page 63.
2. The file system where storage pool volumes are allocated can have an effect on
performance and reliability. For better performance in backing up and restoring
large numbers of small files, allocate storage pool volumes on a FAT file
system. To take advantage of the ability of the operating system to recover from
problems that can occur during I/O to a disk, allocate storage pool volumes on
NTFS.

You can also use a space trigger to automatically create volumes assigned to a
particular storage pool.

Preparing volumes for sequential-access storage pools

For most purposes, in a sequential-access storage pool, the server can use
dynamically acquired scratch volumes, volumes that you define, or a combination
of both.

About this task

For sequential-access storage pools with a FILE or SERVER device type, no labeling
or other preparation of volumes is necessary. For sequential-access storage pools
associated with device types other than a FILE or SERVER, you must prepare
volumes for use.

Chapter 10. Managing storage pools and volumes 255

 When the server accesses a sequential-access volume, it checks the volume name in
the header to ensure that the correct volume is being accessed. To prepare a
volume:

Procedure
1. Label the volume. Table 21 on page 253 shows the types of volumes that
require labels. You must label those types of volumes before the server can use
them.
For details, see:
“Labeling media” on page 129

Tip: When you use the LABEL LIBVOLUME command with drives in an
automated library, you can label and check in the volumes with one command.
2. For storage pools in automated libraries, use the CHECKIN LIBVOLUME
command to check the volume into the library. For details, see:
“Checking media into automated library devices” on page 130.
3. If you have not allowed scratch volumes in the storage pool, you must identify
the volume, by name, to the server. For details, see “Defining storage pool
volumes.”
If you allowed scratch volumes in the storage pool by specifying a value
greater than zero for the MAXSCRATCH parameter, you can let the server use
scratch volumes, identify volumes by name, or do both. See “Acquiring scratch
volumes dynamically” on page 257 for information about scratch volumes.

Results
Defining storage pool volumes
You can use defined volumes to control precisely which volumes are used in the
storage pool. Using defined volumes can also be useful when you want to
establish a naming scheme for volumes.

About this task

Task Required Privilege Class
Define volumes in any storage pool System or unrestricted storage
Define volumes in specific storage pools System, unrestricted storage, or restricted
storage for those pools

When you define a storage pool volume, you inform the server that the volume is
available for storing backup, archive, or space-managed data.

For a sequential-access storage pool, the server can use dynamically acquired
scratch volumes, volumes that you define, or a combination.

To define a volume named VOL1 in the ENGBACK3 tape storage pool, enter:
define volume engback3 vol1

Each volume that is used by a server for any purpose must have a unique name.
This requirement applies to all volumes, whether the volumes are used for storage
pools, or used for operations such as database backup or export. The requirement
also applies to volumes that are in different libraries but that are used by the same
server.

256 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 For storage pools associated with FILE device classes, you can define private
volumes in a single step by using the DEFINE VOLUME command. For example, to
define ten, 5000 MB volumes, in a sequential-access storage pool that uses a FILE
device class, you would enter the following command.
define volume filepool filevol numberofvolumes=10 formatsize=5000

For storage pools associated with the FILE device class, you can also use the
DEFINE SPACETRIGGER and UPDATE SPACETRIGGER commands to have the server
create volumes and assign them to a specified storage pool when predetermined
space-utilization thresholds are exceeded. One volume must be predefined.

Remember: You cannot define volumes for storage pools that are defined with a
Centera device class.

Acquiring scratch volumes dynamically

If you allow sequential-access storage pools to use scratch volumes, you do not
need to define volumes. You can control the maximum number of scratch volumes
that the server can request using the MAXSCRATCH parameter on the DEFINE
STGPOOL and UPDATE STGPOOL command.

About this task

To allow the storage pool to acquire volumes as needed, set the MAXSCRATCH
parameter to a value greater than zero. The server automatically defines the
volumes as they are acquired. The server also automatically deletes scratch
volumes from the storage pool when the server no longer needs them.

Before the server can use a scratch volume with a device type other than FILE or
SERVER, the volume must have a label.

Restriction: Tivoli Storage Manager only accepts tapes labeled with IBM standard
labels. IBM standard labels are similar to ANSI Standard X3.27 labels except that
the IBM standard labels are written in EBCDIC (extended binary coded decimal
interchange code). For a list of IBM media sales contacts who can provide
compatible tapes, go to the IBM Web site. If you are using non-IBM storage devices
and media, consult your tape-cartridge distributor.

For details about labeling, see “Preparing volumes for sequential-access storage
pools” on page 255.

Updating storage pool volumes

You can update a volume to reset an error state to an access mode of read/write.
You can also update a volume to change the its location in a sequential-access
storage pool. or to change the access mode of the volume, for example, if a tape
cartridge is moved off-site or is damaged.

About this task

Task Required Privilege Class
Update volumes System or operator

To change the properties of a volume that has been defined to a storage pool, issue
the UPDATE VOLUME command. For example, suppose you accidentally damage

Chapter 10. Managing storage pools and volumes 257

 a volume named VOL1. To change the access mode to unavailable so that the
server does not try to write or read data from the volume, issue the following
command:
update volume vol1 access=unavailable

For details about access modes, see “Access modes for storage pool volumes” on
page 259.

Volume properties that you can update

Update volume properties by changing the values of those properties in the
volume definition.

Table 22 lists volume properties that you can update.

Table 22. Information for updating a storage pool volume
Information Explanation
Volume name Specifies the name of the storage pool volume to be updated. You can
specify a group of volumes to update by using wildcard characters in
(Required) the volume name. You can also specify a group of volumes by
specifying the storage pool, device class, current access mode, or status
of the volumes you want to update. See the parameters that follow.
New access mode Specifies the new access mode for the volume (how users and server
processes such as migration can access files in the storage pool volume).
See “Access modes for storage pool volumes” on page 259 for
descriptions of access modes.

A random-access volume must be varied offline before you can change

its access mode to unavailable or destroyed. To vary a volume offline, use
the VARY command. See “Varying disk volumes online or offline” on
page 70.

If a scratch volume that is empty and has an access mode of off-site is

updated so that the access mode is read/write, read-only, or unavailable,
the volume is deleted from the database.
Location Specifies the location of the volume. This parameter can be specified
only for volumes in sequential-access storage pools.
Storage pool Restricts the update to volumes in the specified storage pool.
Device class Restricts the update to volumes in the specified device class.
Current access Restricts the update to volumes that currently have the specified access
mode mode.
Status Restricts the update to volumes with the specified status (online, offline,
empty, pending, filling, or full).
Preview Specifies whether you want to preview the update operation without
actually performing the update.

258 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Access modes for storage pool volumes
Access to a volume in a storage pool is determined by the access mode assigned to
that volume. You can manually change the access mode of a volume, or the server
can change the access mode based on what happens when it tries to access a
volume.

For example, if the server cannot write to a volume having read/write access
mode, the server automatically changes the access mode to read-only.

The following access modes apply to storage pool volumes:

Read/write
Allows files to be read from or written to a volume in the storage pool.
If the server cannot write to a read/write access volume, the server
automatically changes the access mode to read-only.
If a scratch volume that is empty and has an access mode of off-site is
updated so that the access mode is read/write, the volume is deleted from
the database.
Read-only
Allows files to be read from but not written to a disk or tape volume.
If a scratch volume that is empty and has an access mode of off-site is
updated so that the access mode is read-only, the volume is deleted from
the database.
Unavailable
Specifies that the volume is not available for any type of access by the
server.
You must vary offline a random-access volume before you can change its
access mode to unavailable. To vary a volume offline, use the VARY
command. See “Varying disk volumes online or offline” on page 70.
If a scratch volume that is empty and has an access mode of off-site is
updated so that the access mode is unavailable, the volume is deleted from
the database.
Destroyed
Specifies that a primary storage pool volume has been permanently
damaged. Neither users nor system processes (like migration) can access
files stored on the volume.
This access mode is used to indicate an entire volume that should be
restored using the RESTORE STGPOOL or RESTORE VOLUME command.
After all files on a destroyed volume are restored to other volumes, the
destroyed volume is automatically deleted from the database. See
Only volumes in primary storage pools can be updated to an access mode
of destroyed.
You must vary offline a random-access volume before you can change its
access mode to destroyed. To vary a volume offline, use the VARY
command. See “Varying disk volumes online or offline” on page 70. Once
you update a random-access storage pool volume to destroyed, you cannot
vary the volume online without first changing the access mode.
If you update a sequential-access storage pool volume to destroyed, the
server does not attempt to mount the volume.

Chapter 10. Managing storage pools and volumes 259

 If a volume contains no files and the UPDATE VOLUME command is used
to change the access mode to destroyed, the volume is deleted from the
database.
Offsite
Specifies that a copy storage pool volume or active-data pool volume is at
an off-site location and therefore cannot be mounted. Use this mode to
help you track volumes that are off-site. The server treats off-site volumes
differently, as follows:
v Mount requests are not generated for off-site volumes.
v Data can be reclaimed or moved from off-site volumes by retrieving files
from other storage pools.
v Empty, off-site scratch volumes are not deleted from the copy storage
pool or from the active-data pool.
You can only update volumes in a copy storage pool or an active-data pool
to off-site access mode. Volumes that have the device type of SERVER
(volumes that are actually archived objects stored on another Tivoli Storage
Manager server) cannot have an access mode of off-site.

Storage pool hierarchies

You can arrange storage pools in a storage hierarchies, which consist of at least one
primary storage pool to which a client node backs up, archives, or migrates data.
Typically, data is stored initially in a disk storage pool for fast client restores, and
then moved to a tape-based storage pool, which is slower to access but which has
greater capacity. The location of all data objects is automatically tracked within the
server database.

About this task

You can set up your devices so that the server automatically moves data from one
device to another, or one media type to another. The selection can be based on
characteristics such as file size or storage capacity. A typical implementation might
have a disk storage pool with a subordinate tape storage pool. When a client backs
up a file, the server might initially store the file on disk according to the policy for
that file. Later, the server might move the file to tape when the disk becomes full.
This action by the server is called migration. You can also place a size limit on files
that are stored on disk, so that large files are stored initially on tape instead of on
disk.

For example, your fastest devices are disks, but you do not have enough space on
these devices to store all data that needs to be backed up over the long term. You
have tape drives, which are slower to access, but have much greater capacity. You
define a hierarchy so that files are initially stored on the fast disk volumes in one
storage pool. This provides clients with quick response to backup requests and
some recall requests. As the disk storage pool becomes full, the server migrates, or
moves, data to volumes in the tape storage pool.

Another option to consider for your storage pool hierarchy is IBM 3592 tape
cartridges and drives, which can be configured for an optimal combination of
access time and storage capacity. For more information, see “Controlling
data-access speeds for 3592 volumes” on page 185.

Migration of files from disk to sequential storage pool volumes is particularly

useful because the server migrates all the files for a group of nodes or a single

260 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 node together. This gives you partial collocation for clients. Migration of files is
especially helpful if you decide not to enable collocation for sequential storage
pools. For details, see “Keeping client files together using collocation” on page 357.

Setting up a storage pool hierarchy

To establish a hierarchy, identify the next storage pool, sometimes called the
subordinate storage pool. The server migrates data to the next storage pool if the
original storage pool is full or unavailable.

About this task

Restrictions:
v You cannot establish a chain of storage pools that leads to an endless loop. For
example, you cannot define StorageB as the next storage pool for StorageA, and
then define StorageA as the next storage pool for StorageB.
v The storage pool hierarchy includes only primary storage pools. It does not
include copy storage pools or active-data pools. See “Backing up the data in a
storage hierarchy” on page 265.
v A storage pool must use the NATIVE or NONBLOCK data formats to be part of
a storage pool hierarchy. For example, a storage pool that uses the
NETAPPDUMP data format cannot be part of a storage pool hierarchy.

For detailed information about how migration between storage pools works, see
“Migrating files in a storage pool hierarchy” on page 271.

Example: Defining a storage pool hierarchy

You determined that an engineering department requires a separate storage
hierarchy. You set up policy so that the server initially stores backed up files for
this department to a disk storage pool. When that pool fills, you want the server to
migrate files to a tape storage pool.

About this task

You want the storage pools to have the following characteristics:

v Primary storage pool on disk
– Name the storage pool ENGBACK1.
– Limit the size of the largest file that can be stored to 5 MB. The server stores
files that are larger than 5 MB in the tape storage pool.
– Files migrate from the disk storage pool to the tape storage pool when the
disk storage pool is 85% full. File migration to the tape storage pool stops
when the disk storage pool is down to 40% full.
– Use caching, so that migrated files stay on disk until the space is needed for
other files.
v Primary storage pool on tape:
– Name the storage pool BACKTAPE.
– Use the device class TAPE, which has already been defined, for this storage
pool.
– Do not set a limit for the maximum file size, because this is the last storage
pool in the hierarchy.
– Use scratch volumes for this pool, with a maximum number of 100 volumes.

Chapter 10. Managing storage pools and volumes 261

 You can define the storage pools in a storage pool hierarchy from the top down or
from the bottom up. Defining the hierarchy from the bottom up requires fewer
steps. To define the hierarchy from the bottom up:

Procedure
1. Define the storage pool named BACKTAPE with the following command:
define stgpool backtape tape
description=’tape storage pool for engineering backups’
maxsize=nolimit collocate=node maxscratch=100
2. Define the storage pool named ENGBACK1 with the following command:
define stgpool engback1 disk
description=’disk storage pool for engineering backups’
maxsize=5M nextstgpool=backtape highmig=85 lowmig=40

Example: Updating a storage pool hierarchy

You already defined the ENGBACK1 disk storage pool. Now you decide to set up
a tape storage pool to which files from ENGBACK1 can migrate.

About this task

If you have already defined the storage pool at the top of the hierarchy, you can
update the storage hierarchy to include a new storage pool. You can update the
storage pool by using the UPDATE STGPOOL command.

To define the new tape storage pool and update the hierarchy:

Procedure
1. Define the storage pool named BACKTAPE with the following command:
define stgpool backtape tape
description=’tape storage pool for engineering backups’
maxsize=nolimit collocate=node maxscratch=100
2. Update the storage-pool definition for ENGBACK1 to specify that BACKTAPE is the
next storage pool defined in the storage hierarchy:
update stgpool engback1 nextstgpool=backtape

How the server groups files before storing

When client files are backed up or archived, the server can group them into an
aggregate of files. By controlling the size of aggregates, you can control the
performance of client operations.

The size of the aggregate depends on the sizes of the client files being stored, and
the number of bytes and files allowed for a single transaction. Two options affect
the number of files and bytes allowed for a single transaction. TXNGROUPMAX,
in the server options file, affects the number of files allowed. TXNBYTELIMIT, in
the client options file, affects the number of bytes allowed in the aggregate.
v The TXNGROUPMAX option in the server options file indicates the maximum
number of logical files (client files) that a client might send to the server in a
single transaction. The server might create multiple aggregates for a single
transaction, depending on how large the transaction is.
It is possible to affect the performance of client backup, archive, restore, and
retrieve operations by using a larger value for this option. When you transfer
multiple small files, increasing the TXNGROUPMAX option can improve
throughput for operations to tape.

262 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 Important: If you increase the value of the TXNGROUPMAX option by a large
amount, watch for possible effects on the recovery log. A larger value for the
TXNGROUPMAX option can result in increased utilization of the recovery log,
and an increased length of time for a transaction to commit. If the effects are
severe enough, they can lead to problems with operation of the server. For more
information, see “Files moved as a group between client and server” on page
685.
You can override the value of the TXNGROUPMAX server option for individual
client nodes by using the TXNGROUPMAX parameter in the REGISTER NODE
and UPDATE NODE commands.
v The TXNBYTELIMIT option in the client options file indicates the total number
of bytes that the client can send to the server in a single transaction.

When a Tivoli Storage Manager for Space Management client (HSM client)
migrates files to the server, the files are not grouped into an aggregate.

Server file aggregation is disabled for client nodes storing data that is associated
with a management class that has a copy group whose destination is a Centera
storage pool.

Where the server stores files

When a client file is backed up, archived, or migrated, the server verifies the
management class that is bound to the file. The management class specifies the
destination storage pool in which to store the file.

The server checks the destination storage pool to determine:

v If it is possible to write file data to the storage pool (access mode).
v If the size of the physical file exceeds the maximum file size allowed in the
storage pool. For backup and archive operations, the physical file may be an
aggregate or a single client file.
v Whether sufficient space is available on the available volumes in the storage
pool.
v What the next storage pool is, if any of the previous conditions prevent the file
from being stored in the storage pool that is being checked.

Using these factors, the server determines if the file can be written to that storage
pool or the next storage pool in the hierarchy.

Subfile backups: When the client backs up a subfile, it still reports the size of the
entire file. Therefore, allocation requests against server storage and placement in
the storage hierarchy are based on the full size of the file. The server does not put
a subfile in an aggregate with other files if the size of the entire file is too large to
put in the aggregate. For example, the entire file is 8 MB, but the subfile is only 10
KB. The server does not typically put a large file in an aggregate, so the server
begins to store this file as a stand-alone file. However, the client sends only 10 KB,
and it is now too late for the server to put this 10 KB file with other files in an
aggregate. As a result, the benefits of aggregation are not always realized when
clients back up subfiles.

Chapter 10. Managing storage pools and volumes 263

 Example: How the server determines where to store files in a
hierarchy
The server determines where to store a file based upon the destination storage
pool specified in the copy group of the management class to which the file is
bound. The server also checks the capacity utilization of the storage pool and the
maximum file size allowed.

Assume a company has a storage pool hierarchy as shown in Figure 20.

DISKPOOL
Read/Write Access
Max File Size=3MB

TAPEPOOL
Read/Write Access

Figure 20. Storage hierarchy example

The storage pool hierarchy consists of two storage pools:

DISKPOOL
The top of the storage hierarchy. It contains fast disk volumes for storing
data.
TAPEPOOL
The next storage pool in the hierarchy. It contains tape volumes accessed
by high-performance tape drives.

Assume a user wants to archive a 5 MB file that is named FileX. FileX is bound to
a management class that contains an archive copy group whose storage destination
is DISKPOOL, see Figure 20.

When the user archives the file, the server determines where to store the file based
on the following process:
1. The server selects DISKPOOL because it is the storage destination specified in
the archive copy group.
2. Because the access mode for DISKPOOL is read/write, the server checks the
maximum file size allowed in the storage pool.
The maximum file size applies to the physical file being stored, which may be a
single client file or an aggregate. The maximum file size allowed in DISKPOOL
is 3 MB. FileX is a 5 MB file and therefore cannot be stored in DISKPOOL.
3. The server searches for the next storage pool in the storage hierarchy.
If the DISKPOOL storage pool has no maximum file size specified, the server
checks for enough space in the pool to store the physical file. If there is not
enough space for the physical file, the server uses the next storage pool in the
storage hierarchy to store the file.

264 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 4. The server checks the access mode of TAPEPOOL, which is the next storage
pool in the storage hierarchy. The access mode for TAPEPOOL is read/write.
5. The server then checks the maximum file size allowed in the TAPEPOOL
storage pool. Because TAPEPOOL is the last storage pool in the storage
hierarchy, no maximum file size is specified. Therefore, if there is available
space in TAPEPOOL, FileX can be stored in it.

Backing up the data in a storage hierarchy

You can use copy storage pools and active-data pools to protect the data in
primary storage pools. Copy storage pools can contain any combination of active
and inactive data, archive data, or space-migrated data. Active-data pools contain
only active versions of client backup data.

Before you begin

Restoring a primary storage pool from an active-data pool might cause some or all
inactive files to be deleted from the database if the server determines that an
inactive file needs to be replaced but cannot find it in the active-data pool.

As a best practice, therefore, and to prevent the permanent loss of inactive versions
of client backup data, you should create a minimum of one active-data pool, which
contains active-data only, and one copy storage pool, which contains both active
and inactive data. To recover from a disaster, use the active-data pool to restore
critical client node data, and then restore the primary storage pools from the copy
storage pool. Do not use active-data pools for recovery of a primary pool or
volume unless the loss of inactive data is acceptable.

About this task

“Setting up copy storage pools and active-data pools” on page 266 describes the
high-level steps for implementation.

Neither copy storage pools nor active-data pools are part of a storage hierarchy,
which, by definition, consists only of primary storage pools. Data can be stored in
copy storage pools and active-data pools using the following methods:
v Including the BACKUP STGPOOL and COPY ACTIVEDATA commands in
administrative scripts or schedules so that data is automatically backed up or
copied at regular intervals.
v Enabling the simultaneous-write function so that data is written to primary
storage pools, copy storage pools, and active-data pools during the same
transaction. Writing data simultaneously to copy storage pools is supported for
backup, archive, space-management, and import operations. Writing data
simultaneously to active-data pools is supported only for client backup
operations and only for active backup versions.
v (copy storage pools only) Manually issuing the BACKUP STGPOOL command,
specifying the primary storage pool as the source and a copy storage pool as the
target. The BACKUP STGPOOL command backs up whatever data is in the
primary storage pool (client backup data, archive data, and space-managed
data).
v (active-data pools only) Manually issuing the COPY ACTIVEDATA command,
specifying the primary storage pool as the source and an active-data pool as the
target. The COPY ACTIVEDATA command copies only the active versions of
client backup data. If an aggregate being copied contains all active files, then the
entire aggregate is copied to the active-data pool during command processing. If

Chapter 10. Managing storage pools and volumes 265

 an aggregate being copied contains some inactive files, the aggregate is
reconstructed during command processing into a new aggregate without the
inactive files.

For efficiency, you can use a single copy storage pool and a single active-data pool
to back up all primary storage pools that are linked in a storage hierarchy. By
backing up all primary storage pools to one copy storage pool and one active-data
pool, you do not need to repeatedly copy a file when the file migrates from its
original primary storage pool to another primary storage pool in the storage
hierarchy.

In most cases, a single copy storage pool and a single active-data pool can be used
for backup of all primary storage pools. However, the number of copy storage
pools and active-data pools you actually need depends on whether you have more
than one primary storage pool hierarchy and on the type of disaster recovery
protection you want to implement. Multiple copy storage pools and active-data
pools might be needed to handle particular situations, including the following:
v Special processing of certain primary storage hierarchies (for example, archive
storage pools or storage pools dedicated to priority clients)
v Creation of multiple copies for multiple locations (for example, to keep one copy
on-site and one copy off-site)
v Rotation of full storage pool backups (See “Backing up primary storage pools”
on page 883.)

Inactive files in volumes in an active-data pool are deleted by reclamation

processing. The rate at which reclaimable space accumulates in active-data pool
volumes is typically faster than the rate for volumes in non-active-data pools. If
reclamation of volumes in an active-data pool is occurring too frequently, requiring
extra resources such as tape drives and libraries to mount and dismount volumes,
you can adjust the reclamation threshold until the rate of reclamation is acceptable.
The default reclamation threshold for active-data pools is 60 percent, which means
that reclamation begins when the storage pool reaches 60 percent of capacity. Note
that accelerated reclamation of volumes has more of an effect on active-data pools
that use removable media and, in particular, on removable media that is taken
off-site.

Setting up copy storage pools and active-data pools

To back up the data in primary storage pools, use copy storage pools, active-data
pools, or combination of the two.

About this task

To set up a copy storage pool or an active-data pool:

Procedure
1. Define a copy storage pool or active-data pool. For details, see “Defining copy
storage pools and active-data pools” on page 409.
2. (active-data pools only) Create a policy domain, and specify the name of the
active-data pool as the value of the ACTIVEDATAPOOL parameter. To learn
more about creating domains and the ACTIVEDATAPOOL parameter, see
“Defining and updating a policy domain” on page 498.
3. (active-data pools only) Identify the nodes whose active backup data is to be
stored in the active-data pool, and then assign the nodes to the domain defined
in step 2. For details about assigning nodes to a domain, see “Assigning client
nodes to a policy domain” on page 512.
266 IBM Tivoli Storage Manager for Windows: Administrator's Guide
4. (optional) If you want to use the simultaneous-write function, update the
primary storage pool definition, specifying the name of the copy storage pool
and active-data pool as the values of the COPYSTGPOOLS and
ACTIVEDATAPOOLS parameters, respectively. For details about the
simultaneous-write function, see “Writing data simultaneously to primary, copy,
and active-data pools” on page 331.
Ensure that client sessions have sufficient mount points. Each session requires
one mount point for the primary storage pool and a mount point for each copy
storage pool and each active-data pool. To allow a sufficient number of mounts
points, use the MAXNUMMP parameter on the REGISTER NODE or UPDATE NODE
commands.
5. Set up administrative schedules or scripts to automatically issue the BACKUP
STGPOOL and COPY ACTIVEDATA commands. See “Automating a basic
administrative command schedule” on page 620 and “Tivoli Storage Manager
server scripts” on page 626.

Example: Setting up an active-data pool for fast client restore:

A sequential-access disk (FILE) device class is used to set up an active-data pool

for fast restore of client-node data.

Before you begin

Decide which client nodes have data that needs to be restored quickly if a disaster
occurs. Only the data belonging to those nodes should be stored in the active-data
pool.

About this task

For the purposes of this example, the following definitions already exist on the
server:
v The default STANDARD domain, STANDARD policy set, STANDARD
management class, and STANDARD copy group.
v A primary storage pool, BACKUPPOOL, and a copy storage pool, COPYPOOL.
BACKUPPOOL is specified in the STANDARD copy group as the storage pool
in which the server initially stores backup data. COPYPOOL contains copies of
all the active and inactive data in BACKUPPOOL.
v Three nodes that are assigned to the STANDARD domain (NODE1, NODE2, and
NODE 3).
v Two mount points assigned for each client session.
v A FILE device class named FILECLASS.

You have identified NODE2 as the only high-priority node, so you need to create a
new domain to direct the data belonging to that node to an active-data pool. To set
up and enable the active-data pool, follow these steps:

Procedure
1. Define the active-data pool:
DEFINE STGPOOL ADPPOOL FILECLASS POOLTYPE=ACTIVEDATA MAXSCRATCH=1000
2. Define a new domain and specify the active-data pool in which you want to
store the data belonging to NODE2:
DEFINE DOMAIN ACTIVEDOMAIN ACTIVEDESTINATION=ADPPOOL
3. Define a new policy set:

Chapter 10. Managing storage pools and volumes 267

 DEFINE POLICYSET ACTIVEDOMAIN ACTIVEPOLICY
4. Define a new management class:
DEFINE MGMTCLASS ACTIVEDOMAIN ACTIVEPOLICY ACTIVEMGMT
5. Define a backup copy group:
DEFINE COPYGROUP ACTIVEDOMAIN ACTIVEPOLICY ACTIVEMGMT DESTINATION=BACKUPPOOL

This command specifies that the active and inactive data belonging to client
nodes that are members of ACTIVEDOMAIN will be backed up to
BACKUPPOOL. Note that this is the destination storage pool for data backed
up from nodes that are members of the STANDARD domain.
6. Assign the default management class for the active-data pool policy set:
ASSIGN DEFMGMTCLASS ACTIVEDOMAIN ACTIVEPOLICY ACTIVEMGMT
7. Activate the policy set for the active-data pool:
ACTIVATE POLICYSET ACTIVEDOMAIN ACTIVEPOLICY
8. Assign the high-priority node, NODE2, to the new domain:
UPDATE NODE NODE2 DOMAIN=ACTIVEDOMAIN

A node can belong to only one domain. When you update a node by changing
its domain, you remove it from its current domain.
9. (optional) Update the primary storage pool, BACKUPPOOL, with the name of
the active-data pool, ADPPOOL, where the server simultaneously will write
data during a client backup operation:
UPDATE STGPOOL BACKUPPOOL ACTIVEDATAPOOLS=ADPPOOL

Only active versions of backup data can be simultaneously written to

active-data pools.
10. To ensure that copies of active data are complete, define a schedule to copy
active data from BACKUPPOOL to ADPPOOL every day at 8:00 p.m.:
DEFINE SCHEDULE COPYACTIVE_BACKUPPOOL TYPE=ADMINISTRATIVE
CMD="COPY ACTIVEDATA BACKUPPOOL ADPPOOL" ACTIVE=YES
STARTTIME=20:00 PERIOD=1

Instead of defining a schedule, you can issue the COPY ACTIVEDATA

command manually whenever it is convenient to copy the active data.

Results

Every time NODE2 stores data into BACKUPPOOL, the server simultaneously
writes the data to ADPPOOL. The schedule, COPYACTIVE_BACKUPPOOL,
ensures that any data that was not stored during simultaneous-write operations is
copied to the active-data pool. When client nodes NODE1 and NODE3 are backed
up, their data is stored in BACKUPPOOL only, and not in ADPPOOL. When the
administrative schedule runs, only the data belonging to NODE2 is copied to the
active-data pool.

Remember: If you want all the nodes belonging to an existing domain to store
their data in the active-data pool, then you can skip steps 2 through 8. Use the
UPDATE DOMAIN command to update the STANDARD domain, specifying the
name of the active-data pool, ADPPOOL, as the value of the
ACTIVEDESTINATION parameter.

268 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Example: Setting up an active-data pool to reduce media resources:

Backup data is simultaneously written to an active-data pool so that volumes in

the pool can be taken off-site.

About this task

In addition to using active-data pools for fast restore of client-node data, you can
also use active-data pools to reduce the number of tape volumes that are stored
either on-site or off-site for the purpose of disaster recovery. This example assumes
that, in your current configuration, all data is backed up to a copy storage pool
and taken off-site. However, your goal is to create an active-data pool, take the
volumes in that pool off-site, and maintain the copy storage pool on-site to recover
primary storage pools.

Attention: Active-data pools should not be considered for recovery of a primary

pool or volume unless the loss of inactive data is acceptable.

The following definitions already exist on the server:

v The default STANDARD domain, STANDARD policy set, STANDARD
management class, and STANDARD copy group.
v A primary storage pool, BACKUPPOOL, and a copy storage pool, COPYPOOL.
BACKUPPOOL is specified in the STANDARD copy group as the storage pool
in which the server initially stores backup data. COPYPOOL contains copies of
all the active and inactive data in BACKUPPOOL.
v An administrative schedule, named BACKUP_BACKUPPOOL, that issues a
BACKUP STGPOOL command to back up the data in BACKUPPOOL to
COPYPOOL. The schedule runs every day at 10:00 p.m.
v Three nodes that are assigned to the STANDARD domain (NODE1, NODE2, and
NODE 3).
v Two mount points for each client session.
v A device class of type 3592 named 3592CLASS.

To set up and enable an active-data pool, follow these steps:

Procedure
1. Define the active-data pool:
DEFINE STGPOOL ADPPOOL 3592CLASS POOLTYPE=ACTIVEDATA MAXSCRATCH=1000
2. Update the STANDARD domain to allow data from all nodes to be stored in
the active-data pool:
UPDATE DOMAIN STANDARD ACTIVEDESTINATION=ADPPOOL
3. (optional) To set up simultaneous write operations, complete the following steps:
v Update the primary storage pool, BACKUPPOOL, with the name of the
active-data pool, ADPPOOL, where the server will write data simultaneously
during client backup operations:
UPDATE STGPOOL BACKUPPOOL ACTIVEDATAPOOL=ADPPOOL

Only active versions of backup data can be simultaneously written to

active-data pools.
v For each node, update the maximum number of mount points allowed to
include the primary storage pool, the copy storage pool, and the active-data
pool:
UPDATE NODE NODE2 MAXNUMMP=3

Chapter 10. Managing storage pools and volumes 269

 4. To ensure that copies of active data are complete, define a schedule to copy
active data from BACKUPPOOL to ADPPOOL every day at 8:00 p.m.:
DEFINE SCHEDULE COPYACTIVE_BACKUPPOOL TYPE=ADMINISTRATIVE
CMD="COPY ACTIVEDATA BACKUPPOOL ADPPOOL" ACTIVE=YES STARTTIME=20:00 PERIOD=1

Instead of defining a schedule, you can issue the COPY ACTIVEDATA

command manually whenever it is convenient to copy the active data.

Results

Every time data is stored into BACKUPPOOL, the data is simultaneously written
to ADPPOOL. The schedule, COPYACTIVE_BACKUPPOOL, ensures that any data
that was not stored during a simultaneous-write operation is copied to the
active-data pool. You can now move the volumes in the active-data pool to a safe
location off-site.

If your goal is to replace the copy storage pool with the active-data pool, follow
the steps below. As a best practice and to protect your inactive data, however, you
should maintain the copy storage pool so that you can restore inactive versions of
backup data if required. If the copy storage pool contains archive or files that were
migrated by a Tivoli Storage Manager for Space Management client, do not delete
it.
1. Stop backing up to the copy storage pool:
DELETE SCHEDULE BACKUP_BACKUPPOOL
UPDATE STGPOOL BACKUPPOOL COPYSTGPOOLS=""
2. After all data has been copied to the active-data pool, delete the copy storage
pool and its volumes.

Staging client data from disk to tape

Typically, client backup data is stored initially in disk-based storage pools. To make
room for additional backups, you can migrate the older data to tape. If you are
using copy storage pools or active-data pools, store data in those pools before
beginning the migration process.

About this task

Typically you need to ensure that you have enough disk storage to process one
night's worth of the clients' incremental backups. While not always possible, this
guideline proves to be valuable when considering storage pool backups.

For example, suppose you have enough disk space for nightly incremental backups
for clients, but not enough disk space for a FILE-type, active-data pool. Suppose
also that you have tape devices. With these resources, you can set up the following
pools:
v A primary storage pool on disk, with enough volumes assigned to contain the
nightly incremental backups for clients
v A primary storage pool on tape, which is identified as the next storage pool in
the hierarchy for the disk storage pool
v An active-data pool on tape
v A copy storage pool on tape

You can then schedule the following steps every night:

270 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 Procedure
1. Perform an incremental backup of the clients to the disk storage pool.
2. After clients complete their backups, back up the active and inactive versions in
the disk primary storage pool (now containing the incremental backups) to the
copy storage pool. Then copy the active backup versions to the active-data
pool.
Backing up disk storage pools before migration processing allows you to copy
as many files as possible while they are still on disk. This saves mount requests
while performing your storage pool backups. If the migration process starts
while active data is being copied to active-data pools or while active and
inactive data is being backed up to copy storage pools, some files might be
migrated before they are copied or backed up.
3. Start the migration of the files in the disk primary storage pool to the tape
primary storage pool (the next pool in the hierarchy) by lowering the high
migration threshold. For example, lower the threshold to 40%.
When this migration completes, raise the high migration threshold back to
100%.
4. To ensure that all files are backed up, back up the tape primary storage pool to
the copy storage pool. In addition, copy the active backup data in the tape
primary storage pool to the active-data pool.
The tape primary storage pool must still be backed up (and active files copied)
to catch any files that might have been missed in the backup of the disk storage
pools (for example, large files that went directly to sequential media).

Results

For more information about storage pool space, see “Estimating space needs for
storage pools” on page 379

Migrating files in a storage pool hierarchy

To maintain free space in primary storage pools, the Tivoli Storage Manager server
can automatically migrate data from one primary pool to the next in the hierarchy.

About this task

The migration process can help to ensure that there is sufficient free space in the
storage pools at the top of the hierarchy, where faster devices can provide the most
benefit to clients. For example, the server can migrate data that is stored in a
random-access disk storage pool to a slower, but less expensive sequential-access
storage pool.

You can control the following options:

When migration begins and ends
Migration thresholds are used to control when migration begins and ends.
Thresholds are set as levels of the space that is used in a storage pool. It is
expressed as a percentage of the total space available in the storage pool.
For random-access and sequential-access disk storage pools, the server
compares the threshold to the amount of data that is stored in the pool as
a percentage of the total data capacity of the volumes in the pool. Total
data capacity for sequential-access disk storage pools includes the capacity
of all scratch volumes that are specified for the pool. For tape storage

Chapter 10. Managing storage pools and volumes 271

 pools, the server compares the threshold to the number of volumes that
contain data as a percentage of the total number of volumes available to
the pool, including scratch volumes.
You can also schedule migration activities to occur at optimal times in your
environment. In addition, you can specify the following options:
v How long migration can run before it is automatically canceled
v Whether the server attempts reclamation before migration
v Whether the migration process runs in the background or foreground
How the server chooses files to migrate
By default, the server does not consider how long a file has been stored in
a storage pool, or how long since the file was accessed before it chooses
files to migrate. There are optional parameters available for you to change
the default behavior of migration. To ensure that files remain in a storage
pool for a minimum number of days before the server migrates them to
another pool, you can set a migration delay period for the storage pool.
For random-access disk storage pools, the last time the file was accessed is
also considered for migration delay. For sequential-access storage pools,
including sequential-access disk storage pools that are associated with a
FILE device class, all files on a volume must exceed migration delay period
before the server can migrate all of the files on the volume.
The number of concurrent migration processes
You can specify a single migration process or multiple concurrent
migration processes for a random-access or sequential-access storage pool.
You can use multiple concurrent processes to make better use of your
available tape drives, and FILE volumes. Since you can run migration
concurrently on different storage pools during auto-migration, you must
carefully consider the resources (for example, drives) you have available
for the operation.
If simultaneous-write operations during migration are enabled during
multiple concurrent-migration processing, each process has the following
requirements:
v A mount point
v A volume for each copy storage pool, and active-data pool that is
defined to the target storage pool, and the primary pool
For details about the simultaneous-write function, see “Writing data
simultaneously to primary, copy, and active-data pools” on page 331.

Migration processing can differ for disk storage pools versus sequential-access
storage pools. If you plan to modify the default migration settings for storage
pools or want to understand how migration works, read the following topics:
v “Migrating disk storage pools” on page 273
v “Migrating sequential-access storage pools” on page 279
v “Starting migration manually or in a schedule” on page 282

Remember:
v Data cannot be migrated in to or out of storage pools that are defined with a
CENTERA device class.
v If you receive an error message during the migration process, refer to IBM Tivoli
Storage Manager Messages, which can provide useful information to diagnose and
fix problems.

272 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 v If a migration process is started from a storage pool that does not have the next
storage pool identified in the hierarchy, a reclamation process is triggered for the
source storage pool. To prevent the reclamation process, define the next storage
pool in the hierarchy. For details, see “Setting up a storage pool hierarchy” on
page 261. As an alternative to prevent automatic migration from running, set the
HIGHMIG parameter of the storage pool definition to 100.

Migrating disk storage pools

Migration thresholds specify when the server should begin and stop migrating
data to the next storage pool in the storage hierarchy. Migration thresholds are
defined as a percentage of total storage-pool data capacity.

About this task

You can use the defaults for the migration thresholds, or you can change the
threshold values to identify the maximum and minimum amount of space for a
storage pool.

To control how long files must stay in a storage pool before they are eligible for
migration, specify a migration delay for a storage pool. For details, see “Keeping
files in a storage pool” on page 278.

If you decide to enable cache for disk storage pools, files can temporarily remain
on disks even after migration. When you use cache, you might want to set lower
migration thresholds.

For more information about migration thresholds, see “How the server selects files
to migrate” and “Migration thresholds” on page 276. For information about using
the cache, see “Minimizing access time to migrated files” on page 279 and
“Caching in disk storage pools” on page 284.

How the server selects files to migrate

When the amount of data in a storage pool is equal to the high migration
threshold, the server must migrate files to the next storage pool. The process for
selecting files to migrate is based on the space that is used by a client node's files
and the migration delay setting.

The server selects the files to migrate as follows:

1. The server checks for the client node that backed up or migrated the largest
single file space or archived files that occupy the most space.
2. The next pool collocation setting determines how data is processed. If the
collocation setting is GROUP, all the filespaces for all nodes in the collocation
group are processed together. If the setting is NODE, all the filespaces for that
node are processed together. If the collocation setting is NO or FILESPACE,
each filespace and data type (backup, archive, or migrate) is processed
separately.
3. For all files from every filespace belonging to the client node that was identified,
the server examines the following statistics:
v The number of days since the files were stored in the storage pool
v The number of days since the files were retrieved from the storage pool
The server compares the number (whichever is less) to the migration delay that
is set for the storage pool. The server migrates any of these files for which the
number is more than the migration delay set for the storage pool.
The server checks the low migration threshold for the storage pool:
Chapter 10. Managing storage pools and volumes 273
 v After the server migrates the files for the node to collocate by GROUP or
NODE, or
v After the server migrates the processing group to the next storage pool
If the amount of space that is used in the storage pool is now below the low
migration threshold, migration ends. If not, the server continues with that
group to collocate by group, if there are nodes that are left in the group. The
server can also choose another group by using criteria that has already been
described, and the migration process continues.

The server might not be able to reach the low migration threshold by migrating
only files that were stored longer than the migration delay period. If so, the server
checks the storage pool characteristic that determines whether migration stops,
even if the pool is still above the low migration threshold. For more information,
see “Keeping files in a storage pool” on page 278.

If multiple migration processes are running, the server might choose the files from
more than one node for migration at the same time. The processes are controlled
by the MIGPROCESS parameter of the DEFINE STGPOOL command.

For example, Table 23 displays information that is contained in the database that is
used by the server to determine which files to migrate. This example assumes that
the storage pool contains no space-managed files. This example also assumes that
the migration delay period for the storage pool is set to zero. When the setting is
zero, any files can be migrated. Regardless of the amount of time they are stored in
the pool or the last time of access, all files are migrated.
Table 23. Database information about files stored in DISKPOOL
Archived Files (All Client File
Client Node Backed-Up File Spaces and Sizes Spaces)
TOMC TOMC/C 200 MB 55 MB
TOMC/D 100 MB
CAROL CAROL 50 MB 5 MB
PEASE PEASE/home 150 MB 40 MB
PEASE/temp 175 MB

274 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 Before Migration During Migration After Migration

High
Migration
Threshold
80%

Low
Migration
Threshold
20% DISKPOOL DISKPOOL DISKPOOL

TAPEPOOL

Figure 21. The migration process and migration thresholds

Figure 21 shows what happens when the following occur:

v The high migration threshold that is defined for the disk storage pool
DISKPOOL is exceeded
v The TAPEPOOL COLLOCATE parameter is set at the default GROUP or NODE
When the amount of data that can be migrated in DISKPOOL reaches 80%, the
server completes the following tasks:
1. Determines that the TOMC/C file space backup data is taking up the most
space in the DISKPOOL storage pool. This space is more than any other single
backed-up or space-managed or archived file space and more than any client
node's archived files.
2. Locates all data that belongs to node TOMC stored in DISKPOOL. In this
example, node TOMC backed up or archived files from file spaces TOMC/C
and TOMC/D that are stored in the DISKPOOL storage pool.
3. Migrates all data from TOMC/C and TOMC/D to the next available storage
pool. In this example, the data is migrated to the tape storage pool,
TAPEPOOL.
The server migrates all of the data from both file spaces that belong to node
TOMC. The migration happens, even if the occupancy of the storage pool drops
below the low migration threshold before the second file space is migrated.
If the cache option is enabled, files that are migrated remain on disk storage
(that is, the files are cached) until space is needed for new files. For more
information about using cache, see “Caching in disk storage pools” on page
284.
4. After all files that belong to TOMC are migrated to the next storage pool, the
server checks the low migration threshold. If the low migration threshold is not
reached, the server determines which client node backed up or migrated the
largest single file space or archived files that occupy the most space. The server
begins migrating the files that belong to that node.
In this example, the server migrates all files that belong to the client node
named PEASE to the TAPEPOOL storage pool.

Chapter 10. Managing storage pools and volumes 275

 5. After all the files that belong to PEASE are migrated to the next storage pool,
the server checks the low migration threshold again. If the low migration
threshold is reached or passed, migration ends.

If the TAPEPOOL COLLOCATE parameter is set to NO or FILESPACE, the server

completes the following tasks:
1. Determines that the TOMC/C file space backup data is taking up the most
space in the DISKPOOL storage pool. This space is more than any other single
backed-up or space-managed file space and more than any client node's
archived files.
2. Locates all backup data that belongs to node TOMC/C that is stored in the
DISKPOOL storage pool.
3. Migrates all backup data from TOMC/C to the next available storage pool. In
this example, the data migrates to the tape storage pool, TAPEPOOL.
The server migrates all of the backup data from filespace C belonging to node
TOMC. The migration happens, even if the occupancy of the storage pool drops
below the low migration threshold of all the data that was migrated.
If the cache option is enabled, files that are migrated remain on disk storage
(files are cached) until space is needed for new files. For more information
about using cache, see “Caching in disk storage pools” on page 284.
4. After all backup files that belong to TOMC/C are migrated to the next storage
pool, the server checks the low migration threshold. If the low migration
threshold is not reached, the server determines which filespace backed up,
migrated, or archived files that occupy the most space. The server begins
migrating files that belong to that filespace/type.
In this example, the server migrates all archive files that belong to the client
node named PEASE filespace /temp to the TAPEPOOL storage pool.
5. After all the archive files that belong to PEASE/temp are migrated to the next
storage pool, the server checks the low migration threshold again. If the low
migration threshold was reached or passed, migration ends.

Migration thresholds
Migration thresholds specify when migration for a storage pool begins and ends.
Setting migration thresholds for disk storage pools ensures sufficient free space on
faster devices, which can lead to better performance.

Choosing thresholds appropriate for your situation takes some experimenting. Start
by using the default high and low values. You need to ensure that migration
occurs frequently enough to maintain some free space but not so frequently that
the device is unavailable for other use.

High-migration thresholds:

Before changing the high-migration threshold, you need to consider the amount of
storage capacity provided for each storage pool and the amount of free storage
space needed to store additional files, without having migration occur.

If you set the high-migration threshold too high, the pool may be just under the
high threshold, but not have enough space to store an additional, typical client file.
Or, with a high threshold of 100%, the pool may become full and a migration
process must start before clients can back up any additional data to the disk
storage pool. In either case, the server stores client files directly to tape until
migration completes, resulting in slower performance.

276 IBM Tivoli Storage Manager for Windows: Administrator's Guide

If you set the high-migration threshold too low, migration runs more frequently
and can interfere with other operations.

Keeping the high-migration threshold at a single value means that migration

processing could start at any time of day, whenever that threshold is exceeded. You
can control when migration occurs by using administrative command schedules to
change the threshold. For example, set the high-migration threshold to 95% during
the night when clients run their backup operations. Lower the high-migration
threshold to 50% during the time of day when you want migration to occur. By
scheduling when migration occurs, you can choose a time when your tape drives
and mount operators are available for the operation.

Low-migration thresholds:

Before setting the low-migration threshold, you must consider the amount of free
disk storage space needed for normal daily processing.

To choose the low-migration threshold, consider:

v The amount of free disk storage space needed for normal daily processing. If
you have disk space to spare, you can keep more data on the disk (a larger low
threshold). If clients' daily backups are enough to fill the disk space every day,
you might need to empty the disk (a smaller low threshold).
If your disk space is limited, try setting the threshold so that migration frees up
enough space for the pool to manage the amount of client data that is stored
every day. Migration then runs about every day, or you can force it to run every
day by lowering the high-migration threshold at a time of your choosing.
You might also want to identify clients that are daily, transferring large amounts
of data. For these clients, you might want to set up a policy (a new copy group
or a new policy domain) so that their data is stored directly to tape. Using a
separate policy like this can optimize disk use for most clients.
v Whether you use cache on disk storage pools to improve how quickly some files
are retrieved. If you use cache, you can set the low threshold lower, yet still
maintain faster retrieval for some data. Migrated data remains cached on the
disk until new client data pushes the data off the disk. Using cache requires
more disk space for the database, however, and can slow backup and archive
operations that use the storage pool.
If you do not use cache, you might want to keep the low threshold at a higher
number so that more data stays on the disk.
v How frequently you want migration to occur, based on the availability of
sequential-access storage devices and mount operators. The larger the low
threshold, the shorter time that a migration process runs (because there is less
data to migrate). But if the pool refills quickly, then migration occurs more
frequently. The smaller the low threshold, the longer time that a migration
process runs, but the process runs less frequently.
You might need to balance the costs of larger disk storage pools with the costs
of running migration (drives, tapes, and either operators or automated libraries).
v Whether data in the next storage pool is being collocated. When you collocate by
GROUP, all the file spaces for all nodes in the collocation group are processed
together. When you collocate by NODE, all file spaces for that node are
processed as a group. When collocation is NO or FILESPACE, each filespace and
data type is processed separately, providing greater granularity.

Chapter 10. Managing storage pools and volumes 277

 Keeping files in a storage pool
For some applications, you might want to delay the migration of files in the
storage pool where they were initially stored by the server. You can delay
migration of files for a specified number of days.

About this task

For example, you might have backups of monthly summary data that you want to
keep in your disk storage pool for faster access until the data is 30 days old. After
the 30 days, the server moves the files to a tape storage pool.

To delay file migration of files, set the MIGDELAY parameter when you define or
update a storage pool. The number of days is counted from the day that a file was
stored in the storage pool or accessed by a client, whichever is more recent. You
can set the migration delay separately for each storage pool. When you set the
delay to zero, the server can migrate any file from the storage pool, regardless of
how short a time the file has been in the storage pool. When you set the delay to
greater than zero, the server checks how long the file has been in the storage pool
and when it was last accessed by a client. If the number of days exceeds the
migration delay, the server migrates the file.

Note: If you want the number of days for migration delay to be counted based
only on when a file was stored and not when it was retrieved, use the
NORETRIEVEDATE server option. For more information about this option, see the
Administrator's Reference.

If you set migration delay for a pool, you must decide what is more important:
either ensuring that files stay in the storage pool for the migration delay period, or
ensuring that there is enough space in the storage pool for new files. For each
storage pool that has a migration delay set, you can choose what happens as the
server tries to move enough data out of the storage pool to reach the low
migration threshold. If the server cannot reach the low migration threshold by
moving only files that have been stored longer than the migration delay, you can
choose one of the following:
v Allow the server to move files out of the storage pool even if they have not been
in the pool for the migration delay (MIGCONTINUE=YES). This is the default.
Allowing migration to continue ensures that space is made available in the
storage pool for new files that need to be stored there.
v Have the server stop migration without reaching the low migration threshold
(MIGCONTINUE=NO). Stopping migration ensures that files remain in the
storage pool for the time you specified with the migration delay. The
administrator must ensure that there is always enough space available in the
storage pool to hold the data for the required number of days.

If you allow more than one migration process for the storage pool and allow the
server to move files that do not satisfy the migration delay time
(MIGCONTINUE=YES), some files that do not satisfy the migration delay time
may be migrated unnecessarily. As one process migrates files that satisfy the
migration delay time, a second process could begin migrating files that do not
satisfy the migration delay time to meet the low migration threshold. The first
process that is still migrating files that satisfy the migration delay time might have,
by itself, caused the storage pool to meet the low migration threshold.

278 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 Minimizing access time to migrated files
Caching is a method of minimizing access time to files on disk storage, even if the
server has migrated files to a tape storage pool. However, cached files are removed
from disk when the space they occupy is required. The files must then be obtained
from the storage pool to which they were migrated

About this task

Important: For information about the disadvantages of using cache, see “Caching
in disk storage pools” on page 284.

To ensure that files remain on disk storage and do not migrate to other storage
pools, use one of the following methods:
v Do not define the next storage pool.
A disadvantage of using this method is that if the file exceeds the space
available in the storage pool, the operation to store the file fails.
v Set the high-migration threshold to 100%.
When you set the high migration threshold to 100%, files will not migrate at all.
You can still define the next storage pool in the storage hierarchy, and set the
maximum file size so that large files are stored in the next storage pool in the
hierarchy.
A disadvantage of setting the high threshold to 100% is that after the pool
becomes full, client files are stored directly to tape instead of to disk.
Performance may be affected as a result.

Migrating sequential-access storage pools

You can set up migration thresholds for sequential-access storage pools. Migrating
data from one sequential-access storage pool to another might be appropriate in
some cases, for example, when you install a tape drive that uses a different type of
tape and want to move data to that tape.

About this task

You probably will not want the server to migrate sequential-access storage pools
on a regular basis. An operation such as tape-to-tape migration has limited benefits
compared to disk-to-tape migration, and requires at least two tape drives.

You can migrate data from a sequential-access storage pool only to another
sequential-access storage pool. You cannot migrate data from a sequential-access
storage pool to a disk storage pool. If you need to move data from a
sequential-access storage pool to a disk storage pool, use the MOVE DATA
command. See “Moving data from one volume to another volume” on page 400.

To control the migration process, set migration thresholds and migration delays for
each storage pool using the DEFINE STGPOOL and UPDATE STGPOOL
commands. You can also specify multiple concurrent migration processes to better
use your available tape drives or FILE volumes. (For details, see “Specifying
multiple concurrent migration processes” on page 282.) Using the MIGRATE
STGPOOL command, you can control the duration of the migration process and
whether reclamation is attempted prior to migration. For additional information,
see “Starting migration manually or in a schedule” on page 282.

Tip: Data in storage pools that have an NDMP format (NETAPPDUMP,

CELERRADUMP, or NDMPDUMP) cannot be migrated. However, in primary

Chapter 10. Managing storage pools and volumes 279

 storage pools that have an NDMP format, you can make space available by using
the MOVE DATA command. The target storage pool must have the same data
format as the source storage pool.

How the server migrates files from sequential-access storage

pools
The server migrates files by volume from sequential-access storage pools. Volumes
that exceed the reclamation threshold are migrated first. Files in the least
frequently referenced volumes are migrated next. The server checks the migration
delay for the storage pool before any files are migrated.

For tape storage pools, the server begins the migration process when the ratio of
volumes containing data to the total number of volumes in the storage pool,
including scratch volumes, reaches the high migration threshold. For
sequential-access disk (FILE) storage pools, the server starts the migration process
when the ratio of data in a storage pool to the pool's total estimated data capacity
reaches the high migration threshold. The calculation of data capacity includes the
capacity of all the scratch volumes that are specified for the pool.

Tip: When Tivoli Storage Manager calculates the capacity for a sequential-access
disk storage pool, it considers the amount of disk space available in the file
system. For this reason, ensure that you have enough disk space in the file system
to hold all the defined and scratch volumes that are specified for the storage pool.
For example, suppose that the capacity of all the scratch volumes that are specified
for a storage pool is 10 TB. (There are no predefined volumes.) However, only 9 TB
of disk space is available in the file system. The capacity value that is used in the
migration threshold is 9 TB, not 10 TB. If the high migration threshold is set to
70%, migration begins when the storage pool contains 6.3 TB of data, not 7 TB.

When migrating files by volume from sequential-access storage pools, including

sequential-access disk storage pools associated with a FILE device class, the server
completes the following procedure:
1. The server first reclaims volumes that exceed the reclamation threshold.
Reclamation is a server process of consolidating files from several volumes onto
one volume. (See “Reclaiming space in sequential-access storage pools” on page
367.)
2. After reclamation processing, the server compares the space that is used in the
storage pool to the low migration threshold.
3. If the space used is now below the low migration threshold, the server stops
processing. If the space used is still above the low migration threshold, the
server determines which volume is the least recently referenced volume.
4. If the amount of time a file has been in the storage pool exceeds the amount of
time that is specified as the migration delay for the storage pool, the file is
eligible for migration. The server selects the volume for migration only when
all files on the volume are eligible for migration.
5. The server repeats steps 3 and 4 until the storage pool reaches the low
migration threshold.

Migration delay can prevent volumes from being migrated. The server can migrate
files from all eligible volumes but still find that the storage pool is above the low
migration threshold. If you set migration delay for a pool, you must decide which
of the following is of greater importance. Either ensuring that files stay in the
storage pool for the amount of time that is specified as the migration delay, or
ensuring there is enough space in the storage pool for new files. For each storage
pool that has a migration delay that is set, you can choose what happens as the

280 IBM Tivoli Storage Manager for Windows: Administrator's Guide

server tries to move enough files out of the storage pool to reach the low migration
threshold. If the server cannot reach the low migration threshold by migrating only
volumes that meet the migration delay requirement, you can choose one of the
following:
v Allow the server to migrate volumes from the storage pool even if they do not
meet the migration delay criteria (MIGCONTINUE=YES). This is the default.
Allowing migration to continue ensures that space is made available in the
storage pool for new files that must be stored there.
v Have the server stop migration without reaching the low migration threshold
(MIGCONTINUE=NO). Stopping migration ensures that volumes are not
migrated for the time you specified with the migration delay. The administrator
must ensure that there is always enough space available in the storage pool to
hold the data for the required number of days.

Migration criteria for sequential-access storage pools

If you are planning to use migration for sequential-access storage pools, consider
the time that is required to mount tapes into drives and whether collocation is
enabled.

When you define migration criteria for sequential-access storage pools, consider:
v The capacity of the volumes in the storage pool
v The time that is required to migrate data to the next storage pool
v The speed of the devices that the storage pool uses
v The time that is required to mount media, such as tape volumes, into drives
v Whether operator presence is required
v The number of concurrent migration processes

If you decide to migrate data from one sequential-access storage pool to another,
ensure that:
v Two drives (mount points) are available, one in each storage pool.
v The access mode for the next storage pool in the storage hierarchy is set to
read/write.
For information about setting an access mode for sequential-access storage pools,
see “Defining storage pools” on page 245.
v Collocation is set the same in both storage pools. For example, if collocation is
set to NODE in the first storage pool, then set collocation to NODE in the next
storage pool. If collocation is set to FILESPACE in the first storage pool, then set
collocation to FILESPACE in the next storage pool.
When you enable collocation for a storage pool, the server tries to keep all files
on a minimal number of volumes. The files can belong to a single client node, a
group of client nodes, a client file space, or a group of file spaces. For
information about collocation for sequential-access storage pools, see “Keeping
client files together using collocation” on page 357.
v You have sufficient resources (for example, staff) available to manage any
necessary media mount and dismount operations. (This is especially true for
multiple concurrent processing, For details, see “Specifying multiple concurrent
migration processes” on page 282.) More mount operations occur because the
server attempts to reclaim space from sequential-access storage pool volumes
before it migrates files to the next storage pool.
If you want to limit migration from a sequential-access storage pool to another
storage pool, set a high, high-migration threshold, such as 95%.

Chapter 10. Managing storage pools and volumes 281

 For information about setting a reclamation threshold for tape storage pools, see
“Reclaiming space in sequential-access storage pools” on page 367.

There is no straightforward way to selectively migrate data for a specific node

from one sequential storage pool to another. You can use the MOVE NODEDATA
command to move file spaces for a node from one storage pool to another. See
“Moving data belonging to a client node” on page 406.

Starting migration manually or in a schedule

To gain more control over how and when the migration process occurs, you can
use the MIGRATE STGPOOL command. Issuing this command starts migration from
one storage pool to the next storage pool in the hierarchy, regardless of the value
of the HIGHMIG parameter of the storage pool definition.

About this task

You can specify the maximum number of minutes the migration runs before
automatically canceling. If you prefer, you can include this command in a schedule
to perform migration when it is least intrusive to normal production needs.

For example, to migrate data from a storage pool named ALTPOOL to the next
storage pool, and specify that it end as soon as possible after one hour, issue the
following command:
migrate stgpool altpool duration=60

Do not use this command if you are going to use automatic migration. To prevent
automatic migration from running, set the HIGHMIG parameter of the storage pool
definition to 100. For details about the MIGRATE STGPOOL command, refer to the
Administrator's Reference.

Results

Restriction: Data cannot be migrated into or out of storage pools that are defined
with a CENTERA device class.

Specifying multiple concurrent migration processes

Running multiple migration processes concurrently lets you make better use of
your available tape drives or FILE volumes. When calculating the number of
concurrent processes to run, you must carefully consider available resources.

About this task

Each migration process requires at least two simultaneous volume mounts (at least
two mount points) and, if the device type is not FILE, at least two drives. One of
the drives is for the input volume in the storage pool from which files are being
migrated. The other drive is for the output volume in the storage pool to which
files are being migrated.

When calculating the number of concurrent processes to run, carefully consider the
resources you have available, including the number of storage pools that will be
involved with the migration, the number of mount points, the number of drives
that can be dedicated to the operation, and (if appropriate) the number of mount
operators available to manage migration requests. The number of available mount
points and drives depends on other Tivoli Storage Manager and system activity
and on the mount limits of the device classes for the storage pools that are
involved in the migration. For more information about mount limit, see:

282 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 “Controlling the number of simultaneously mounted volumes” on page 181

For example, suppose that you want to migrate data on volumes in two sequential
storage pools simultaneously and that all storage pools involved have the same
device class. Each process requires two mount points and, if the device type is not
FILE, two drives. To run four migration processes simultaneously (two for each
storage pool), you need a total of at least eight mount points and eight drives if
the device type is not FILE. The device class must have a mount limit of at least
eight.

If the number of migration processes you specify is more than the number of
available mount points or drives, the processes that do not obtain mount points or
drives will wait indefinitely or until the other migration processes complete and
mount points or drives become available.

To specify one or more migration processes for each primary sequential-access

storage pool, use the MIGPROCESS parameter on the DEFINE STGPOOL and
UPDATE STGPOOL commands.

The Tivoli Storage Manager server starts the specified number of migration
processes regardless of the number of volumes that are eligible for migration. For
example, if you specify ten migration processes and only six volumes are eligible
for migration, the server will start ten processes and four of them will complete
without processing a volume.

Multiple concurrent migration processing does not affect collocation. If you specify
collocation and multiple concurrent processes, the Tivoli Storage Manager server
attempts to migrate the files for each collocation group, client node, or client file
space onto as few volumes as possible. If files are collocated by group, each
process can migrate only one group at a single time. In addition, if files belonging
to a single collocation group (or node or file space) are on different volumes and
are being migrated at the same time by different processes, the files could be
migrated to separate output volumes.

If simultaneous-write operations during migration are enabled during multiple

concurrent-migration processing, each process has the following requirements:
v A mount point
v A volume for each copy storage pool and active-data pool that is defined to the
target storage pool and the primary pool

For details about the simultaneous-write function, see “Writing data

simultaneously to primary, copy, and active-data pools” on page 331.

The effect of migration on copy storage pools and active-data

pools
Files in copy storage pools and active-data pools cannot be migrated. Migration of
files between primary storage pools does not affect copy storage pool files or
active-data pool files. Neither copy storage pool files nor active-data pool files
move when primary storage pool files move.

For example, suppose a copy of a file is made while it is in a disk storage pool.
The file then migrates to a primary tape storage pool. If you then back up the
primary tape storage pool to the same copy storage pool, a new copy of the file is
not needed. The server knows it already has a valid copy of the file.

Chapter 10. Managing storage pools and volumes 283

 The only way to store files in copy storage pools is by backing up (the BACKUP
STGPOOL command) or by using the simultaneous-write function. The only way to
store files in active-data pools is by copying active data (the COPY ACTIVEDATA
command) or by using the simultaneous-write function.

Caching in disk storage pools

When cache is enabled, the migration process leaves behind duplicate copies of
files after the server migrates these files to the next storage pool in the storage
hierarchy. Using cache can improve the speed with which the server retrieves some
files. Consider enabling cache for space-managed files that are frequently accessed
by clients.

About this task

If space is needed to store new data in the disk storage pool, cached files are
erased and the space they occupied is used for the new data.

Using cache has some important disadvantages:

v Using cache can increase the time required for client backup operations to
complete. Performance is affected because, as part of the backup operation, the
server must erase cached files to make room for storing new files. The effect can
be severe when the server is storing a very large file and must erase cached files.
For the best performance for client backup operations to disk storage pools, do
not use cache.
v Using cache can require more space for the server database. When you use
cache, more database space is needed because the server has to keep track of
both the cached copy of the file and the new copy in the next storage pool.
v If you want to use caching, you cannot also enable shredding for that disk
storage pool. See “Securing sensitive client data” on page 539 for more
information about shredding.

When cache is disabled and migration occurs, the server migrates the files to the
next storage pool and erases the files from the disk storage pool. By default, the
system disables caching for each disk storage pool because of the potential effects
of cache on backup performance. If you leave cache disabled, consider higher
migration thresholds for the disk storage pool. A higher migration threshold keeps
files on disk longer because migration occurs less frequently.

If fast restores of active client data is your objective, you can also use active-data
pools, which are storage pools containing only active versions of client backup
data. For details, see “Active-data pools” on page 241.

To enable cache, specify CACHE=YES when defining or updating a storage pool.

How the server removes cached files

When space is needed, the server reclaims space occupied by cached files. Files
that have the oldest retrieval date are overwritten first.

For example, assume that two files, File A and File B, are cached files that are the
same size. If File A was last retrieved on 05/16/08 and File B was last retrieved on
06/19/08, then File A is deleted to reclaim space first.

284 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 If you do not want the server to update the retrieval date for files when a client
restores or retrieves the file, specify the server option NORETRIEVEDATE in the
server options file. If you specify this option, the server removes copies of files in
cache regardless how recently the files were retrieved.

Effect of caching on storage pool statistics

The space-utilization statistic for the pool (Pct Util) includes the space used by any
cached copies of files in the storage pool. The migratable-data statistic (Pct Migr)
does not include space occupied by cached copies of files.

The server compares the migratable-data statistic with migration-threshold

parameters to determine when migration should begin or end. For more
information about storage pool statistics, see “Monitoring storage-pool and volume
usage” on page 382.

Deduplicating data
Data deduplication is a method for eliminating redundant data in order to reduce
the storage that is required to retain the data. Only one instance of the data is
retained in a deduplicated storage pool. Other instances of the same data are
replaced with a pointer to the retained instance.

About this task

Deduplicated data must be stored in sequential-access disk (FILE) primary, copy,

and active-data storage pools that you enable for data deduplication. Because you
can store more data on disk than on tape, data deduplication can reduce the
overall amount of time that is required to retrieve data.

Restriction: When a client backs up or archives a file, the data is written to the
primary storage pool specified by the copy group of the management class that is
bound to the data. To deduplicate the client data, the primary storage pool must be
a sequential-access disk (FILE) storage pool that is enabled for data deduplication.

Data deduplication overview

IBM Tivoli Storage Manager provides two options for performing data
deduplication; server-side data deduplication and client-side data deduplication.
Both methods use the same process to identify redundant data, however the time
and location of the deduplication processing is different. In server-side data
deduplication, processing takes place exclusively on the server after the data is
backed up. In client-side data deduplication, the processing is distributed between
the server and the backup-archive client during the backup process.

The ability to deduplicate data on either the backup-archive client or the server
provides flexibility in terms of resource utilization, policy management, and
security. You can also combine both client-side and server-side data deduplication
in the same production environment. For example, you can specify certain nodes
for client-side data deduplication and certain nodes for server-side data
deduplication. You can store the data for both sets of nodes in the same
deduplicated storage pool.

Backup-archive clients that can deduplicate data can also access data that was
deduplicated by server-side processes. Similarly, data that was deduplicated by
client-side processes can be accessed by the server. Furthermore, duplicate data can
be identified across objects regardless of whether the data deduplication is

Chapter 10. Managing storage pools and volumes 285

 performed on the client or the server.

Server-side data deduplication

Server-side data deduplication is a two-phase process. In the first phase, the server
identifies duplicate data. In the second phase, duplicate data is removed by certain
server processes.

In addition to whole files, IBM Tivoli Storage Manager can also deduplicate parts
of files that are common with parts of other files. Data becomes eligible for
duplicate identification as volumes in the storage pool are filled. A volume does
not have to be full before duplicate identification starts.

Duplicate data is removed by one of the following processes:

v Reclaiming volumes in the primary storage pool, copy storage pool, or
active-data pool
v Backing up a primary storage pool to a copy storage pool that is also set up for
data deduplication
v Copying active data in the primary storage pool to an active-data pool that is
also set up for data deduplication
v Migrating data from the primary storage pool to another primary storage pool
that is also set up for data deduplication
v Moving data from the primary storage pool to a different primary storage pool
that is also set up for data deduplication
v Moving data within the same copy storage pool or moving data within the same
active-data pool

Client-side data deduplication

In client-side data deduplication, the backup-archive client and the server identify
and remove duplicate data to save storage space on the server.

Benefits

Client-side data deduplication provides the following advantages:

v It can reduce the amount of data that is sent over the local area network (LAN).
v Extra processing power and time that is required to remove duplicate data on
the server is eliminated.
v Space savings occur immediately on the server because the client removed the
duplicated data.
v Extra reclamation processing is not required to remove the redundant data from
the server.
v It is no longer necessary to identify duplicates on the server.

Client-side data deduplication stores data directly in a deduplicated format. If

storage pool backup is used to create secondary copies to a non-deduplicated
storage pool, client extents are reassembled into contiguous files. (Extents are parts
of a file that are created during the data-deduplication process.) This reassembly
can cause storage pool backup processing to take longer when compared to
processing data that was not previously deduplicated.

Requirements

When you configure client-side data deduplication, the following requirements

must be met:

286 IBM Tivoli Storage Manager for Windows: Administrator's Guide

v The client and server must be at version 6.2.0 or later.
v Client-side deduplication cannot be used in combination with LAN-free backups.
v The primary storage pool must be a sequential-acces disk (FILE) storage pool
that is enabled for data deduplication.
v The value of the DEDUPLICATION option on the client must be set to yes. You can
set the DEDUPLICATION option in the client options file, in the preference editor of
the Tivoli Storage Manager client GUI, or in the client option set on the Tivoli
Storage Manager server.
v Client-side data deduplication must be enabled on the server by using the
DEDUPLICATION parameter on the REGISTER NODE or UPDATE NODE server command.
v Files that are intended for deduplication must not be excluded.
v Files that are intended for deduplication must not be encrypted. Encrypted files
and files from encrypted file systems cannot be deduplicated.

Configuration options for client-side deduplication

To take advantage of the client-side data deduplication feature, the following

options are available:
v Exclude specific files on a client from data deduplication by using the
[Link] client option.
v Enable a data deduplication cache, which reduces network traffic between the
client and the server. The cache on the client can be enabled through the client
options file.
Specify a size and location for a client cache.

Restriction: For applications that use the Tivoli Storage Manager API, do not
use the data deduplication cache because backup failures might occur when the
cache is out of sync with the Tivoli Storage Manager server. If multiple,
concurrent Tivoli Storage Manager client sessions are configured, you must
configure a separate cache for each session.
v Enable both client-side data deduplication and compression to reduce the
amount of data that is stored on the server. Each extent is compressed before it
is sent to the server. However, you must balance the benefits of storage savings
versus the processing power that is required to compress client data. In general,
if you compress and deduplicate data on the client system, you typically use
about twice as much processing power as data deduplication alone.
The server can process compressed data that has been deduplicated. In addition,
backup-archive clients earlier than V6.2 can restore deduplicated, compressed
data.

Client-side data deduplication and storage pools

If client-side data deduplication is enabled and the primary destination storage

pool is full, and another storage pool is in the hierarchy, the server stops the
transaction. Client-side data deduplication is disabled, and the client tries the
transaction again with files that are not deduplicated.

If the backup operation is successful and if the next storage pool is enabled for
data deduplication, the files are deduplicated by the server. If the next storage pool
is not enabled for data deduplication, the files are not deduplicated.

To ensure that client-side data deduplication can complete processing, maintain

sufficient free storage in your primary destination storage pool.

Chapter 10. Managing storage pools and volumes 287

 For details about client-side data deduplication, including options for controlling
data deduplication, see the Backup-Archive Clients Installation and User's Guide.

LAN-free access to storage pools that contain client-side deduplicated

data

Only V6.2 and later storage agents can use LAN-free data movement to access
storage pools that contain data that was deduplicated by clients. V6.1 storage
agents or later can complete operations over the LAN.
Table 24. Paths for data movement
Storage pool
Storage pool contains a mixture of Storage pool
contains only client-side and contains only
client-side server-side server-side
deduplicated data deduplicated data deduplicated data
V6.1 or earlier Over the LAN Over the LAN LAN-free
storage agent
V6.2 storage agent LAN-free LAN-free LAN-free

V6.2 backup-archive clients are compatible with V6.2 storage agents and provide
LAN-free access to storage pools that contain client-side deduplicated data.
Related concepts:
“Server-side data deduplication” on page 286

Data deduplication limitations

Before implementing data deduplication, be aware that certain limitations apply.

Version support

Server-side data deduplication is available only with IBM Tivoli Storage Manager
V6.1 or later servers. For optimal efficiency when using server-side data
deduplication, upgrade to the backup-archive client V6.1 or later.

Client-side data deduplication is available only with Tivoli Storage Manager V6.2
or later servers and backup-archive clients V6.2 or later.

Eligible storage pools

Data on random-access disk or on tape cannot be deduplicated. Only data in

storage pools that are associated with sequential-access disk devices (FILE) can be
deduplicated. You must enable FILE storage pools for data deduplication.

Client files must be bound to a management class that specifies a

deduplication-enabled storage pool.

Encrypted files

The Tivoli Storage Manager server and the backup-archive client cannot
deduplicate encrypted files. If an encrypted file is encountered during data
deduplication processing, the file is not deduplicated, and a message is logged.

288 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Tip: You do not have to process encrypted files separately from files that are
eligible for client-side data deduplication. Both types of files can be processed in
the same operation. However, they are sent to the server in different transactions.

As a security precaution, you can take one or more of the following steps:
v Enable storage-device encryption together with client-side data deduplication.
v Use client-side data deduplication only for nodes that are secure.
v If you are uncertain about network security, enable Secure Sockets Layer (SSL).
v If you do not want certain objects (for example, image objects) to be processed
by client-side data deduplication, you can exclude them on the client. If an
object is excluded from client-side data deduplication and it is sent to a storage
pool that is set up for data deduplication, the object is deduplicated on server.
v Use the SET DEDUPVERIFICATIONLEVEL command to detect possible security
attacks on the server during client-side data deduplication. Using this command,
you can specify a percentage of client extents for the server to verify. If the
server detects a possible security attack, a message is displayed.

File size

Only files that are more than 2 KB are deduplicated. Files that are 2 KB or less are
not deduplicated.

Operations that preempt client-side data deduplication

The following operations take precedence over client-side data deduplication:

v LAN-free data movement
v Subfile backup operations
v Simultaneous-write operations
v Server-initiated sessions

Do not schedule or enable any of those operations during client-side data

deduplication. If any of those operations occur during client-side data
deduplication, client-side data deduplication is turned off, and a message is issued
to the error log.

The following is an example of how simultaneous operations can create problems

when using client-side data deduplication.

When client-side data deduplication is used to back up or archive files, it is

possible for a server operation to mark a file extent on the server as expired, before
the client operation completes a backup or an archive that involves the same
extent. If this happens, the client is notified of the expired extent. The client makes
three attempts to copy the extent to the server before it gives up and processes the
other files in the backup operation.

A return code (RC=254) and message are written to the [Link] file. The
message is also displayed in the command-line client. The error message is:
ANS7899E The client referenced a duplicated extent that does not exist
on the Tivoli Storage Manager server.

The workaround for this situation is to ensure that processes that can cause files to
expire are not run at the same time that back up or archive operations with
client-side data deduplication are performed.

Chapter 10. Managing storage pools and volumes 289

 Data deduplication of hierarchical storage management data

Microsoft Windows HSM uses the Tivoli Storage Manager application

programming interface (API), which can deduplicate client data. Server-side
deduplication of HSM data from Microsoft Windows clients is allowed.

Collocation

You can use collocation for storage pools that are set up for data deduplication.
However, collocation might not have the same benefit as it does for storage pools
that are not set up for data deduplication.

By using collocation with storage pools that are set up for data deduplication, you
can control the placement of data on volumes. However, the physical location of
duplicate data might be on different volumes. No-query-restore, and other
processes remain efficient in selecting volumes that contain non-deduplicated data.
However, the efficiency declines when additional volumes are required to provide
the duplicate data.

Moving or copying data from a deduplicated storage pool to a

non-deduplicated storage pool

When you copy or move data from a deduplicated storage pool to a

non-deduplicated storage pool, the data is reconstructed. However, after the data
movement or copy operation, the amount of data that is reported as moved or
copied is the amount of deduplicated data. For example, suppose that a storage
pool contains 20 GB of deduplicated data that represents 50 GB of total file data. If
the data is moved or copied, the server reports that 20 GB was moved or copied,
even though 50 GB of data was sent.

When data deduplication is not appropriate

Using Tivoli Storage Manager data deduplication can provide several advantages.
However, there are some situations where data deduplication is not appropriate.
Those situations are:
v Your primary storage of backup data is on a Virtual Tape Library or physical
tape. If regular migration to tape is required, the benefits of using data
deduplication are lessened, since the purpose of data deduplication is to reduce
disk storage as the primary location of backup data.
v You have no flexibility with the backup processing window. Tivoli Storage
Manager data deduplication processing requires additional resources, which can
extend backup windows or server processing times for daily backup activities.
v Your restore processing times must be fast. Restore performance from
deduplicated storage pools is slower than from a comparable disk storage pool
that does not use data deduplication. If fast restore performance from disk is a
high priority, restore performance benchmarking must be done to determine
whether the effects of data deduplication can be accommodated.
Related tasks:
“Keeping client files together using collocation” on page 357
“Detecting possible security attacks during client-side data deduplication” on page
302

290 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Planning guidelines for data deduplication
Planning for data deduplication is important because there are many factors to
consider. These factors include database and log size capacity, data deduplication
location, storage pool setup, and additional hardware requirements.

As part of the planning process, ensure that you will benefit from using data
deduplication. In the following situations, IBM Tivoli Storage Manager data
deduplication can provide a cost-effective method for reducing the amount of disk
storage that is required for backups:
v You have to reduce the disk space that is required for backup storage.
v You must perform remote backups over limited bandwidth connections.
v You are using Tivoli Storage Manager node replication for disaster recovery
across geographically dispersed locations.
v You either have disk-to-disk backup configured (where the final destination of
backup data is on a deduplicating disk storage pool), or data is stored in the
FILE storage pool for a significant time (for example 30 days), or until
expiration.
v For guidance on the scalability of data deduplication with Tivoli Storage
Manager, see Effective Planning and Use of IBM Tivoli Storage Manager V6
Deduplication at [Link]
wikis/home/wiki/Tivoli Storage Manager/page/Effective Planning and Use of
IBM Tivoli Storage Manager V6 Deduplication.
Related concepts:
“Example: Estimating active and archive log sizes for data deduplication
operations” on page 663
Related tasks:
“Detecting possible security attacks during client-side data deduplication” on page
302

Implementing data deduplication

In order to implement data deduplication, you must make several decisions that
are based on the outcome that you want to achieve.

About this task

v Determine database capacity requirements. When you use data deduplication,
considerably more database space is required as a result of storing the metadata
that is related to duplicate data.
v Determine database log size requirements. It is essential that you properly size
the storage capacity for the database active log and archive log.
v Determine which client nodes have data that you want to deduplicate.
v Determine whether you want to implement server-side data deduplication,
client-side data deduplication, or a combination of both. To help you make that
determination, consider the following factors:
– Server-side data deduplication is a two-step process in which duplicate data
is identified and then storage space is reclaimed to remove the duplicate data.
Client-side data deduplication stores the data directly in a deduplicated
format.
– Data deduplication and data compression can be combined on the
backup-archive client to reduce data storage. This reduction is typically more
than you can achieve by using server-side data deduplication alone.
– If bandwidth is not restrictive, client-side data deduplication processing
typically causes an increase in time for backup operations to complete.
Chapter 10. Managing storage pools and volumes 291
 Consider doubling the time that you allow for backups when you use
client-side data deduplication in an environment that is not limited by the
network. If you are creating a secondary copy by using storage pool backup,
where the copy storage pool is not using data deduplication, it takes longer
for data to be moved because of the extra processing that is required to
reconstruct the deduplicated data.
– Duplicate identification processing is handled by client systems when
client-side deduplication is used. However, the Tivoli Storage Manager server
still requires processing to handle the lookup requests from clients and to
store data deduplication metadata that is produced by client systems.
– Client-side data deduplication cannot be combined with LAN-free data
movement that uses the Tivoli Storage Manager for Storage Area Networks
feature. If you are implementing a Tivoli Storage Manager supported
LAN-free to disk solution, consider server-side data deduplication.
v If you choose client-side data deduplication, decide what, if any, security
precautions to take.
v Decide whether you want to define a new storage pool exclusively for data
deduplication or update an existing storage pool. The storage pool must be a
sequential-access disk (FILE) pool. Data deduplication occurs at the storage-pool
level, and all data within a storage pool, except encrypted data, is deduplicated.
v If you want to implement server-side data deduplication, decide how best to
control duplicate-identification processes. For example, you might want to run
duplicate-identification processes automatically all the time. Alternatively, you
might want to start and stop duplicate-identification processes manually. You
can also start duplicate-identification processes automatically and then increase
or decrease the number of processes depending on your server workload.
Whatever you decide, you can change the settings later, after the initial setup, to
meet the requirements of your operations.
The following table lists the options that you can use to control duplicate
identification processes.
Table 25. Options for controlling duplicate-identification processes
If you create a storage pool for data
deduplication... If you update an existing storage pool...
You can specify 1 - 20 duplicate- You can specify 0 - 20 duplicate-
identification processes to start identification processes to start
automatically. The Tivoli Storage Manager automatically. If you do not specify any
server does not start any processes if you duplicate-identification processes, you must
specify zero. start and stop processes manually.

If you are creating a primary The Tivoli Storage Manager server does not
sequential-access storage pool and you do start any duplicate-identification processes
not specify a value, the server starts one automatically by default.
process automatically. If you are creating a
copy storage pool or an active-data pool and
you do not specify a value, the server does
not start any processes automatically.

After the storage pool is created, you can

increase and decrease the number of
duplicate-identification processes manually.
You can also start, stop, and restart
duplicate-identification processes manually.

v Decide whether to define or update a storage pool for data deduplication, but
not actually perform data deduplication. For example, suppose that you have a

292 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 primary sequential-access disk storage pool and a copy sequential-access disk
storage pool. Both pools are set up for data deduplication. You might want to
run duplicate-identification processes for only the primary storage pool. In this
way, only the primary storage pool reads and deduplicates data. However, when
the data is moved to the copy storage pool, the data deduplication is preserved,
and no duplicate identification is required.
v Determine the best time to use data deduplication for the storage pool. The
duplicate identification (IDENTIFY) processes can increase the workload on the
processor and system memory. Schedule duplicate identification processes at the
following times:
– When the process does not conflict with other processes such as reclamation,
migration, and storage pool backup
– Before node replication (if node replication is being used) so that node
replication can be used in combination with deduplication
Related tasks:
“Detecting possible security attacks during client-side data deduplication” on page
302

Checklist for data deduplication

Data deduplication requires additional processing resources on the server or client.
Use the checklist to verify that hardware and your Tivoli Storage Manager
configuration have characteristics that are key to good performance.

Tasks, characteristics, options, or

Question settings More information
Are you using fast disk storage for Use high-performance disk for the Checklist for server database disks
the Tivoli Storage Manager database Tivoli Storage Manager database. At
as measured in terms of a minimum, use 10000-rpm drives for
input/output operations per second smaller databases that are 200 GB or
(IOPS)? less. For databases over 500 GB, use
15000-rpm drives or solid-state
drives.

Tivoli Storage Manager database

should have a minimum capability of
3,000 IOPS. For each TB of data that
is backed up daily (before data
deduplication), include an additional
1,000 IOPS to this minimum.

For example, a Tivoli Storage

Manager server that is ingesting 3 TB
of data per day would need 6,000
IOPS for the database disks:
3,000 IOPS minimum + 3,000 (3
TB x 1,000 IOPS) = 6,000 IOPS
Do you have enough memory for Use a minimum of 64 GB of system
the size of your database? memory for Tivoli Storage Manager
servers that are deduplicating data. If
the retained capacity of backup data
grows, the memory requirement
might need to be as high as 128 GB.

Monitor memory usage regularly to

determine whether more memory is
required.

Chapter 10. Managing storage pools and volumes 293

 Tasks, characteristics, options, or
Question settings More information
Have you properly sized your disk For a rough estimate, plan for 150 GB
space for the database, logs, and of database storage for every 10 TB
storage pools? of data that is to be protected in
deduplicated storage pools. Protected
data is the amount of data before
deduplication, including all versions
of objects stored.

Configure the server to have the

maximum active log size of 128 GB
by setting the ACTIVELOGSIZE server
option to a value of 131072.

Use a directory for the database

archive logs with an initial free
capacity of at least 500 GB. Specify
the directory by using the
ARCHLOGDIRECTORY server option.

Define space for the archive failover

log by using the
ARCHFAILOVERLOGDIRECTORY server
option.
Are the Tivoli Storage Manager The Tivoli Storage Manager database See Server database and recovery log
database and logs on separate disk must not share disk volumes with configuration and tuning
volumes (LUNs)? Tivoli Storage Manager database logs
or storage pools, or with any other
Is the disk that is used for the application or file system.
database configured according to
best practices for a transactional
database?
Are you using a minimum of 8 (2.2 If you are planning to use client-side
GHz or equivalent) processor cores data deduplication, verify that client
for each Tivoli Storage Manager systems have adequate resources
server that you plan to use with available during a backup operation
data deduplication? to perform data deduplication
processing. Use a processor that is at
least the minimum equivalent of one
2.2 GHz processor core per backup
process with client-side data
deduplication.

294 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 Tasks, characteristics, options, or
Question settings More information
Have you estimated storage pool You can estimate storage pool For an example of using this
capacity to configure enough space capacity requirements for a technique, see Effective Planning and
for the size of your environment? deduplicated storage pool by using Use of IBM Tivoli Storage Manager V6
the following technique: Deduplication at [Link]
1. Estimate the base size of the developerworks/mydeveloperworks/
source data. wikis/home/wiki/Tivoli Storage
Manager/page/Effective Planning and
2. Estimate the daily backup size by
Use of IBM Tivoli Storage Manager V6
using an estimated change and
Deduplication.
growth rate.
3. Determine retention requirements.
4. Estimate the total amount of
source data by factoring in the
base size, daily backup size, and
retention requirements.
5. Apply the deduplication ratio
factor.
6. Round up the estimate to
consider transient storage pool
usage.
Have you distributed disk I/O over Use arrays that consist of as many See Checklist for storage pools on
many disk devices and controllers? disks as possible, which is sometimes disk.
referred to as wide striping.

Specify 8 or more file systems for the

deduplicated storage pool device
class so that I/O is distributed across
as many LUNs and physical devices
as possible.
Do you have adequate resources to If you are planning to use client-side
cover client-side data deduplication data deduplication, verify that client
requirements? systems have adequate resources
available during a backup operation
to perform data deduplication
processing. Use a processor that is at
least the minimum equivalent of one
2.2 GHz processor core per backup
process with client-side data
deduplication.

Best practices for data deduplication

Implement data deduplication in IBM Tivoli Storage Manager by using the best
practices to ensure that the benefits of data deduplication are optimized on your
system.

When you use data deduplication, your system can achieve benefits such as these:
v Reduction in the storage capacity that is required for storage pools on the server
that are associated with a FILE-type device class. This reduction applies for both
server-side and client-side data deduplication.
v Reduction in the network traffic between the client and server. This reduction
applies for client-side deduplication only.

Using data deduplication requires an increase in resources on the Tivoli Storage

Manager server and clients. The upper limit on the size of objects that are stored in

Chapter 10. Managing storage pools and volumes 295

 deduplicated storage pools is one primary consideration. However, in general, a
server that uses data deduplication must be allocated additional resources.

When you implement the suggested practices for data deduplication, you can help
to avoid problems such as these on your system:
v Server outages that are caused by running out of active log space or archive log
space
v Server outages or client backup failures that are caused by exceeding the IBM
DB2 internal lock list limit
v Process failures and hangs that are caused during server data management

Properly size the server database, recovery log, and system memory:

When you use data deduplication, considerably more database space is required as
a result of storing the metadata that is related to duplicate data. Data
deduplication also tends to cause longer-running transactions and a related larger
peak in recovery log usage.

In addition, more system memory is required for caching database pages that are
used during duplicate data lookup for both server-side and client-side data
deduplication.

Tips:
v Ensure that the Tivoli Storage Manager server has a minimum of 64 GB of
system memory.
v Allocate a file system with two-to-three times more capacity for the server
database than you would allocate for a server that does not use data
deduplication. You can plan for 150 GB of database storage for every 10 TB of
data that is protected in the deduplicated storage pools.
v Configure the server to have the maximum active log size of 128 GB by setting
the ACTIVELOGSIZE server option to a value of 131072.
v Use a directory for the database archive logs with an initial free capacity of at
least 500 GB. Specify the directory by using the ARCHLOGDIRECTORY server option.

For more information about managing resources such as the database and recovery
log, see the Installation Guide. Search for database and recovery log capacity
planning.

Avoid the overlap of server maintenance tasks with client backup windows:

When you schedule client backups for a period during which server maintenance
tasks are not running, you create a backup window. This practice is important when
you use data deduplication. Use this process regardless of whether data
deduplication is used with Tivoli Storage Manager.

Perform server maintenance tasks in a sequence that avoids contention between

the different types of processing. The server maintenance tasks are:
v Migration
v Storage pool backup
v Duplicate identification
v Database backup
v Node replication
v Reclamation

296 IBM Tivoli Storage Manager for Windows: Administrator's Guide

v Expiration

Migration and reclamation are the tasks most likely to interfere with the success of
client backups.

Tips:
v Schedule client backups in a backup window that is isolated from data
maintenance processes, such as migration and reclamation.
v Schedule each type of data maintenance task with controlled start times and
durations so that they do not overlap with each other.
v If storage-pool backup is used to create a secondary copy, schedule storage-pool
backup operations before you start data deduplication processing to avoid
restoring objects that are sent to a non-deduplicated copy storage pool.
v If you are using node replication to keep a secondary copy of your data,
schedule the REPLICATE NODE command to run after duplicate identification
processes are completed.

For more information about tuning the schedule for daily server maintenance
tasks, see the Optimizing Performance guide. Search for tuning the schedule for daily
operations.

Modify DB2 lock list management:

The lock list storage of DB2 that is automatically managed can become insufficient.
If you deduplicate data that includes large files or large numbers of files
concurrently, the data deduplication can cause insufficient storage. When the lock
list storage is insufficient, backup failures, data management process failures, or
server outages can occur.

File sizes greater than 500 GB that are processed by data deduplication are most
likely to cause storage to become insufficient. However, if many backups use
client-side data deduplication, this problem can also occur with smaller-sized files.

The following activities can use more lock list storage:

v Client backups that use client-side data deduplication
v Data movement within a deduplicated storage pool that uses the reclamation
and MOVE DATA commands
v Data movement out of a deduplicated storage pool that uses the migration and
MOVE DATA commands

See Technote 1430874 at [Link]

[Link]?uid=swg21430874 for an explanation about how to estimate the peak
volume of deduplication transactions that are processed. This document also
contains information about the corresponding lock list requirements for managing
the volume and information about how to change the DB2 limit, if necessary.

Tip: When you estimate the lock list storage requirements, follow the information
described in the technote to manage storage for loads that are much larger than
expected.

Chapter 10. Managing storage pools and volumes 297

 Limit the impact of large objects on data deduplication processing:

You can use controls to limit the potential effect of large objects on data
deduplication processing on the Tivoli Storage Manager server.

You can use the following controls when you deduplicate large-object data:
v Server controls that limit the size of objects. These controls limit the size of
objects that are processed by data deduplication.
v Controls on the data management processes of the server. These controls limit
the number of processes that can operate concurrently on the server.
v Scheduling options that control how many clients run scheduled backups
simultaneously. These scheduling options can be used to limit the number of
clients that perform client-side data deduplication at the same time.
v Client controls whereby larger objects can be processed as a collection of smaller
objects. These controls are primarily related to the Tivoli Storage Manager data
protection products.

Server controls to limit the data deduplication of large objects:

Use the server controls that are available on Tivoli Storage Manager server to
prevent large objects from being processed by data deduplication.

Use the following parameter and server options to limit the object size for data
deduplication:
MAXSIZE
For storage pools, the MAXSIZE parameter can be used to prevent large
objects from being stored in a deduplicated storage pool. Use the default
NOLIMIT parameter value, or set the value to be greater than
CLIENTDEDUPTXNLIMIT and SERVERDEDUPTXNLIMIT option values.
Use the MAXSIZE parameter with a deduplicated storage pool to prevent
objects that are too large to be eligible for data deduplication from being
stored in a deduplicated storage pool. The objects are then redirected to the
next storage pool in the storage pool hierarchy.
SERVERDEDUPTXNLIMIT
The SERVERDEDUPTXNLIMIT server option limits the total size of objects that
can be deduplicated in a single transaction by duplicate identification
processes. This option limits the maximum file size that is processed by
server-side data deduplication. The default value for this option is 300 GB,
and the maximum value is 2048 GB. Because less simultaneous activity is
typical with server-side data deduplication, consider having a limit larger
than 300 GB on the object size for server-side data deduplication.
CLIENTDEDUPTXNLIMIT
The CLIENTDEDUPTXNLIMIT server option restricts the total size of all objects
that can be deduplicated in a single client transaction. This option limits
the maximum object size that is processed by client-side data
deduplication. However, there are some methods to break up larger
objects. The default value for this option is 300 GB, and the maximum
value is 1024 GB.

Tips:
v Set the MAXSIZE parameter for deduplicated storage pools to a value slightly
greater than CLIENTDEDUPTXNLIMIT and SERVERDEDUPTXNLIMIT option values.

298 IBM Tivoli Storage Manager for Windows: Administrator's Guide

v If you increase CLIENTDEDUPTXNLIMIT or SERVERDEDUPTXNLIMIT option values
beyond the defaults, reevaluate your sizing or the server recovery log. Also,
reevaluate your sizing for the IBM DB2 lock list.
v If you plan to run many simultaneous client backups that use client-side data
deduplication, consider lowering the setting of the CLIENTDEDUPTXNLIMIT option
to the minimum setting of 32 GB. Lower the setting if large objects will not be
deduplicated.

Controls for data management processes:

Use the controls for the data management processes of the Tivoli Storage Manager
server. These controls limit the number of large objects that are simultaneously
processed by the server during data deduplication.

Use the following commands and parameters to limit the number of large objects
that are simultaneously processed by the server:
v The storage pool parameters on the DEFINE STGPOOL command or the UPDATE
STGPOOL command.
– The MIGPROCESS parameter controls the number of migration processes for a
specific storage pool.
– The RECLAIMPROCESS parameter controls the number of simultaneous processes
that are used for reclamation.
v The IDENTIFYPROCESS parameter on the IDENTIFY DUPLICATES command. The
parameter controls the number of duplicate identification processes that can run
at one time for a specific storage pool.

Tips:
v You can safely run duplicate identification processes for more than one
deduplicated storage pool at the same time. However, specify the
IDENTIFYPROCESS parameter with the IDENTIFY DUPLICATES command to limit the
total number of all simultaneous duplicate identification processes. Limit the
total number to a number less than or equal to the number of processors that are
available in the system.
v Schedule duplicate identification processes to run when the additional load does
not affect client operations or conflict with other server processes. For example,
schedule the duplicate identification process to run outside the client backup
window. The duplicate identification processes for the server intensively use the
database and system resources. These processes place additional processing on
the processor and memory of the system.
v Do not overlap different types of operations, such as expiration, reclamation,
migration, and storage pool backup.
v Read the information about data deduplication and the server storage pool. The
effect of data deduplication on system resources is also related to the size of the
file for deduplication. As the size of the file increases, more processing time,
processor resources, memory, and active log space are needed on the server.
Review the document for information about data deduplication and the server
storage pool.

Chapter 10. Managing storage pools and volumes 299

 Schedule controls for client backups:

For scheduled backups, you can limit the number of client backup sessions that
perform client-side data deduplication at the same time.

You can use any of the following approaches to limit the number of client backup
sessions:
v Clients can be clustered in groups by using different schedule definitions that
run at different times during the backup window. Consider spreading clients
that use client-side deduplication among these different groups.
v Increase the duration for scheduled startup windows and increase the
randomization of schedule start times. This limits the number of backups that
use client-side data deduplication that start at the same time.
v Separate client backup destinations by using the server policy definitions of the
Tivoli Storage Manager server, so that different groups of clients use different
storage pool destinations:
– Clients for which data is never to be deduplicated cannot use a management
class that has as its destination a storage pool with data deduplication
enabled.
– Clients that use client-side data deduplication can use storage pools where
they are matched with other clients for which there is a higher likelihood of
duplicate matches. For example, all clients that run Microsoft Windows
operating systems can be set up to use a common storage pool. However,
they do not necessarily benefit from sharing a storage pool with clients that
perform backups of Oracle databases.

Client controls to limit the data deduplication of large objects:

A primary source of large objects that are processed by client-side data

deduplication are backups by Tivoli Storage Manager data protection products.
You can use controls to have the clients break these objects into multiple smaller
objects that are within the object size limits for data deduplication.

Many of the data protection products process objects with sizes in the range of
several hundred GBs to one TB. This range exceeds the maximum object size that
is acceptable for data deduplication.

You can reduce large objects into multiple smaller objects by using the following
methods:
v Use Tivoli Storage Manager client features that back up application data with
the use of multiple streams. For example, a 1 TB database is not eligible for data
deduplication as a whole. However, when backed up with four parallel streams,
the resulting four 250 GB objects are eligible for deduplication. For Tivoli Storage
Manager Data Protection for SQL, you can specify a number of stripes to change
the backup into multiple streams.
v Use application controls that influence the maximum object size that is passed
through to Tivoli Storage Manager. Tivoli Storage Manager Data Protection for
Oracle has several RMAN configuration parameters that can cause larger
databases to be broken into smaller objects. These configuration parameters
include the use of multiple channels, or the MAXPIECESIZE option, or both.

Restriction: In some cases, large objects cannot be reduced in size, and therefore
cannot be processed by Tivoli Storage Manager data deduplication:

300 IBM Tivoli Storage Manager for Windows: Administrator's Guide

v The backup-archive clients always send large files in a single transaction, which
cannot be broken into smaller transactions.
v Image backups of a large file system are sent within a single transaction and
cannot be broken into smaller components.

Performance considerations for data deduplication

Finding duplicate data is a processor-intensive process. When you use client-side
data deduplication, the processor consumption occurs on the client system during
the backup. With server-side data deduplication, the processor consumption occurs
on the server during the duplicate identification process. Consider factors such as
processor usage, network bandwidth, restore performance, and compression when
you decide to use data deduplication.

Processor usage

The amount of processor resources that are used depends on how many client
sessions or server processes are simultaneously active. Additionally, the amount of
processor usage is increased because of other factors, such as the size of the files
that are backed up. When I/O bandwidth is available and the files are large, for
example 1 MB, finding duplicates can use an entire processor during a session or
process. When files are smaller, other bottlenecks can occur. These bottlenecks can
include reading files from the client disk or the updating of the Tivoli Storage
Manager server database. In these bottleneck situations, data deduplication might
not use all of the resources of the processor.

You can control processor resources by limiting or increasing the number of client
sessions for a client or a server duplicate identification processes. To take
advantage of your processor and to complete data deduplication faster, you can
increase the number of identification processes or client sessions for the client. The
increase can be up to the number of processors that are on the system. It can be
more than that number if the processors support multiple hardware-assisted
threads for the core, such as with simultaneous multithreading. Consider a
minimum of at least 8 (2.2Ghz or equivalent) processor cores in any Tivoli Storage
Manager server that is configured for data deduplication.

Client-side data deduplication can use a large amount of processor resources.

Therefore, verify that the additional workload does not affect the primary
workload of the client system.

Compressing the data, in addition to deduplicating it on the client, uses additional

processor resources. However, it lowers the network bandwidth that is required if
the data is compressible.

Network bandwidth

A primary reason to use client-side data deduplication is to reduce the bandwidth

that is required to transfer data to a Tivoli Storage Manager server. Client
compression can reduce this bandwidth further. The amount that the bandwidth is
reduced by is directly related to how much of the data is duplicate that is already
stored on the server. It is also directly related to how compressible this data is.

Network bandwidth for the queries for data from the Tivoli Storage Manager client
to the server can be reduced by using the enablededupcache client option. The
cache stores information about extents that have been previously sent to the server.
If an extent is found that was previously sent, it is not necessary to query the

Chapter 10. Managing storage pools and volumes 301

 server again for that extent. Therefore, bandwidth and performance are not
additionally reduced.

Restore performance

During a restore operation, performance for a deduplicated storage pool can be

slower than a restore from a non-deduplicated pool. When data deduplication is
used, the extents for a given file can be spread across multiple volumes on the
server. This spreading of extents makes the reads from the volumes more random
in nature, and also slower than during a sequential operation. In addition, more
database operations are required.

Compression

Data deduplication is not performed on directories or file metadata. In comparison,

compression can be performed on these types of data. Therefore, the reduction
percentages do not typically add up to the total data-reduction percentage. When
client-side data deduplication is used, the calculation of the compression-reduction
percentage is performed differently. The calculation includes only the actual data
reduction that results from use of the compression engine. Client-side data
deduplication, when performed, occurs before compression.

For the quickest backups on an unconstrained network, choose server-side data

deduplication. For the largest storage savings, choose client-side data deduplication
combined with compression. Avoid performing client-compression in combination
with server-side data deduplication.

For additional information about data deduplication and performance

considerations, see the Optimizing Performance guide.

Detecting possible security attacks during client-side data

deduplication
A rogue application that is part of a client system and that imitates the client, API,
or GUI application can initiate an attack on the server. To reduce server
vulnerability to such attacks, you can specify a percentage of client extents for the
server to verify.

About this task

If the server detects that a security attack is in progress, the current session is
canceled. In addition, setting of the node DEDUPLICATION parameter is changed from
CLIENTORSERVER to SERVERONLY. The SERVERONLY setting disables client-side data
deduplication for that node.

The server also issues a message that a potential security attack was detected and
that client-side data deduplication was disabled for the node.

If client-side data deduplication is disabled, all other client operations (for

example, backup operations) continue. Only the client-side data deduplication
feature is disabled. If client-side data deduplication is disabled for a node because
a potential attack was detected, the server deduplicates the data that is eligible for
client-side data deduplication.

302 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 Procedure

To detect a possible security attack when client-side data deduplication is enabled,

issue the SET DEDUPVERIFICATIONLEVEL command. Specify an integer value 1 - 100
to indicate the percentage of client extents to be verified. The default value is 0.
This value indicates that no extents are verified.

What to do next

Tip: Verifying extents consumes processing power and adversely affects server
performance. For optimal performance, do not specify values greater than 10 for
the SET DEDUPVERIFICATIONLEVEL command. Other methods for protecting the
server include:
v Enabling client-side data deduplication only for clients that are secure. If you
choose this method, do not change the default setting of SET
DEDUPVERIFICATIONLEVEL command.
v Creating automated scripts to enable client-side data deduplication only during
certain time periods.
v Using storage-device encryption together with client-side data deduplication
v Enabling Secure Sockets Layer (SSL)
v Deduplicating data that uses only server-side data deduplication. Server-side
data deduplication does not expose the server to security attacks from the client.

To show the current value for SET DEDUPVERIFICATIONLEVEL, issue the QUERY STATUS
command. Check the value in the Client-side Deduplication Verification Level
field.

Evaluating data deduplication in a test environment

Testing can give you important information about the possible benefits of
server-side and client-side data deduplication in your production environment.
Space savings and restore-and-retrieve times are two key indicators that you can
test.

Restore and retrieve operations from server-side and client-side

deduplicated storage pools
Restore-and-retrieve operations from a sequential-access disk (FILE) storage pool
that is set up for data deduplication have different performance characteristics than
restore-and-retrieve operations from a FILE storage pool that is not set up for data
deduplication. To ensure that performance objectives can be met, test your restore
scenarios.

In a FILE storage pool that is not set up for data deduplication, files on a volume
that are being restored or retrieved are read sequentially from the volume before
the next volume is mounted. This process ensures optimal I/O performance and
eliminates the need to mount a volume multiple times.

In a FILE storage pool that is set up for data deduplication, however, extents that
comprise a single file can be distributed across multiple volumes. To restore or
retrieve the file, each volume containing a file extent must be mounted. As a result,
the I/O is more random, which can lead to slower restore-and-retrieve times.
These results occur more often with small files that are less than 100 KB. In
addition, more processor resources are consumed when restoring or retrieving
from a deduplicated storage pool. The additional consumption occurs because the
data is checked to ensure that it has been reassembled properly.

Chapter 10. Managing storage pools and volumes 303

 Although small-file, restore-and-retrieve operations from a deduplicated storage
pool might be relatively slow, these operations are still typically faster than
small-file restore-and-retrieve operations from tape because of the added tape
mount-and-locate time. If you have data for which fastest-possible
restore-and-retrieval time is critical, you can use a sequential-access disk storage
pool that is not set up for data deduplication.

Tip: To reduce the mounting and removing of FILE storage pool volumes, the
server allows for multiple volumes to remain mounted until they are no longer
needed. The number of volumes that can be mounted at a time is controlled by the
NUMOPENVOLSALLOWED option.

Estimating space savings from server-side data deduplication

Before you set up data deduplication in your production environment, you can
estimate the amount of storage space that can be saved. Directions are provided for
backing up the data in a primary storage pool to a temporary copy storage pool
that is set up for data deduplication.

Procedure

To estimate space savings, complete the following steps:

1. Create a sequential-access disk (FILE) copy storage pool and enable the pool for
data deduplication.
2. Back up the contents of the primary storage pool that you want to test to the
copy storage pool.
3. Run the duplicate-identification processes against the volumes in the copy
storage pool.
If you specified one or more duplicate-identification processes when you
created the copy storage pool, those processes start automatically. If you did
not specify any processes, you must specify and start duplicate-identification
processes manually.
4. After all the data in the copy storage pool is identified, start reclamation by
changing the reclamation percentage on the copy storage pool to 1%.
5. When reclamation finishes, use the QUERY STGPOOL command to check the copy
storage-pool statistics to determine the amount of space that was saved.

Results

If the results are satisfactory, complete one of the following tasks:

v If the primary storage pool is a sequential-access disk storage pool, update the
storage, specifying data deduplication.
v If the primary storage pool is not a sequential-access disk storage pool, create a
new primary sequential-access disk storage pool, specifying data deduplication.
Move the data or migrate the data from the original storage pool to the new
storage pool.

304 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Verifying your implementation
After you configured data deduplication, there are some actions that you can take
to verify whether your solution is processing effectively.

About this task

Use the following questions to evaluate your data deduplication configuration.

Have you scheduled data deduplication processing based on your backup
strategy?
If you are not creating a secondary copy of backup data or if you are using
node replication for the second copy, client backup and duplicate
identification can be overlapped. This can reduce the total elapsed time for
these operations, but might increase the time that is required for client
backup.
If you are using storage pool backup, do not overlap client backup and
duplicate identification. The best practice sequence of operations is client
backup, storage pool backup, and then duplicate identification.
For data that is not stored with client-side data deduplication, schedule
storage-pool backup operations to complete before you start data
deduplication processing. Set up your schedule this way to avoid
reconstructing objects that are deduplicated to make a non-deduplicated
copy to a different storage pool.
See Scheduling data deduplication and node replication processes for more
information.
Are the processes for identifying duplicates able to handle all new data that is
backed up each day?
If the process completes, or goes into an idle state before the next
scheduled operation begins, then all new data is being processed.
Is reclamation able to run to a sufficiently low threshold?
If a low threshold cannot be reached, consider the following actions:
v Increase the number of processes that are used for reclamation.
v Upgrade to faster hardware.
Is deduplication cleanup processing able to clean out the dereferenced extents to
free disk space before the start of the next backup cycle?
Run the SHOW DEDUPDELETE command. The output shows that all threads
are idle when the workload is complete.
If cleanup processing cannot complete, consider the following actions:
v Increase the number of processes that are used for duplicate
identification.
v Upgrade to faster hardware.
v Determine if you are ingesting more data that your Tivoli Storage
Manager server can process with data deduplication and consider
deploying an additional Tivoli Storage Manager server.
Related reference:
“Checklist for data deduplication” on page 293

Chapter 10. Managing storage pools and volumes 305

 Managing deduplication-enabled storage pools
You can create a storage pool for data deduplication or you can update an existing
storage pool. If you are implementing server-side data deduplication, IBM Tivoli
Storage Manager provides the option of running duplicate-identification processes
automatically or manually.

Before you begin

Before you set up a storage pool:

v Determine which client nodes have data that you want to deduplicate. Decide
whether you want to deduplicate data on a node-by-node basis, on either the
client or the server.
v Decide whether you want to define a new storage pool exclusively for data
deduplication or update an existing storage pool. If you update a storage pool
for data deduplication, Tivoli Storage Manager deduplicates the data that is
already stored. No additional backup, archive, or migration is required. You can
also define or update a storage pool for data deduplication, but not actually
deduplicate data.
v Decide how you want to control duplicate-identification processes.

About this task

You can create a storage pool for data deduplication or update an existing storage
pool for data deduplication. You can store client-side deduplicated data and
server-side deduplicated data in the same storage pool.

Procedure

To set up a storage pool for data deduplication, complete the following steps:
v If you are defining a new storage pool:
1. Use the DEFINE STGPOOL command and specify the DEDUPLICATE=YES
parameter.
2. Define a new policy domain to direct eligible client-node data to the storage
pool.
v If you are updating an existing storage pool:
1. Determine whether the storage pool contains data from one or more client
nodes that you want to exclude from data deduplication. If it does:
a. Using the MOVE DATA command, move the data that belongs to the
excluded nodes from the storage pool to be converted to another storage
pool.
b. Direct data that belongs to the excluded nodes to the other storage pool.
The easiest way to complete this task is to create another policy domain
and designate the other storage pool as the destination storage pool.
2. Change the storage-pool definition with the UPDATE STGPOOL command.
Specify the DEDUPLICATE and NUMPROCESSES parameters.

Results

As data is stored in the pool, the duplicates are identified. When the reclamation
threshold for the storage pool is reached, reclamation begins, and the space that is
occupied by duplicate data is reclaimed.

306 IBM Tivoli Storage Manager for Windows: Administrator's Guide

In the storage pool definition, you can specify as many as 50 duplicate-
identification processes to start automatically. However, the number of
duplicate-identification processes must not exceed the number of processor cores
available on the Tivoli Storage Manager server. If you do not specify any
duplicate-identification processes in the storage pool definition, you must control
data deduplication manually. Duplicate identification requires extra disk I/O and
processor resources. To mitigate the effects on server workload, you can manually
increase or decrease the number of duplicate-identification processes, along with
their duration.

Attention: By default, the Tivoli Storage Manager server requires that you back
up deduplication-enabled primary storage pools before volumes in the storage
pool are reclaimed and before duplicate data is discarded.

Creating a copy of primary storage pools with data deduplication

enabled
Tivoli Storage Manager provides two methods that you can use to create a copy of
data in your deduplicated storage pools.

About this task

You can create a copy of the data by using BACKUP STGPOOL or REPLICATE NODE
command. When you back up a primary storage pool, you create a copy of the
entire storage pool. When you replicate data by using node replication, you copy
data from one or more nodes from primary storage pools to a primary storage
pool on another Tivoli Storage Manager server.

For deduplicated storage pools, you must determine whether the

DEDUPREQUIRESBACKUP should be set to yes or no.

The value of the DEDUPREQUIRESBACKUP option specifies whether the following

actions will occur:
v Volumes in primary sequential-access storage pools that are set up for data
deduplication can be reclaimed
v Duplicate data can be discarded before the storage pools are backed up
The default value is Yes.

Important: If you do not want to create a copy of the storage pool data, you must
change the DEDUPREQUIRESBACKUP option to No. Failure to set the option correctly
can cause deduplication to not work correctly and cause your data to not be
reclaimed.

Table 26 describes the different scenarios that you can use to create a copy of data
in your deduplicated storage pools, and which value of DEDUPREQUIRESBACKUP to
use.
Table 26. Setting the value for the DEDUPREQUIRESBACKUP option
Creating a copy of your primary storage pool DEDUPREQUIRESBACKUP
data value Method
Back up your primary storage pool data to a Yes BACKUP STGPOOL
non-deduplicated copy pool, such as a copy
pool that uses tape.
Back up your primary storage pool data to a No BACKUP STGPOOL
deduplicated copy pool.

Chapter 10. Managing storage pools and volumes 307

 Table 26. Setting the value for the DEDUPREQUIRESBACKUP option (continued)
Creating a copy of your primary storage pool DEDUPREQUIRESBACKUP
data value Method
Use node replication to create a copy of your No REPLICATE NODE
data on another Tivoli Storage Manager server.
No copy is created. No

Depending on the method that you chose to create a copy of the data in the
primary storage pools, complete one of the following actions:

Procedure
v Use the storage pool backup command to back up data:
1. Issue the BACKUP STGPOOL command. If you set the DEDUPREQUIRESBACKUP
option to yes, you must back up data to a copy storage pool that is not set
up for data deduplication.

Tip: When you copy data to an active data pool, it does not provide the
same level of protection that occurs when you create a storage pool backup
or use node replication.
2. Issue the IDENTIFY DUPLICATES command to identify duplicate data.

Tip: If you backup storage pool data after duplicate data is identified, the
copy process can take longer because the data must be reconstructed to find
any duplicate data.
v Use the node replication command to back up data:
1. Issue the IDENTIFY DUPLICATES command to identify duplicate data.
2. Issue the REPLICATE NODE command to start node replication.

Effects on data deduplication when moving or copying data

You can move or copy data between storage pools regardless of whether they are
set up for data deduplication.

The following table illustrates what happens to data deduplication when data
objects are moved or copied.
Table 27. Effects when moving or copying data
...and you move or copy
If the source storage data to a target storage
pool is... pool that is... The result is...
Set up for data Set up for data All data objects in the source pool are
deduplication deduplication examined for existence in the target
pool. If an object exists in the target
pool, information about data
deduplication is preserved so that the
data does not need to be deduplicated
again. If an object does not exist in the
target pool, it is moved or copied.
Not set up for data The data is not deduplicated in the
deduplication target storage pool. This rule applies to
any type of storage pool, including
storage pools that use virtual volumes.

308 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Table 27. Effects when moving or copying data (continued)
...and you move or copy
If the source storage data to a target storage
pool is... pool that is... The result is...
Not set up for data Set up for data Normal data deduplication processing
deduplication deduplication takes place after the data is moved or
copied.
Not set up for data No data deduplication occurs.
deduplication

When you copy or move data from a deduplicated storage pool to a

non-deduplicated storage pool, the data is reconstructed. However, after the data
movement or copy operation, the amount of data that is reported as moved or
copied is the amount of deduplicated data. For example, suppose a storage pool
contains 20 GB of deduplicated data that represents 50 GB of total file data. If the
data is moved or copied, the server reports that 20 GB was moved or copied, even
though 50 GB of data was sent.

Turning data deduplication on or off for a storage pool

If you turn data deduplication off for a storage pool by updating the storage pool
definition, new data that enters the storage pool is not deduplicated.

About this task

Deduplicated data, which was in the storage pool before you turned off data
deduplication, is not reassembled. Deduplicated data continues to be removed due
to normal reclamation and deletion. All information about data deduplication for
the storage pool is retained.

Procedure

To turn off data deduplication for a storage pool, use the UPDATE STGPOOL
command and specify DEDUPLICATE=NO.

What to do next

If you turn data deduplication on for the same storage pool, duplicate-
identification processes resume, skipping any files that are previously processed.
You can change the number of duplicate-identification processes. When you are
calculating the number of duplicate-identification processes to specify, consider the
workload on the server and the amount of data that requires data deduplication.
The number of duplicate-identification processes must not exceed the number of
processor cores available on the IBM Tivoli Storage Manager server.

To enable data deduplication for a storage pool, specify DEDUPLICATE=YES on the

UPDATE STGPOOL command.

Chapter 10. Managing storage pools and volumes 309

 Controlling data deduplication
If client files are bound to a management class that specifies a
deduplication-enabled storage pool, the files are, by default, deduplicated on the
server. Client-side data deduplication is enabled by using a combination of settings
on the client and the server.

About this task

The following table shows how the data deduplication settings on the client
interact with the data deduplication settings on the Tivoli Storage Manager server.
Table 28. Data deduplication settings: Client and server
Value of the
DEDUPLICATION Value of the client
parameter for REGISTER NODE DEDUPLICATION option
or UPDATE NODE in the client options file Data deduplication location
SERVERONLY Yes Server

The Yes value of the client

option is ignored by the
server.
CLIENTORSERVER Yes Client
CLIENTORSERVER No Server
SERVERONLY No Server

You can set the DEDUPLICATION option in the client options file, in the preference
editor of the Tivoli Storage Manager client GUI, or in the client option set on the
Tivoli Storage Manager server. Use the DEFINE CLIENTOPT command to set the
DEDUPLICATION option in a client option set. To prevent the client from overriding
the value in the client option set, specify FORCE=YES.

Controlling server-side data deduplication

If client files are bound to a management class that specifies a
deduplication-enabled storage pool, the files are, by default, deduplicated on the
server.

Procedure

To enable server-side data deduplication, specify SERVERONLY as the value of the

DEDUPLICATION parameter on the REGISTER NODE or UPDATE NODE command.
If you specify SERVERONLY, the values of the client DEDUPLICATION option are
ignored.
Related concepts:
“Server-side data deduplication” on page 286

310 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Controlling duplicate-identification processes:

In server-side data deduplication, client data is deduplicated on the server. When

you define or update a storage pool for data deduplication, you can specify 0 - 20
duplicate-identification processes to start automatically and run indefinitely. To
avoid resource impacts during server operations (for example, client backups), you
can also control data deduplication processing manually.

About this task

For example, suppose you specify eight duplicate-identification processes in your

storage pool definition. These processes start automatically and run indefinitely.
However, you decide that you want to reduce the number of processes during
client backups, which take 60 minutes. You can manually reduce the number of
duplicate-identification processes to four and set a duration of 60 minutes. After
the backup is complete, the IBM Tivoli Storage Manager server automatically
restarts four processes so that the eight processes are running again.

Alternatively, you can identify duplicates manually. Specify 0 as the number of

duplicate-identification processes in your storage pool definition. With this setting,
the Tivoli Storage Manager server does not automatically start any
duplicate-identification processes. Then, depending on your schedule and server
workload, specify the number of duplicate-identification processes and their
duration for one or more storage pools.

Remember: The number of duplicate-identification processes should not exceed

the number of processor cores available on the Tivoli Storage Manager server.

Duplicate-identification processing states:

Duplicate-identification processes are different from other server processes. When

other server processes finish a task, they end. When duplicate-identification
processes finish processing available files, they go into an idle state.

Duplicate-identification processes can be either active or idle. Processes that are

currently working on files are active. Processes that are waiting for files to work on
are idle. Processes remain idle until volumes with data to be deduplicated become
available. Processes end only when canceled or when you change the number of
duplicate-identification processes for the storage pool to a value less than the
number that is currently specified.

The output of the QUERY PROCESS command for a duplicate-identification process

includes the total number of bytes and files that have been processed since the
process first started. For example, if a duplicate-identification process processes
four files, idles, and then processes five more files, the total number of files
processed is nine.

Chapter 10. Managing storage pools and volumes 311

 Interaction of manual data deduplication controls:

You can change the number of duplicate-identification processes used during

server-side data deduplication. You can also change the length of time that
processes are allowed to run. You can adjust these settings as often as you want.

Table 29 shows how these two controls, the number and duration of processes,
interact for a particular storage pool.

Remember:
v When the amount of time that you specify as a duration expires, the number of
duplicate-identification processes always reverts to the number of processes
specified in the storage pool definition.
v When the server stops a duplicate-identification process, the process completes
the current physical file and then stops. As a result, it might take several
minutes to reach the value that you specify as a duration.
v To change the number of duplicate-identification processes, you can also update
the storage pool definition using the UPDATE STGPOOL command. However, when
you update a storage pool definition, you cannot specify a duration. The
processes that you specify in the storage pool definition run indefinitely, or until
you issue the IDENTIFY DUPLICATES command, update the storage pool definition
again, or cancel a process.

In this example, you specified three duplicate-identification processes in the

storage pool definition. You use the IDENTIFY DUPLICATES command to change the
number of processes and to specify the amount of time the change is to remain in
effect.
Table 29. Controlling duplicate-identification processes manually
Using the IDENTIFY DUPLICATES
command, you specify... ...and a duration of... The result is...
2 duplicate-identification None specified One duplicate-identification processes finishes the file
processes it is working on, if any, and then stops. Two processes
run indefinitely, or until you reissue the IDENTIFY
DUPLICATES command, update the storage pool
definition, or cancel a process.
60 minutes One duplicate-identification process finishes the file it
is working on, if any, and then stops. After 60
minutes, the server starts one process so that three are
running.
4 duplicate-identification None specified The server starts one duplicate-identification process.
processes Four processes run indefinitely, or until you reissue
the IDENTIFY DUPLICATES command, update the
storage pool definition, or cancel a process.
60 minutes The server starts one duplicate-identification process.
At the end of 60 minutes, one process finishes the file
it is working on, if any, and then stops. The
additional process started by this command might not
be the one that stops when the duration has expired.

312 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Table 29. Controlling duplicate-identification processes manually (continued)
Using the IDENTIFY DUPLICATES
command, you specify... ...and a duration of... The result is...
0 duplicate-identification None specified All duplicate-identification processes finish the files
processes that they are working on, if any, and stop. This
change lasts indefinitely, or until you reissue the
IDENTIFY DUPLICATES command, update the storage
pool definition, or cancel a process.
60 minutes All duplicate-identification processes finish the files
that they are working on, if any, and stop. At the end
of 60 minutes, the server starts three processes.
None specified Not available The number of duplicate-identification processes
resets to the number of processes specified in the
storage pool definition. This change lasts indefinitely,
or until you reissue the IDENTIFY DUPLICATES
command, update the storage pool definition, or
cancel a process.

The following example illustrates how you can control data deduplication using a
combination of automatic and manual duplicate-identification processes. Suppose
you create two new storage pools for data deduplication, A and B. When you
create the pools, you specify two duplicate-identification processes for A and one
process for B. The IBM Tivoli Storage Manager server is set by default to run those
processes automatically. As data is stored in the pools, duplicates are identified and
marked for removal. When there is no data to deduplicate, the
duplicate-identification processes go into an idle state, but remain active.

Suppose you want to avoid resource impacts on the server during client-node
backups. You must reduce the number of duplicate-identification processes
manually. For A, you specify a value of 1 for the number of duplicate-identification
process. For B, you specify a value of 0. You also specify that these changes remain
in effect for 60 minutes, the duration of your backup window.

Specifying these values causes two of the three running processes to finish the files
on which they are working and to stop. One duplicate-identification process is
now running for A. No duplicate-identification processes are running for B. After
60 minutes, the Tivoli Storage Manager server automatically resets the
data-deduplication processes to the values specified in the storage pool definition.
One process starts for A, for a total of two running processes. One process also
starts for B.

Starting and stopping duplicate-identification processes:

In server-side data deduplication, you can start additional duplicate-identification

processes and stop some or all active processes. You can also specify an amount of
time that the change remains in effect. If you did not specify any
duplicate-identification processes in the storage pool definition, you can start new
processes and stop them manually.

Procedure

To specify the number and duration of duplicate-identification processes for a

storage pool, issue the IDENTIFY DUPLICATES command.

Chapter 10. Managing storage pools and volumes 313

 Example

For example, suppose that you have four storage pools: stgpoolA, stgpoolB,
stgpoolC, and stgpoolD. All the storage pools are associated with a particular IBM
Tivoli Storage Manager server. Storage pools A and B are each running one
duplicate-identification process, and storage pools C and D are each running two.
A 60-minute client backup is scheduled to take place, and you want to reduce the
server workload from these processes by two-thirds.

To accomplish this task, issue the following commands:

IDENTIFY DUPLICATES STGPOOLA DURATION=60 NUMPROCESS=0
IDENTIFY DUPLICATES STGPOOLB DURATION=60 NUMPROCESS=0
IDENTIFY DUPLICATES STGPOOLC DURATION=60 NUMPROCESS=1
IDENTIFY DUPLICATES STGPOOLD DURATION=60 NUMPROCESS=1

Now two processes are running for 60 minutes, one third of the number running
before the change. At the end of 60 minutes, the Tivoli Storage Manager server
automatically restarts one duplicate-identification process in storage pools A and B,
and one process in storage pools C and D.

Enabling deduplication of NetApp file-server data:

To deduplicate data that is stored by a network-attached storage (NAS) file server,

use the ENABLENASDEDUP server option. This option applies only to NetApp file
servers. The default is NO. Deduplication for the storage pool must be enabled. If
you are using NDMP backup of NetApp file servers in your environment, use the
YES option.

Controlling client-side data deduplication

The combination of client option and server parameter values determines whether
data deduplication occurs on the client or the server. The default data
deduplication setting is server-side data deduplication.

Procedure

To enable client-side data deduplication, complete both of the following steps:

v Specify YES as the value of the DEDUPLICATION option for the client.
You can set the DEDUPLICATION option in the client options file, in the preference
editor of the IBM Tivoli Storage Manager client GUI, or in the client option set
on the Tivoli Storage Manager server. Use the DEFINE CLIENTOPT command to set
the DEDUPLICATION option in a client option set. To prevent the client from
overriding the value in the client option set, specify FORCE=YES.
v Specify DEDUPLICATION=CLIENTORSERVER on the REGISTER NODE or UPDATE NODE
command.

Results

For details about client-side data deduplication options, see the Backup-Archive
Clients Installation and User's Guide.
Related concepts:
“Client-side data deduplication” on page 286

314 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Enabling client-side data deduplication for a single node:

In this example, you enable client-side data deduplication for a single node. You
have a policy domain that you use to manage deduplicated data.

About this task

The name of the domain that you use to manage deduplicated data is
dedupdomain1. The primary storage pool that is specified by the copy group of
the default management class is a deduplication-enabled storage pool. The client,
MATT, that you want to enable for data deduplication uses a default management
class for backup operations.

Procedure

To enable client-side data deduplication for a single node, complete the following
steps:
v On the server, assign client node MATT to dedupdomain1:
update node matt domain=dedupdomain1 deduplication=clientorserver
The setting of the DEDUPLICATION parameter must be CLIENTORSERVER
v Add the following option to the [Link] file:
deduplication yes
You can also set the DEDUPLICATION option in the preference editor of the Tivoli
Storage Manager client GUI or in the client option set on the Tivoli Storage
Manager server. Use the DEFINE CLIENTOPT command to set the DEDUPLICATION
option in a client option set. To prevent the client from overriding the value in
the client option set, specify FORCE=YES.

What to do next

To determine the amount of data that was deduplicated, start a backup or archive
operation. At the end of the operation, check the backup or archive report.

Enabling client-side data deduplication for multiple client nodes:

In this example, you enable client-side data deduplication for more than one client
node.

Procedure

Complete the following steps on the server. No configuration is necessary on the

client.
1. Assign all the client nodes to a domain (DEDUPDOMAIN1) that has a
deduplication-enabled destination storage pool:
update node node1 domain=dedupdomain1 deduplication=clientorserver
...
update node noden domain=dedupdomain1 deduplication=clientorserver

You can automate this step with a script or a macro.

2. To enable client-side data deduplication, define a client option set:
define cloptset client_dedup
define clientopt client_dedup deduplication yes force=yes

Chapter 10. Managing storage pools and volumes 315

 You can also add the following client-side data deduplication options to the
client option set:
v ENABLEDEDUPCACHE
v DEDUPCACHEPATH
v DEDUPCACHESIZE
v [Link]
v [Link]
3. Assign the client option set to the client nodes:
update node node1 cloptset=client_dedup
...
update node noden cloptset=client_dedup

You can automate this step with a script or a macro.

Changing data deduplication location for a single client:

In this scenario, a client is deduplicating data. However, because of security

concerns, you want the IBM Tivoli Storage Manager server to deduplicate the data
belonging to the client.

About this task

The data belonging client MATT is bound to a management class with a copy
group that specifies a deduplication-enabled destination storage pool.

Procedure

To change the data deduplication location from the client to the server, issue the
following command:
update node matt deduplication=serveronly

No configuration is necessary on the client.

Specifying transaction and object size for deduplication

When you deduplicate large objects, intensive database activity can result from
long-running transactions that are required to update the database. Server options
are available to limit the size of transactions when client-side deduplicated data is
backed up or archived and to limit the size of objects that are deduplicated by the
server.

About this task

High levels of database activity can produce the following symptoms:

v Reduced throughput for client backup and archive operations
v Resource contention that results from concurrent server operations
v Excessive recovery log activity

The extent to which these symptoms occur depends on the number and size of the
objects that are being processed, the intensity, and the type of concurrent
operations that are taking place on the IBM Tivoli Storage Manager server, and the
Tivoli Storage Manager server configuration.

316 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Procedure

With the SERVERDEDUPTXNLIMIT server option, you can limit the size of objects that
can be deduplicated on the server. With the CLIENTDEDUPTXNLIMIT server option,
you can limit the size of transactions when client-side deduplicated data is backed
up or archived.

Results

Tip: To control which objects are deduplicated, you can also use the MAXSIZE
parameter of the DEFINE STGPOOL and UPDATE STGPOOL commands. Using the
MAXSIZE parameter, you can force large objects to the NEXT storage pool for
storage.

What to do next

For details about these options, see the Administrator's Reference.

Tiered data deduplication

With tiered data deduplication, you can use larger average extent sizes as objects
reach a greater size.

Data deduplication uses an average extent size of 256 KB. When deduplicating
large objects, for example, over 200 GB, the number of extents for an object can
grow large. Assuming extents are 256 KB, there are 819,200 extents for a 200 GB
object. When you need to restore this object, all 819,200 database records must be
read before the object is accessible.

Tiered data deduplication can manage larger objects because a larger average
extent size is used when deduplicating the data. For example, after an object
reaches 200 GB, the Tivoli Storage Manager server uses 1 MB as the average extent
size, instead of 256 KB. 819,200 extents become 204,800 extents.

Using the server options, DEDUPTIER2FILESIZE and DEDUPTIER3FILESIZE, you can

configure three different tier sizes for data deduplication:
Tier 1 All objects greater than 0 and less than the value set for the
DEDUPTIER2FILESIZE option. Tier 1 uses an average extent size of 256 KB.
Tier 2 All objects greater than or equal to the value of DEDUPTIER2FILESIZE and
less than the value for DEDUPTIER3FILESIZE. Tier 2 uses an average extent
size of 1 MB.
Tier 3 All objects greater than or equal to the value for DEDUPTIER3FILESIZE. Tier
3 uses an average extent size of 2 MB.

Note: By default, objects under 100 GB in size are processed at Tier 1. Objects in
the range of 100 GB to under 400 GB are processed in Tier 2. All objects 400 GB
and larger are processed in Tier 3.

Chapter 10. Managing storage pools and volumes 317

 Setting options for tiered deduplication:

Depending on your environment, you can set different options for using tiered
data deduplication. However, if possible, avoid changing the default tier settings.
Small changes might be tolerated, but frequent changes to these settings can
prevent matches between previously stored backups and future backups.

Example

If you want to use two tiers for data deduplication instead of three, you can set the
DEDUPTIER2FILESIZE and DEDUPTIER3FILESIZE accordingly.
Use Tier 1 and Tier 2 only
To have two tiers with an average extent size of 256 KB and 1 MB, specify
these values:
DEDUPTIER2FILESIZE 100
DEDUPTIER3FILESIZE 9999
Use Tier 1 and Tier 3 only
To have two tiers with an average extent size of 256 KB and 2 MB, specify
these values:
DEDUPTIER2FILESIZE 100
DEDUPTIER3FILESIZE 100

If you do not want to use tiered data deduplication and instead preserve your
existing environment, set the value for both of the tiered data deduplication
options to 9999. For example:
DEDUPTIER2FILESIZE 9999
DEDUPTIER3FILESIZE 9999

If both options are set to 9999, then all files that are 10 TB or less are processed
with the default extent size of 256 KB.

Displaying statistics about server-side data deduplication

Important statistics about data deduplication are available by querying the server
for information about storage pools or duplicate-identification processes.

About this task

You can also obtain statistics about client-side data deduplication. For details, see
Backup-Archive Clients Installation and User's Guide.

Querying a storage pool for statistics about data deduplication

You can query a storage pool for important statistics about data deduplication.

About this task

Querying a storage pool provides the following statistics:

v Whether the storage pool has been set up for data deduplication
v The number of duplicate-identification processes specified when the storage pool
was created
v The amount of data that was removed from the storage pool by reclamation
processing

318 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Procedure

To query a storage pool for statistics about data deduplication, issue the QUERY
STGPOOL command.

What to do next

You might notice a discrepancy between the number of duplicate-identification

processes specified as the default for a storage pool and the number of
duplicate-identification processes currently running. This discrepancy occurs when
you manually increase or decrease the number of duplicate-identification processes
for the storage pool.

If you run a query before reclamation of the storage pool, the Duplicate Data Not
Stored value in the command output is inaccurate and does not reflect the most
recent data reduction.

Remember: Querying a storage pool displays storage-pool utilization as a

percentage of its assigned capacity. (Storage-pool utilization is expressed as Pct Util
in the command output.) This field does not display a value for storage pools that
are set up for data deduplication. If you turn off data deduplication for a storage
pool, a value for percentage utilized is not displayed until all duplicate data is
removed from the storage pool.

Querying a duplicate-identification process

Querying a duplicate-identification process displays the total number of bytes and
total number of files processed.

Procedure

To query a duplicate-identification process, issue the QUERY PROCESS command.

Querying a volume for information about linked files

You can query a volume for information about client files that link to files on other
volumes. This information is useful when file extents created by data deduplication
are distributed on different volumes.

About this task

You can display information only about files that are linked to a volume or only
about files that are stored on a volume. You can also display information about
both stored files and linked files.

Procedure

To display information about files on a volume, issue the QUERY CONTENT command
and specify the FOLLOWLINKS parameter.
For example, suppose a volume in a deduplicated storage pool is physically
destroyed. You must restore this volume. Before you do, you want to determine
whether other volumes in the storage pool have files that are linked to files in the
destroyed volume. With that information, you can decide whether to restore the
other volumes. To identify links, you issue the QUERY CONTENT command for the
destroyed volume and specify the FOLLOWLINKS parameter to list all the files with
links to files on the destroyed volume.

Chapter 10. Managing storage pools and volumes 319

 Displaying statistics about client-side data deduplication
Use client statistics to compare the amount of data that is transferred during a
backup with the amount of data that would be transferred if no deduplication took
place. Client statistics compare the amount of transferred data that was unique
with the amount of data that was inspected by the client.

About this task

You can use the activity log to view client statistics about data deduplication. The
activity log can show historical information about one or more nodes. You can also
view data reduction information for data deduplication by using the Tivoli Storage
Manager API.

Procedure

To view client statistics for data deduplication, see the activity log, or use the
Tivoli Storage Manager API.

Example

The following client statistics are taken from the activity log:
tsm> incremental c:\test\* -sub=yes
Incremental backup of volume ’c:\test\*’
Normal File--> 43,387,224 \\naxos\c$\test\newfile [Sent]
Successful incremental backup of ’\\naxos\c$\test\*’

Total number of objects inspected: 7

Total number of objects backed up: 1
Total number of objects updated: 0
Total number of objects rebound: 0
Total number of objects deleted: 0
Total number of objects expired: 0
Total number of objects failed: 0
Total objects deduplicated: 1
Total number of bytes inspected: 143.29 MB <= sum of all bytes inspected
Total number of bytes processed: 21.07 MB <= bytes after client-side
deduplication
Total number of bytes transferred: 21.07 MB <= bytes after compression
Data transfer time: 1.76 sec
Network data transfer rate: 12,217.36 KB/sec
Aggregate data transfer rate: 4,223.10 KB/sec
Objects compressed by: 0%
Deduplication reduction: 49.10% <== 1 - (21.07 / 41.37)
Total data reduction ratio: 85.30% <== 1 - (21.07 / 143.29)
Elapsed processing time: [Link]
ANS1900I Return code is 0.
ANS1901I Highest return code was 0.

The \\naxos\c$\test directory uses approximately 143.29 MB of space. All files are
already stored on the Tivoli Storage Manager server except the c:\test\newfile
file, which is 41.37 MB (43,387,224 bytes). After client-side data deduplication, it is
determined that only approximately 21 MB will be sent to the server.

The following client statistics are produced using the Tivoli Storage Manager API:
typedef struct tsmEndSendObjExOut_t
{
dsUint16_t stVersion; /* structure version */
dsStruct64_t totalBytesSent; /* total bytes read from app */
dsmBool_t objCompressed; /* was object compressed */

320 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 dsStruct64_t totalCompressSize; /* total size after compress */
dsStruct64_t totalLFBytesSent; /* total bytes sent LAN free */
dsUint8_t encryptionType; /* type of encryption used */
dsmBool_t objDeduplicated; /* was object processed for dist. data dedup */
dsStruct64_t totalDedupSize; /* total size after de-dup */
} tsmEndSendObjExOut_t;

The API returns the totalDedupSize member of the tsmEndSendObjExOut_t

structure. The calling application can calculate the client-side data deduplication
reduction. It is calculated by comparing the total number of bytes that are sent to
the server with the total deduplication size structure members. If data
deduplication took place but no duplicate extents were found, the following
calculation is used:
totalBytesSent == totalDedupSize

To determine whether client-side data deduplication processing took place,

examine the Boolean value of the objDeduplicated member of the
tsmEndSendObjExOut_t structure.

Querying the server activity log for client-side data deduplication

statistics
You can use the Tivoli Storage Manager client reports for data deduplication
statistics after each backup or archive operation. To find out the cumulative data
reduction for a node or the reduction across multiple nodes, you can query the
server activity log.

About this task

After each backup or archive operation, the Tivoli Storage Manager client reports
the data deduplication statistics in the server activity log. For details about the
activity log, see the Tivoli Storage Manager Information Center, and search for
activity log.

Procedure

To query the data deduplication statistics for the client, issue the QUERY ACTLOG
command.

Example

See the following example for sample information provided by the QUERY ACTLOG
command:
tsm: HALLEY>q actlog nodename=’modo’

Date/Time Message
-------------------- ----------------------------------------------------------
03/15/10 [Link] ANE4952I (Session: 406, Node: MODO)
Total number of objects inspected: 1 (SESSION: 406)
03/15/10 [Link] ANE4954I (Session: 406, Node: MODO)
Total number of objects backed up: 1 (SESSION: 406)
03/15/10 [Link] ANE4958I (Session: 406, Node: MODO)
Total number of objects updated: 0 (SESSION: 406)
03/15/10 [Link] ANE4960I (Session: 406, Node: MODO)
Total number of objects rebound: 0 (SESSION: 406)
03/15/10 [Link] ANE4957I (Session: 406, Node: MODO)
Total number of objects deleted: 0 (SESSION: 406)
03/15/10 [Link] ANE4970I (Session: 406, Node: MODO)
Total number of objects expired: 0 (SESSION: 406)
03/15/10 [Link] ANE4959I (Session: 406, Node: MODO)

Chapter 10. Managing storage pools and volumes 321

 Total number of objects failed: 0 (SESSION: 406)
03/15/10 [Link] ANE4982I (Session: 406, Node: MODO)
Total objects deduplicated: 1(SESSION: 406)
03/15/10 [Link] ANE4977I (Session: 406, Node: MODO)
Total number of bytes inspected: 7.05 MB(SESSION: 406)
03/15/10 [Link] ANE4975I (Session: 406, Node: MODO)
Total number of bytes processed: 33 B(SESSION: 406)
03/15/10 [Link] ANE4961I (Session: 406, Node: MODO)
Total number of bytes transferred: 33 B (SESSION: 406)
03/15/10 [Link] ANE4963I (Session: 406, Node: MODO)
Data transfer time: 0.00 sec (SESSION: 406)
03/15/10 [Link] ANE4966I (Session: 406, Node: MODO)
Network data transfer rate: 77.09 KB/sec (SESSION: 406)
03/15/10 [Link] ANE4967I (Session: 406, Node: MODO)
Aggregate data transfer rate: 0.01 KB/sec (SESSION: 406)
03/15/10 [Link] ANE4968I (Session: 406, Node: MODO)
Objects compressed by: 0% (SESSION: 406)
03/15/10 [Link] ANE4981I (Session: 406, Node: MODO)
Deduplication reduction: 100.00%(SESSION: 406)
03/15/10 [Link] ANE4976I (Session: 406, Node: MODO)
Total data reduction ratio: 100.00%(SESSION: 406)
03/15/10 [Link] ANE4964I (Session: 406, Node: MODO)
Elapsed processing time: [Link] (SESSION: 406)

The following example shows how to use the activity log to gather the data
reduction information across all nodes that belong to the DEDUP domain:
dsmadmc -id=admin -password=admin -displaymode=list -scrollprompt=no "select
DISTINCT [Link], [Link] from ACTLOG A1, ACTLOG A2 where [Link]
in (select NODE_NAME from nodes where domain_name=’DEDUP’) and
[Link]=[Link] and [Link]=4977 and [Link]=4961 and EXISTS
(select [Link] from ACTLOG A3 where [Link]=[Link] and [Link]=4982)"
| grep ’MESSAGE:’ | sed -r ’s/MESSAGE:.*:\s+([0-9]+(\.[0-9]+)?)\s+
(B|KB|MB|GB|TB).*(SESSION: .*)/\1 \3/’ | sed -r ’s/\.//’ | awk -f [Link]

The [Link] file contains the following statements:

BEGIN { bytesInspected = 0; bytesTransferred = 0; }

{ if ($2=="B") valueInKB = 0;
if ($2=="KB") valueInKB = $1;
if ($2=="MB") valueInKB = $1 * 1024;
if ($2=="GB") valueInKB = $1 * 1024 * 1024;
if ($2=="TB") valueInKB = $1 * 1024 * 1024 *1024;

if (NR % 2 == 1) bytesInspected += valueInKB;

if (NR % 2 == 0) bytesTransferred += valueInKB;
}

END {printf("Number of bytes inspected:

%d KB\n Number of bytes transferred: %d
KB\nData reduction ratio: %d \%\n", bytesInspected,
bytesTransferred, (1-bytesTransferred/bytesInspected)*100)}

The QUERY ACTLOG command gives a summary, as shown in the following example:
Number of bytes inspected: 930808832 KB
Number of bytes transferred: 640679936 KB
Data reduction ratio: 31 %

322 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Querying about data deduplication in file spaces
You can view the amount of physical space that is occupied by a file space after
removing the deduplication savings that are gained by removing duplicated data
from the file space.

Procedure

To query where client file spaces are stored and how much space they occupy,
issue the QUERY OCCUPANCY command.

Example

In the following example, 10 MB of data is placed in the FS1 file space, and 2 MB
is marked for expiration and is removed during the next expiration process.
Therefore, Physical Space Occupied reports 10 MB and Logical Space Occupied
reports 8 MB. The Physical Space Occupied value for storage pools that use data
deduplication is not shown.
tsm: SERVER1>q occupancy dedup*

Node Name Type Filespace FSID Storage Number of Physical Logical

Name Pool Name Files Space Space
Occupied Occupied
(MB) (MB)
---------- ---- ---------- ----- ---------- --------- --------- ---------
DEDUP_N Bkup \\abcde\c$ 17 NODEDUPOOL 63 10.00 8.00
DEDUP_Y Bkup \\abcde\d$ 18 YADEDUPOOL 63 - 8.00

The occupancy table shows how much physical space is occupied by a file space
after the removal of the deduplication savings. These savings are gained by
removing duplicated data from the file space. You can use select * from
occupancy to get LOGICAL_MB and REPORTING_MB values.

LOGICAL_MB is the amount of space that is used by this file space. REPORTING_MB is
the amount of space that is occupied when the data is not placed in a
deduplication-enabled storage pool.

For example, the \\brian\c$ file space manages 30 MB of space, but 10 MB is

eliminated by either client-side or server-side data deduplication. The resulting
occupancy tables shows:
tsm: SERVER1>select * from occupancy where NODE_NAME=’BRIAN’

NODE_NAME: BRIAN
TYPE: Bkup
FILESPACE_NAME: \\brain\c$
STGPOOL_NAME: MYFILEPOOL
NUM_FILES: 63
PHYSICAL_MB: 0.00
LOGICAL_MB: 10.00
REPORTING_MB: 30.00
FILESPACE_ID: 17

Tip: The LOGICAL_MB value takes into account only the amount of data that is
removed or not stored because the data is identified as a duplicate of data that is
stored elsewhere.

For example, the \\brian\c$ file space backs up 20 MB of unique data, which
means that LOGICAL_MB is 20 MB. The \\jim\c$ file space must not back up any
data because it is an exact duplicate of the \\brian\c$ file space, which means that

Chapter 10. Managing storage pools and volumes 323

 LOGICAL_MB is 0 MB. In this case, the \\brian\c$ file space does not report any
duplication efficiencies, even though the \\jim\c$ file space is able to identify
duplicate data between the file spaces.

Scenarios for data deduplication

You can benefit from client-side data deduplication when it is used in combination
with other IBM Tivoli Storage Manager functions and products. You can reduce
storage requirements when you use client-side data deduplication with image
backups, system state backups, and file-level backups. When you use caching with
data deduplication, you can reduce network traffic.

For example, IBM Tivoli Storage Manager for Mail and IBM Tivoli Storage
Manager for Databases can use client-side data deduplication through the Tivoli
Storage Manager API to create backup sets and export node data.

The progressive incremental backup technology in Tivoli Storage Manager provides

an effective method of efficiently reducing the amount of data that is processed in
each backup. This technology can also be effectively combined with data
deduplication. When used in combination, data is initially reduced by the
incremental processing which is able to skip unchanged objects without applying
deduplication processing against them. For those objects which do require a
backup, data deduplication is applied.

Another Tivoli Storage Manager backup model is the use of incremental-forever

backups. With this technology, each subsequent backup achieves significant storage
savings by not having to send unchanged objects. These unchanged objects then do
not need to go through deduplication processing, which is an efficient method of
reducing data.

Reducing the storage requirement of image backups

With client-side data deduplication, you can further reduce the storage requirement
of image backups. Tivoli Storage Manager can perform a snapshot-based image
backup during which the volume is available to other system applications.

Image backup can be full or incremental. In a typical scenario, full image backups
are scheduled less frequently than incremental image backups. For example, a full
image backup is scheduled weekly and incremental backups are scheduled daily,
except for the day of the selective image backup. The frequency of full image
backups is often driven by the available storage space. For example, each image
backup of a 50 GB volume might need 50 GB of space in a storage pool.

One approach to reduce the required storage space is to use compression. To

further reduce the storage footprint of image backups, you can use client-side data
deduplication. With data deduplication, you can perform more frequent full image
backups without significant use of storage space. The reduced need for storage
space is because only unique data is sent to the server during an image backup
operation.

324 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Reducing the storage requirement of system state backups
With client-side data deduplication, you can reduce the amount of duplicate data
that is stored on the server during system state backups. System state consists of all
system state and system services components that can be started. Tivoli Storage
Manager uses Microsoft Volume Shadow Copy Service (VSS) to back up all system
state components as a single object, which provides a consistent point-in-time
snapshot of the system state.

You can use VSS on Windows Server 2003, Windows Server 2008, and Windows
Vista operating systems. For details about backing up the Windows system state,
see Tivoli Storage Manager: Client Installation and User Guide.

System state can contain thousands of objects and take a large amount of storage
space on the server. It is likely that system state objects do not change much
between backups. This results in a large amount of duplicate data being stored on
the server. In addition, similar systems are likely to have a similar system state.
Therefore, when you perform system state backups on these systems, there is an
increase in duplicate data.

System state backup, when used in combination with client-side data

deduplication, reduces the amount of duplicate data that is stored on the server.
The reduction amount can vary, depending on the makeup of your data.

In the following example, a backup of the system state was performed on two
similar systems that run Windows Server 2008. There was no data backed up to
the storage pool. On the first system, the system-state data was deduplicated by
45%, as shown in Figure 22. A backup of the system state yielded a deduplication
reduction of 98% on the second system, as shown in Figure 23 on page 326.

This example shows a sample deduplication reduction of 45% for the system state
data:

Total number of objects inspected: 63,800

Total number of objects backed up: 63,800
Total number of objects updated: 0
Total number of objects rebound: 0
Total number of objects deleted: 0
Total number of objects expired: 0
Total number of objects failed: 0
Total objects deduplicated: 43,302
Total number of bytes inspected: 9.77 GB
Total number of bytes processed: 5.32 GB
Total number of bytes transferred: 5.32 GB
Data transfer time: 70.30 sec
Network data transfer rate: 79,431.06 KB/sec
Aggregate data transfer rate: 482.50 KB/sec
Objects compressed by: 0%
Deduplication reduction: 45.75%
Total data reduction ratio: 45.51%
Elapsed processing time: [Link]

Figure 22. Final statistics of the system state backup on system 1

This example shows a sample deduplication reduction of 98% for the system state
data:

Chapter 10. Managing storage pools and volumes 325

 Total number of objects inspected: 55,007
Total number of objects backed up: 55,007
Total number of objects updated: 0
Total number of objects rebound: 0
Total number of objects deleted: 0
Total number of objects expired: 0
Total number of objects failed: 0
Total objects deduplicated: 34,606
Total number of bytes inspected: 6.19 GB
Total number of bytes processed: 152.25 MB
Total number of bytes transferred: 152.25 MB
Data transfer time: 3.68 sec
Network data transfer rate: 42,354.40 KB/sec
Aggregate data transfer rate: 63.47 KB/sec
Objects compressed by: 0%
Deduplication reduction: 98.04%
Total data reduction ratio: 97.60%
Elapsed processing time: [Link]

Figure 23. Final statistics of the system state backup on system 2

Reducing the storage requirement of Tivoli Storage Manager for

Virtual Environments backups
A Tivoli Storage Manager for Virtual Environments backup can take a large
amount of storage space on the server. Using client-side data deduplication, you
can reduce the amount of duplicate data that is stored during a backup operation.

Tivoli Storage Manager for Virtual Environments backups consist of all virtual
machines in the environment. Often, large portions of individual backups are
common with other backups. Therefore, when you perform backup operations,
there is an increase in duplicate data.

When you use client-side data deduplication in combination with backups for
Tivoli Storage Manager for Virtual Environments, you can reduce the amount of
duplicate data that is stored on the server. The reduction amount varies, depending
on the makeup of your data.

Reducing the storage requirement of Tivoli Storage Manager for

Mail and Tivoli Storage Manager for Databases
You can use client-side data deduplication with Tivoli Storage Manager for Mail
and Tivoli Storage Manager for Databases through the Tivoli Storage Manager API.
Using client-side deduplication with these products reduces space requirements.

Before you use data deduplication, ensure that your system meets all prerequisites.

You can turn on client-side data deduplication by adding DEDUPLICATION YES to the
[Link] file.
Related concepts:
“Client-side data deduplication” on page 286

326 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Verifying data deduplication for Tivoli Storage Manager for Mail and Tivoli
Storage Manager for Databases:

In Tivoli Storage Manager V6.1 or earlier, data protection clients do not provide
data deduplication reduction statistics in the graphical user interface. In this
situation, you can verify that data deduplication occurs.

Procedure

To verify data deduplication in V6.1 or earlier, complete the following steps:

1. Add the following statements to the [Link] file:
TRACEFILE [Link]
TRACEFLAGS dedup api
2. Locate a trace entry for the tsmEndSendObjEx API call. The entry looks like the
following example:
03/05/2010 [Link].286 : [Link] (2052): tsmEndSendObjEx: Total bytes
send 0 2105534, encryptType is NO encryptAlg is NONE compress is 0, dedup
is 1, totalCompress is 0 0 totalLFBytesSent 0 0 totalDedupSize 0 15001
In the example, the dedup parameter of the TRACEFLAGS statement indicates
whether the objects were processed for client-side data deduplication. The
totalDedupSize parameter of the TRACEFLAGS statement reports the number of
bytes that were sent to the server after deduplication.
With image and system-state backups, you can have more frequent full
backups. Similarly, with client-side data deduplication you can have more
frequent full backups of the application data, for example, with backups of
databases and storage groups.

Reducing the storage requirement of file-level backups

The Tivoli Storage Manager backup-archive client provides several data reduction
methods. These methods include compression, incremental backup processing,
subfile backup, and data deduplication. With client-side data deduplication, you
can reduce the amount of data that is sent to the server during file-level backup.

When only the metadata of the file is changed, for example, with access control
lists or extended attributes, typically the whole file is backed up again. With
client-side data deduplication, although the file is backed up again, only the
metadata is sent to the server.

Reducing network traffic by using data deduplication cache

You can avoid queries to the Tivoli Storage Manager server during data
deduplication when you cache hash sums that are already sent to the server. It is
beneficial with client-side data deduplication to query the server, for example,
when extents are identified and hash sums calculated. Also, data deduplication
cache can be used to your advantage in some situations when multiple processes
run concurrently. When you use data deduplication cache, consider factors such as
insufficient file space and synchronization with the server that might impact on
data deduplication.

Client-side data deduplication identifies extents in the data stream and calculates
the associated hash sums. Data deduplication determines whether a data extent
with the same hash sum is already stored on the server. If it is already stored, the
backup-archive client only needs to notify the server about the hash sum, and can
avoid sending the corresponding data extent. This process reduces the amount of
data that is exchanged between the Tivoli Storage Manager backup-archive client
and the server.

Chapter 10. Managing storage pools and volumes 327

 You can avoid having to query the server when the hash sums that are already
sent to the server are cached. When additional data extents are identified with
identical hash sums, the server can be notified of the hash sum, saving additional
network traffic. The time spent on the query and response is also reduced. In most
cases, it is beneficial to use client deduplication cache for these hash sums.

Multiple processes and API applications

When multiple processes, such as concurrent backups or Tivoli Storage Manager

API applications, transfer content concurrently, they cannot use the cache. In such
a situation, only a single process between the open and close processes can access
the client data deduplication cache. Multi-session backups from the same process
can use the deduplication cache. In general, processes that do not get the lock on
the cache can perform client-side data deduplication. However, the processes
cannot query or store hash sums that are found in the cache.

Insufficient file space

The Tivoli Storage Manager client cannot use a cache for data deduplication if
there is not enough file space for a hash sum cache. Client-side data deduplication
can take place, but it has no memory of hash sums that are already sent by the
client or already found on the server. Data deduplication, generally, must query the
server to find out if hash sums are duplicates. Hash sum lists are maintained in
memory for the life of a transaction. If a hash sum is encountered multiple times
within the same transaction, the hash sum is detectable without a cache.

Synchronization of the cache for client-side data deduplication

The cache for client-side data deduplication can become unsynchronized with the
deduplicated disk storage pool of the server. Object expiration, file space deletion,
or overflow to an associated tape storage pool can cause the cache to be
unsynchronized. When the client cache contains entries that are no longer in the
deduplicated storage pool of the Tivoli Storage Manager server, the client cache
resets. The client cache cannot delete specific entries when objects are deleted from
the storage pool of the server.

In environments where it is likely that the client deduplication cache would

become unsynchronized with the deduplicated storage pool of the server, use
server-side data deduplication. The cache can become unsynchronized during a
backup or an API application function. The Tivoli Storage Manager client can back
up data again, but the Tivoli Storage Manager API cannot. Wherever it is difficult
to perform a backup again, do not use the cache of client-side data deduplication.

Restoring deduplicated data from backup sets

You can restore deduplicated objects to the same storage pool or to another storage
pool that uses or does not use data deduplication.

When a backup set is created for a node by using the GENERATE BACKUPSET
command, all associated node data is placed onto the backup media. It is also
placed on the backup media when node data is exported for a node by the EXPORT
NODE command. This placement ensures that the associated objects can be restored
without any server dependencies, apart from the backup media.

When you restore deduplicated objects, consider the following situations:

328 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 v If the object is restored to the same storage pool, for example, by the RECLAIM
STGPOOL command, any duplicate data blocks are replaced with references to
deduplicated extents. These duplicate data blocks were identified by the
IDENTIFY processes.
v If a deduplicated object is moved to another storage pool that uses
deduplication, duplicate extents are processed as the object is moved. The object
is placed in the target storage pool in a deduplicated state, and no additional
processing is needed on that object. The deduplicated object is moved, for
example, by the MOVE VOLUME command or by storage pool migration.
v If a deduplicated object is moved to a storage pool that does not use
deduplication, that object is created in the target storage pool. All associated
data is also in the target storage pool.

Data deduplication and data compatibility

With client-side data deduplication, data is compressed extent by extent, which
means data compatibility is possible between the client and the server and between
different versions of Tivoli Storage Manager. Also, data extents that are created by
different operations are compatible, which can mean that a greater proportion can
be deduplicated.

Compression

With client-side data deduplication, compression is done extent by extent. When

you use client-side data deduplication, data is compatible between compressed
extents and extents that are not compressed. This compatibility means that extents
can be shared between the client and server, regardless of how the data was
backed up. Compression can save storage space, however, it can increase processor
usage on the client system.

Consider the following factors when you use data compression in an environment
that uses multiple clients:
v Extents that are compressed by a backup-archive client that uses Tivoli Storage
Manager V6.1 or earlier are not compatible with compressed extents from a V6.2
client. Extents are also not compatible with uncompressed extents because each
version uses a different compression algorithm.
v With a deduplication storage pool that contains data from clients that are V6.2
and earlier, there is a mixture of compressed and non-compressed extents. For
example, assume that a restore operation is run from a client that is earlier than
V6.2. Compressed extents from a client at a later version of Tivoli Storage
Manager are uncompressed by the server during the restore operation.
v When backup sets are generated for clients that are at a version earlier than
V6.2, V6.2 compressed extents that are also part of the data to be backed up are
uncompressed.

Even though most data is compatible when using compression, ensure that all
clients are at V6.2 and later. This method minimizes the need for data compression
when you restore data or create a backup set.

Compatibility with data from earlier client versions

Data that is stored by earlier client versions and processed for deduplication
extents by the server is compatible with new extents. For example, an extent that is
identified by the server from an earlier client version matches the query from
client-side data deduplication to the server. The extent is not sent to the server,

Chapter 10. Managing storage pools and volumes 329

 only a reference to it is sent. Therefore, in the same storage pool, extents from
different levels of the client are equivalent. Regardless of where extents are
processed, whether on the server or on the client, they are compatible. Migration of
client and server versions is possible without the data being sent again to the
server.

File image and Tivoli Storage Manager FastBack® mount backups

Data extents that are created by different operations are compatible. For example,
data extents are compatible that are created by file-level, image, or IBM Tivoli
Storage Manager FastBack mount backups. This can mean that a greater proportion
of the extents can be deduplicated.

Image and Tivoli Storage Manager mount backups

Assume that you integrate the Tivoli Storage Manager FastBack mount with Tivoli
Storage Manager to back up volumes to a Tivoli Storage Manager server. The
Tivoli Storage Manager client backs up the Tivoli Storage Manager FastBack
repository to a remote server. You previously performed an image or a file-level
backup of this data with Tivoli Storage Manager client. Then it is likely that the
Tivoli Storage Manager FastBack mount backup can use many data extents that are
already stored on the server.

For example, you perform an image backup of a volume that uses the Tivoli
Storage Manager client. Then you back up the same volume with Tivoli Storage
Manager FastBack. You can expect a greater amount of data deduplication when
you back up the Tivoli Storage Manager FastBack mount.

Image and file-level backups

Data extents that are created by a file-level backup can be used by the Tivoli
Storage Manager client during an image backup. For example, you perform a full
incremental backup of the C drive on your computer. Then you run an image
backup of the same drive. You can expect a greater amount of data deduplication
during the image backup. You can also expect a greater amount of data
deduplication during a file-level backup or an archive operation that immediately
follows an image backup.

Data deduplication and disaster recovery management

Use the following scenarios to help you consider factors that can affect disaster
recovery management when you use deduplicated primary or copy storage pools.

Data deduplication is only permitted for storage pools that are associated with a
devtype=FILE device class. The following scenarios show how you can implement
the data deduplication of storage pools to ensure that you can restore data if a
failure occurs.
Primary storage pool is deduplicated and a single copy storage pool is not
deduplicated
The amount of time required to back up the primary storage pool to a
non-deduplicated copy storage pool can increase. While data is copied to
the copy storage pool, the deduplicated data that represents a file must be
read. The file must be recreated and stored in the copy storage pool.

330 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 With this scenario, to recover data at a disaster recovery site, you can use
disaster recovery management and copy storage pool volumes that are
off-site.
Primary storage pool is deduplicated and multiple copy storage pools are not
deduplicated
The amount of time required to back up the primary storage pool can
increase. This is because of the reads that are necessary to get the various
data that represent an object written to one of these copy storage pools.
With this scenario, to recover data at a disaster recovery site, you can use
disaster recovery management and copy storage pool volumes that are
off-site. There is also a copy of the storage pool data on-site, in addition to
the off-site support.
Primary storage pool is deduplicated and a single copy storage pool is
deduplicated
The main aim of this scenario is space savings at the primary server
location. However, you cannot use disaster recovery management. Also,
you cannot use the off-site rotation of copy storage pool volumes because
the deduplicated copy storage pool can be achieved only by using
devtype=FILE volumes.
With this scenario, you can have a risk in that there is no off-site copy of
the data. The off-site copy could be used if the primary server location,
hardware, or data center is damaged or destroyed. This scenario can result
in the loss of data or the inability to recover that data.
Primary storage pool is deduplicated, a single copy storage pool is not
deduplicated, and a single copy storage pool is deduplicated
The deduplicated copy storage pool provides space savings for the primary
server location and the duplication of the data in the primary storage pool.
You use server-side data deduplication at the level of a single storage pool.
Therefore, a deduplicated primary storage pool and a deduplicated copy
storage pool reference different data. They also maintain different database
references to track and manage the data that represent a given file.
Alternatively, you can use the non-deduplicated copy storage pool, which
is probably tape, along with disaster recovery management as an off-site
data protection site. You can use the copy storage pool volumes that are
off-site, together with an appropriate database backup, to restore the
server. You can also use the volumes to restore or retrieve data.

Writing data simultaneously to primary, copy, and active-data pools

With IBM Tivoli Storage Manager, you can write data simultaneously to a primary
storage pool, copy storage pools, and active-data pools. The simultaneous-write
function increases your level of data protection and reduces the amount of time
required for storage pool backup.

About this task

You can write data simultaneously during any of the following operations:
v Client store sessions, for example:
– Backup and archive sessions by Tivoli Storage Manager backup-archive
clients.
– Backup and archive sessions by application clients using the Tivoli Storage
Manager API.

Chapter 10. Managing storage pools and volumes 331

 – Migration processes by hierarchical storage management (HSM) clients.
Migrated data is simultaneously written only to copy storage pools. Migrated
data is not permitted in active-data pools.
v Server migration of data within a storage pool hierarchy.
v Server import processes that involve copying exported file data from external
media to a primary storage pool that is configured for the simultaneous-write
function. Imported data is simultaneously written to copy storage pools.
Imported data is not simultaneously written to active-data pools. To store newly
imported data into an active-data pool, use the COPY ACTIVEDATA command.

The maximum number of copy storage pools and active-data pools to which data
can be simultaneously written is three. For example, you can write data
simultaneously to three copy storage pools, or you can write data simultaneously
to two copy storage pools and one active-data pool.

Attention: Do not use the simultaneous-write function to replace the task of

regularly backing up storage pools. If you use the function to simultaneously write
to copy storage pools, active-data pools, or both, ensure that the copy of each
primary storage pool is complete by regularly issuing the BACKUP STGPOOL
command and the COPY ACTIVEDATA command. If you fail to regularly back up
storage pools. you can lose the ability to recover primary storage pool data. For
example, if a copy storage pool fails during a write operation and the
COPYCONTINUE parameter is set to YES, the Tivoli Storage Manager server removes
the failed copy storage pool from the copy pool list for the remainder of the client
session. After the copy storage pool is removed, the Tivoli Storage Manager server
continues to write to the primary storage pool and to any remaining copy storage
pools and active-data pools. If these pools become damaged or lost, and if you did
not issue the BACKUP STGPOOL command for the copy storage pool that failed, you
might not be able to recover your data.

Data that is simultaneously written to copy storage pools or active-data pools

during migration is not copied when primary storage pools are backed up or when
active data is copied.

Guidelines for using the simultaneous-write function

The goal of the simultaneous-write function is to minimize the amount of time that
is required for storage-pool backup operations. IBM Tivoli Storage Manager
provides several options for accomplishing this goal. Your choice depends on how
you want to manage your environment.

You can specify the simultaneous-write function for a primary storage pool if it is
the target for client store sessions, server import processes, or server
data-migration processes. You can also specify the simultaneous-write function for
a primary storage pool when it is the target for all of the eligible operations.

Writing data simultaneously during client store sessions might be the logical choice
if you have sufficient time for mounting and removing tapes during the client store
session. However, if you choose this option you must ensure that a sufficient
number of mount points and drives are available to accommodate all the client
nodes that are storing data.

As a best practice, you are probably issuing the BACKUP STGPOOL and COPY
ACTIVEDATA commands for all the storage pools in your storage pool hierarchy. If
you are, and if you migrate only a small percentage of data from the primary
storage pool daily, writing data simultaneously during client store sessions is the

332 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 most efficient option. This method is efficient because data is stored in copy
storage pools and active-data pools when the client stores the data. Little or no
data is left to copy during backup storage pool or copy active-data operations.

Writing data simultaneously during server data-migration processes is another

option. It is the most efficient method of writing data simultaneously if you
migrate all the data in your primary storage pool nightly and then back up the
primary storage pools. The reason for the efficiency is that data written
simultaneously to copy storage pools or active-data pools during migration is not
copied during backup storage pool or copy active-data operations.

Use the simultaneous-write function during migration if you have many client
nodes and the number of mount points that are required to write data
simultaneously during client store sessions is unacceptable. Similarly, mounting
and removing tapes when writing data simultaneously during client store sessions
might be taking too much time. If so, consider writing data simultaneously during
migration.

Another option is to specify the simultaneous-write function for a primary storage

pool if it is the target for any of the eligible operations (client store, server import,
server migration). This choice is preferable if, for example, you have large files to
back up (for example, image files, database backup files, or Tivoli Data Protection
files). Instead of backing up these files to the random-access disk storage pool at
the top of the storage hierarchy, you can back them up to the sequential-access
disk storage pool that is the next storage pool in the hierarchy. If you specify the
simultaneous-write function for any of the eligible operations for the next storage
pool, the following events occur:
v Large files that are backed up to the next storage pool are simultaneously
written to copy storage pools (and active-data pools, if you have them).
v Other files that migrate to the next storage pool are simultaneously written to
the same copy storage pools (and active-data pools, if you have them).

By default, the Tivoli Storage Manager server writes data simultaneously during
client store sessions if you have copy storage pools or active-data pools defined to
the target storage pool.

You can also disable the simultaneous-write function. This option is useful if you
have copy storage pools or active-data pools defined, but you want to disable the
simultaneous-write function without deleting and redefining the pools.

Limitations that apply to simultaneous-write operations

Using the simultaneous-write function requires consideration of factors such as
storage devices and network configuration.

The following limitations apply:

v Data cannot be written simultaneously to copy storage pools and active-data
pools during server data movements such as:
– Reclamation
– MOVE DATA processes
– MOVE NODEDATA processes
– Storage pool backup
– Copy of active data to an active-data storage pool

Chapter 10. Managing storage pools and volumes 333

 The only server-data movement process that can write data simultaneously is
migration.
v Simultaneous-write operations take precedence over LAN-free data movement.
The operations go over the LAN, and the simultaneous-write configuration is
accepted.
v You can back up or archive copies of files that were migrated by a Tivoli Storage
Manager for Space Management client to the same Tivoli Storage Manager
server to which they were migrated. However, the files are stored only in the
primary storage pool. As a best practice, create current backup and archive
versions of the files before the Tivoli Storage Manager for Space Management
client migrates them.
v Target storage pools that are used for simultaneous-write operations can have
different device classes if the device classes have compatible data formats.
Performance is limited by the speed of the slowest device.
v Data deduplication has the following effects on simultaneous-write operations:
Simultaneous write during client store operations
If client-side deduplication is enabled, it is disabled before
simultaneous-write operations occur. Non-deduplicated data is
simultaneously written to target storage pools.
Simultaneous write during server migration of data
If a primary storage pool is set up for server-side deduplication, all copy
storage pools and active-data pools are removed from
simultaneous-write operations. For example, suppose that your
simultaneous-write configuration comprises a primary storage pool, a
copy storage pool, and an active-data pool. The primary storage pool is
set up for server-side deduplication. Data is written only to the primary
storage pool.
If a primary storage pool is not set up for server-side deduplication, data
is written simultaneously only to copy storage pools or active data pools
that are not set up for server-side deduplication. For example, suppose
that your simultaneous-write configuration comprises a primary storage
pool, a copy storage pool, and an active-data pool. The primary storage
pool and the active-data pool are not set up for server-side
deduplication. However, the copy storage pool is set up for server-side
deduplication. When simultaneous-write operations occur, data is
written only to the primary storage pool and the active-data pool.
v You cannot use the simultaneous-write function with Centera storage devices.
v The COPYSTGPOOLS and ACTIVEDATAPOOLS parameters are available only to primary
storage pools that use NATIVE or NONBLOCK data format. This parameter is
not available for storage pools that use the following data formats:
– NETAPPDUMP
– CELERRADUMP
– NDMPDUMP
v Limitations apply when a NAS backup operation is writing a TOC file. If the
primary storage pool that is specified in the TOCDESTINATION in the copy
group of the management class has copy storage pools or active-data pools
defined, the copy storage pools and active-data pools are ignored. The data is
stored only in the primary storage pool.

334 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Controlling the simultaneous-write function
You control the simultaneous-write function by specifying certain parameters when
you define or update primary storage pools. You can control when data is written
simultaneously. You can also specify the copy storage pools and active-data pools
to which data is simultaneously written.

Specifying when the simultaneous-write operations occur

You can specify simultaneous-write operations for any primary storage pool that is
the target of client store sessions, server import processes, or server data-migration
processes. You can also disable the simultaneous-write function.

About this task

To control the simultaneous-write function, use the AUTOCOPY parameter on the

DEFINE STGPOOL or UPDATE STGPOOL commands for primary storage pools.

Remember:
v Specify a value for the AUTOCOPY parameter on the primary storage pool that is
the target of data movement. (The default is to write data simultaneously during
client store sessions and server import processes.) For example, if you want to
write data simultaneously only during server data-migration processes, specify
AUTOCOPY=MIGRATION in the definition of the next storage pool in the storage pool
hierarchy.
v The AUTOCOPY parameter is not available for copy storage pools or active-data
pools.

IBM Tivoli Storage Manager provides the following options for controlling when
simultaneous-write operations occur:

Procedure
v To disable the simultaneous-write function, specify AUTOCOPY=NONE.
This option is useful, if, for example, you have copy storage pools or active-data
pools defined, and you want to temporarily disable the simultaneous-write
function without having to delete and then redefine the pools.
v To specify simultaneous-write operations only during client store sessions and
server import processes, specify AUTOCOPY=CLIENT.
During server import processes, data is simultaneously written only to copy
storage pools. Data is not written to active-data pools during import processes.
v To specify that simultaneous-write operations take place only during server
data-migration processes, specify AUTOCOPY=MIGRATION.
During server data migration, data is simultaneously written to copy storage
pools and active-data pools only if the data does not exist in those pools.
v To specify that simultaneous-write operations take place during client store
sessions, server data-migration processes, and server import processes, specify
AUTOCOPY=ALL.
A primary storage pool can be the target for more than one type of data
movement. For example, the next storage pool in a storage pool hierarchy can be
the target for data migration from the primary storage pool at the top of the
hierarchy. The next storage pool can also be the target for direct backup of
certain types of client files (for example, image files). The AUTOCOPY=ALL setting
on a primary storage pool ensures that data is written simultaneously during
both server data-migration processes and client store sessions.

Chapter 10. Managing storage pools and volumes 335

 Example

The following table provides examples of AUTOCOPY settings for some common
scenarios in which the simultaneous-write function is used.
Table 30. AUTOCOPY settings
Set the AUTOCOPY parameter
for the primary storage pool Set the AUTOCOPY parameter
at the top of the storage for the next primary storage
If your goal is... hierarchy to... pool to...
To disable the NONE NONE
simultaneous-write function
To enable simultaneous-write CLIENT NONE
operations only during client
store sessions and server
import processes
To enable simultaneous-write NONE MIGRATION
operations only during
server data-migration
processes
To simultaneously write NONE ALL
client files to copy storage
pools during migration. You
also want simultaneous-write
operations to occur for files
that are directly backed up to
the next storage pool.
To enable simultaneous-write ALL ALL
operations during any of the
following operations: client
store sessions, server import
processes, and server
data-migration processes.

For details about the DEFINE STGPOOL and UPDATE STGPOOL commands and
parameters, see the Administrator's Reference.

Specifying copy pools and active-data pools for

simultaneous-write operations
The maximum number of copy storage pools and active-data pools to which data
can be simultaneously written is three. For example, you can write data
simultaneously to three copy storage pools. You can also write data simultaneously
to two copy storage pools and one active-data pool, and so on.

About this task

The parameters that are used to specify copy storage pools and active-data pools
are on the DEFINE STGPOOL and UPDATE STGPOOL commands.

Procedure
v To specify copy storage pools, use the COPYSTGPOOLS parameter.
v To specify active-data pools, use the ACTIVEDATAPOOLS parameter.

336 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Results

Ensure that client sessions have sufficient mount points. Each session requires one
mount point for the primary storage pool and a mount point for each copy storage
pool and each active-data pool. To allow a sufficient number of mounts points, use
the MAXNUMMP parameter on the REGISTER NODE or UPDATE NODE commands.

Example

For details about the DEFINE STGPOOL and UPDATE STGPOOL commands, refer to the
Administrator's Reference.
Related concepts:
“Rules of inheritance for the simultaneous-write function” on page 338

Specifying how the server reacts to a write failure during

simultaneous-write operations
Write failures to copy or active-data pools might occur during while data is being
simultaneously written. If a write failure occurs during a client store session, you
can specify whether to continue or discontinue the operation.

About this task

Use the COPYCONTINUE parameter on the DEFINE STGPOOL command to specify how
the server reacts to a write failure to copy storage pools during client store
sessions:

Procedure
v To stop writing to failing copy storage pools for the remainder of the session,
but continue storing files into the primary pool and any remaining copy pools or
active-data pools, specify COPYCONTINUE=YES.
The copy storage pool list is active only for the life of the session and applies to
all the primary storage pools in a particular storage pool hierarchy.
v To fail the transaction and discontinue the store operation, specify
COPYCONTINUE=NO.

What to do next

Restrictions:
v The setting of the COPYCONTINUE parameter does not affect active-data pools. If a
write failure occurs for any of active-data pools, the server stops writing to the
failing active-data pool for the remainder of the session, but continues storing
files into the primary pool and any remaining active-data pools and copy
storage pools. The active-data pool list is active only for the life of the session
and applies to all the primary storage pools in a particular storage pool
hierarchy.
v The setting of the COPYCONTINUE parameter does not affect the simultaneous-write
function during server import. If data is being written simultaneously and a
write failure occurs to the primary storage pool or any copy storage pool, the
server import process fails.
v The setting of the COPYCONTINUE parameter does not affect the simultaneous-write
function during migration. If data is being written simultaneously and a write
failure occurs to any copy storage pool or active-data pool, the failing storage
pool is removed and the data migration process continues. Write failures to the
primary storage pool cause the migration process to fail.

Chapter 10. Managing storage pools and volumes 337

 For details about the DEFINE STGPOOL and UPDATE STGPOOL commands and
parameters, refer to the Administrator's Reference.
Related concepts:
“Rules of inheritance for the simultaneous-write function”

Rules of inheritance for the simultaneous-write function

When switching primary storage pools during client store sessions or server import
processes, certain rules of inheritance apply to copy storage pool lists, active-data
pool lists, and the setting of the COPYCONTINUE parameter.

When a client backs up, archives, or migrates a file, or when the server imports
data, the data is written to the primary storage pool that is specified by the copy
group of the management class that is bound to the data. If a data storage
operation or a server import operation switches from the primary storage pool at
the top of a storage hierarchy to a next primary storage pool in the hierarchy, the
next storage pool inherits the list of copy storage pools, the list of active-data
pools, and the value of the COPYCONTINUE parameter from the primary storage pool
at the top of the storage pool hierarchy.

The following rules apply during a client store session or a server import process
when the server must switch primary storage pools:
v If the destination primary storage pool has one or more copy storage pools or
active-data pools defined using the COPYSTGPOOLS or ACTIVEDATAPOOLS
parameters, the server writes the data to the next storage pool and to the copy
storage pools and active-data pools that are defined to the destination primary
pool, regardless whether the next pool has copy pools defined.
The setting of the COPYCONTINUE of the destination primary storage pool is
inherited by the next primary storage pool. The COPYCONTINUE parameter
specifies how the server reacts to a copy storage-pool write failure for any of the
copy storage pools listed in the COPYSTGPOOLS parameter. If the next pool has
copy storage pools or active-data pools defined, they are ignored as well as the
value of the COPYCONTINUE parameter.
v If no copy storage pools or active-data pools are defined in the destination
primary storage pool, the server writes the data to the next primary storage
pool. If the next pool has copy storage pools or active-data pools defined, they
are ignored.
These rules apply to all the primary storage pools within the storage pool
hierarchy.
Related tasks:
“Specifying copy pools and active-data pools for simultaneous-write operations”
on page 336
“Specifying how the server reacts to a write failure during simultaneous-write
operations” on page 337

338 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Simultaneous-write operations: Examples
Illustrated examples of simultaneous-write operations show how the function
works in various storage pool configurations. Other examples show what happens
when an error occurs during a simultaneous-write operation.

Examples of simultaneous-write operations during client store

operations
Examples show how the simultaneous-write function works during client store
operations. In all the examples, client nodes, whose files require fast restore, are
members of a policy domain that specifies an active-data pool.

For these examples, assume the following conditions:

v Primary storage pools DISKPOOL and TAPEPOOL are linked to form a storage
hierarchy. DISKPOOL is at the top of the storage hierarchy and TAPEPOOL is
the next pool in the storage hierarchy.
v The value of the AUTOCOPY parameter for DISKPOOL is CLIENT. The value of the
AUTOCOPY parameter for TAPEPOOL is NONE.
v The active backup data belonging to certain clients must be restored as quickly
as possible if a disaster occurs. These clients are members of policy domain
FASTRESTORE, which specifies an active-data pool as the destination for active
backup data. Files A and B belong to a node in this domain and are bound to
management class STANDARD. The destination specified in its backup copy
group is DISKPOOL.
v The data belonging to other nodes is less critical. Restore times are flexible.
These nodes are assigned to policy domain NORMAL, which does not have an
active-data pool specified. Files C, D, and E belong to one of the nodes in this
domain and are bound to management class STANDARD. The destination that
is specified in its backup copy group is DISKPOOL.
v DISKPOOL has enough space to store only files C and D, but its next pool
(TAPEPOOL) has enough space for file E.
Related concepts:
Chapter 13, “Implementing policies for client data,” on page 475

Writing data simultaneously during a simple client store session:

In this example, the simultaneous-write function automatically copies client data to

two copy storage pools and an active-data pool during a client store operation.

With DISKPOOL and TAPEPOOL already defined as your storage pool hierarchy,
issue the following commands to enable the simultaneous-write function:
define stgpool copypool1 mytapedevice pooltype=copy
define stgpool copypool2 mytapedevice pooltype=copy
define stgpool activedatapool mydiskdevice pooltype=activedata
update stgpool diskpool copystgpools=copypool1,copypool2 copycontinue=yes
activedatapools=activedatapool

where MYTAPEDEVICE is the device-class name associated with the copy storage
pools and MYDISKDEVICE is the device-class name associated with the
active-data pool.

The storage pool hierarchy and the copy storage pools and active-data pool
associated with DISKPOOL are displayed in Figure 24 on page 340.

Chapter 10. Managing storage pools and volumes 339

 Clients Clients

Policy Domain - NORMAL Policy Domain - FASTRESTORE

Policy Set Policy Set
STANDARD STANDARD
Management Class Management Class
Backup Backup
Copy Copy
Group Group

to
Po

ts
in

in
ts

Po
to

ACTIVEDATAPOOL

DISKPOOL

COPYPOOL2
TAPEPOOL
COPYPOOL1

Figure 24. Example of storage pool hierarchy with copy storage pools defined for DISKPOOL

During a simultaneous-write operation, the next storage pool TAPEPOOL inherits

the list of copy storage pools (COPYPOOL1 and COPYPOOL2) and the value of
the COPYCONTINUE parameter from DISKPOOL, the primary pool at the top of
the storage pool hierarchy. TAPEPOOL also inherits the list of active-data pools
(ACTIVEDATAPOOL). When files A, B, C, D, and E are backed up, the following
events occur:
v A and B are written to DISKPOOL, COPYPOOL1, COPYPOOL2, and
ACTIVEDATAPOOL.
v C and D are written to DISKPOOL, COPYPOOL1, and COPYPOOL2.
v File E is written to TAPEPOOL, COPYPOOL1 and COPYPOOL2.
See Figure 25 on page 341.

340 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 Client in ACTIVEDATAPOOL
FASTRESTORE
domain Server
A
B
COPYPOOL2
B C B A
A E D

E
D
C
B
C C A COPYPOOL1
D D
E B
A

E
Client in DISKPOOL
NORMAL domain

next pool

TAPEPOOL

Figure 25. Inheriting a list of copy storage pools

As a precaution, issue the BACKUP STGPOOL and COPY ACTIVEDATA commands after
the backup operation has completed. Data that is simultaneously written to copy
storage pools or active-data pools during migration is not copied when storage
pools are backed up or when active data is copied.

Inheritance of an empty copy storage-pool list during a simultaneous-write

operation:

In this example, the next storage pool in a hierarchy inherits empty copy storage
pool and active-data pool lists from the primary storage pool at the top of the
storage hierarchy.

You do not specify a list of copy storage pools for DISKPOOL. However, you do
specify copy storage pools for TAPEPOOL (COPYPOOL1 and COPYPOOL2) and
an active-data pool (ACTIVEDATAPOOL). You also specify a value of YES for the
COPYCONTINUE parameter. Issue the following commands to enable the
simultaneous-write function:
define stgpool copypool1 mytapedevice pooltype=copy
define stgpool copypool2 mytapedevice pooltype=copy
define stgpool activedatapool mydiskdevice pooltype=activedata
update stgpool tapepool copystgpools=copypool1,copypool2
copycontinue=yes activedatapools=activedatapool

where MYTAPEDEVICE is the device-class name associated with the copy storage
pools and MYDISKDEVICE is the device-class name associated with the
active-data pool. Figure 26 on page 342 displays this configuration.

Chapter 10. Managing storage pools and volumes 341

 Clients Clients

Policy Domain - NORMAL Policy Domain - FASTRESTORE

Policy Set Policy Set
STANDARD STANDARD
Management Class Management Class
Backup Backup
Copy Copy
Group Group

Po

to
int

ts
s

in
to

Po
DISKPOOL

TAPEPOOL ACTIVEDATAPOOL

COPYPOOL2
COPYPOOL1

Figure 26. Example of storage pool hierarchy with copy storage pools defined for TAPEPOOL

When files A, B, C, D, and E are backed up, the following events occur:
v A, B, C, and D are written to DISKPOOL.
v File E is written to TAPEPOOL.
See Figure 27 on page 343.

342 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 Client in ACTIVEDATAPOOL
FASTRESTORE
domain Server

COPYPOOL2
B
A

C C COPYPOOL1
D D
E B
A

E
Client in DISKPOOL
NORMAL domain

next pool

TAPEPOOL

Figure 27. Inheriting an empty copy storage pool list

Although TAPEPOOL has copy storage pools and an active-data pool defined, file
E is not copied because TAPEPOOL inherits empty copy storage pool and
active-data pool lists from DISKPOOL.

As a precaution, issue the BACKUP STGPOOL and COPY ACTIVEDATA commands after
the backup operation has completed. Data that is simultaneously that is written to
copy storage pools or active-data pools during migration is not copied when
primary storage pools are backed up or when active data is copied.

A simultaneous-write error during a client store operation:

In this example, data is not written to one copy storage pool in a

simultaneous-write operation.

You specify COPYPOOL1 and COPYPOOL2 as copy storage pools for DISKPOOL
and you set the value of the COPYCONTINUE parameter to YES. You also specify
ACTIVEDATAPOOL as the active-data pool for DISKPOOL. This configuration is
identical to the configuration in the first example.

When files A, B, C, D, and E are backed up, the following events occur:
v An error occurs while writing to COPYPOOL1, and it is removed from the copy
storage pool list that is held in memory by the server. The transaction fails.
v Because the value of the COPYCONTINUE parameter is YES, the client tries the
backup operation again. The in-memory copy storage pool list, which is retained
by the server for the duration of the client session, no longer contains
COPYPOOL1.
v Files A and B are simultaneously written to DISKPOOL, ACTIVEDATAPOOL,
and COPYPOOL2.
v Files C and D are simultaneously written to DISKPOOL and COPYPOOL2.
v File E is simultaneously written to TAPEPOOL and COPYPOOL2.
See Figure 28 on page 344.

Chapter 10. Managing storage pools and volumes 343

 Client in
ACTIVEDATAPOOL
FASTRESTORE
domain
Server A
B
COPYPOOL2
B B A
A D C
retry E
tran
sact
ion
COPYPOOL1
D (removed for
C
D C the duration of
on
E acti B the session)
t rans
retry A

E
DISKPOOL
Client in
NORMAL domain
next pool

TAPEPOOL

Figure 28. Inheriting a list of copy storage pools

In this scenario, if the primary storage pools and COPYPOOL2 become damaged
or lost, you might not be able to recover your data. For this reason, issue the
following BACKUP STGPOOL command for the copy storage pool that failed:
backup stgpool diskpool copystgpool1
backup stgpool tapepool copystgpool1

Suppose, in this scenario, that an error occurred while writing to

ACTIVEDATAPOOL, rather than COPYPOOL1. In this situation,
ACTIVEDATAPOOL is removed from the active-data pool list held in memory by
the server, and the transaction fails. The client tries the backup operation again.
The in-memory active-data pool list does not contain ACTIVEDATAPOOL. Files A,
B, C, and D are written simultaneously to DISKPOOL, COPYPOOL1, and
COPYPOOL2. File E is written to TAPEPOOL, COPYPOOL1, and COPYPOOL2.
However, files A and B are not written to the active-data pool.

You can still recover the primary storage pools from COPYPOOL1 and, if
necessary, COPYPOOL2. However, if you want active backup data available in the
active-data pool for fast client restores, you must issue the following command:
copy activedata diskpool activedatapool

As a precaution, issue the BACKUP STGPOOL and COPY ACTIVEDATA commands after
the backup operation has completed. Data that is simultaneously written to copy
storage pools or active-data pools during migration is not copied when primary
storage pools are backed up or when active data is copied.

344 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Examples of simultaneous-write operations during server
data-migration processes
Examples show how the simultaneous-write function works during server
migration of data within a storage pool hierarchy.

For these examples, assume the following conditions:

v Primary storage pools FILEPOOL, which is associated with a sequential-access
device (device type FILE), and TAPEPOOL are linked to form a storage
hierarchy. FILEPOOL is at the top of the storage hierarchy. TAPEPOOL is the
next pool in the storage hierarchy.
v The files in FILEPOOL are eligible to be migrated.
v One or more copy storage pools are defined to FILEPOOL and TAPEPOOL.
There are no active-data pools.
v Copies of one or more of the files in FILEPOOL exist in a copy storage pool.

Simultaneous-write operation in a simple migration scenario:

In this example, the storage pool hierarchy contains two primary storage pools.
The next storage pool has two copy storage pools defined. A copy of one of the
files to be migrated to the next storage pool exists in one of the copy storage pools.

FILEPOOL and TAPEPOOL are defined in your storage pool hierarchy. Two copy
storage pools, COPYPOOL1 and COPYPOOL2, are defined to TAPEPOOL. Files A,
B, and C are in FILEPOOL and eligible to be migrated. A copy of file C exists in
COPYPOOL2.

The storage pool hierarchy and the copy storage pools that are associated with
TAPEPOOL are displayed in Figure 29.

FILEPOOL A B
C
COPYPOOL2

Next pool

TAPEPOOL

COPYPOOL1
C

Figure 29. A storage pool hierarchy with files eligible to be migrated

To specify simultaneous-write operations only during migration, issue the

following command for TAPEPOOL:
update stgpool tapepool autocopy=migration

Tip: In this example, the setting of the AUTOCOPY parameter for FILEPOOL is not
relevant. TAPEPOOL is the target of the data migration.

Chapter 10. Managing storage pools and volumes 345

 When files A, B and C are migrated, the following events occur:
v Files A and B are simultaneously written to TAPEPOOL, COPYPOOL1 and
COPYPOOL2.
v File C is simultaneously written to TAPEPOOL and COPYPOOL2. File C is not
written to COPYPOOL1 because COPYPOOL1 has a copy of the file.
See Figure 30.

FILEPOOL

COPYPOOL2
up
a ck
lb
o C
Next pool
e po
A ag B
or
B St A
C

TAPEPOOL Sto A
rag B
ep
oo
l ba
cku COPYPOOL1
p
C

Figure 30. Simultaneous-write operation during migration to two copy storage pools

As a precaution, issue the BACKUP STGPOOL and COPY ACTIVEDATA commands after
the migration operation has completed. Data that is simultaneously written to copy
storage pools or active-data pools during migration is not copied when primary
storage pools are backed up or when active data is copied.

A simultaneous-write error during server data migration:

In this example, the storage pool hierarchy contains two primary storage pools.
The next storage pool has two copy storage pools defined. A copy of one of the
files to be migrated to the next storage pool exists in a copy storage pool. A write
error to the pool occurs.

FILEPOOL and TAPEPOOL are defined in the storage pool hierarchy. Two copy
storage pools, COPYPOOL1 and COPYPOOL2, are defined to TAPEPOOL. Files A,
B, and C are in FILEPOOL and are eligible to be migrated. A copy of file C exists
in COPYPOOL1.

The storage pool hierarchy and the copy storage pools that are associated with
TAPEPOOL are displayed in Figure 31 on page 347.

346 IBM Tivoli Storage Manager for Windows: Administrator's Guide

FILEPOOL A B
C
COPYPOOL2

Next pool

TAPEPOOL

COPYPOOL1
C

Figure 31. A storage pool hierarchy with files eligible to be migrated

To specify simultaneous-write operations only during migration, issue the

following command for TAPEPOOL:
update stgpool tapepool autocopy=migration

Tip: In this example, the setting of the AUTOCOPY parameter for FILEPOOL is not
relevant. TAPEPOOL is the target of the data migration.

When files A, B and C are migrated, the following events occur:

v An error occurs writing to COPYPOOL1.
v COPYPOOL1 is removed from the in-memory list. The in-memory list is kept for
the duration of the migration process.
v The transaction fails and the server tries the operation again.
v Files A, B, and C are simultaneously written to TAPEPOOL and COPYPOOL2.
See Figure 32.

FILEPOOL

COPYPOOL2
up
a ck
o lb C
Next pool
e po
A ag B
or
B St A
C

COPYPOOL1
TAPEPOOL C
(Removed for the
duration of the
migration process)

Figure 32. An error occurs during simultaneous-write operation during migration

Chapter 10. Managing storage pools and volumes 347

 As a precaution, issue the BACKUP STGPOOL and COPY ACTIVEDATA commands after
the migration operation has completed. Data that is simultaneously written to copy
storage pools or active-data pools during migration is not copied when primary
storage pools are backed up or when active data is copied.

Inheriting a list of copy storage pools during a simultaneous-write operation:

In this example, three primary storage pools are linked to form a storage pool
hierarchy. The next storage pool in the hierarchy has a storage pool list. The last
pool in the hierarchy inherits the list during a simultaneous-write operation.

FILEPOOL1, FILEPOOL2, and TAPEPOOL are defined in your storage pool

hierarchy. One copy storage pool, COPYPOOL, is defined to FILEPOOL2.
v Files A, B, and C on FILEPOOL1 are eligible to be migrated.
v FILEPOOL2 has sufficient space only for files B and C, but not A. TAPEPOOL
has enough space for file A.
v A copy of file C exists in COPYPOOL.

The storage pool hierarchy and the copy storage pool are displayed in Figure 33.

FILEPOOL1 A B
C

Next pool

COPYPOOL1
C

FILEPOOL2

Next pool

TAPEPOOL

Figure 33. Three-tiered storage pool hierarchy with one copy storage pool

Issue the following commands for FILEPOOL2 and TAPEPOOL to enable the
simultaneous-write function only during migration:
update stgpool filepool2 autocopy=migration
update stgpool tapepool autocopy=migration

Tip: In this example, the setting of the AUTOCOPY parameter for FILEPOOL1 is not
relevant. FILEPOOL2 and TAPEPOOL are the targets of the data migration.

When files A, B, and C are migrated, the follow events occur:

v File B is migrated to FILEPOOL2 and simultaneously written to COPYPOOL.
348 IBM Tivoli Storage Manager for Windows: Administrator's Guide
v File C is migrated to FILEPOOL2. It is not written to COPYPOOL because a
copy of that file exists in COPYPOOL.
v File A is migrated to TAPEPOOL. TAPEPOOL inherits the copy storage pool list
from FILEPOOL2 and simultaneously writes File A to COPYPOOL.

See Figure 34.

FILEPOOL1

Next pool

B
C FILEPOOL2

Migration COPYPOOL1
Next pool C

A A

TAPEPOOL

Figure 34. Inheriting a list of copy storage pools

As a precaution, issue the BACKUP STGPOOL and COPY ACTIVEDATA commands after
the migration operation has completed. Data that is simultaneously written to copy
storage pools or active-data pools during migration is not copied when primary
storage pools are backed up or when active data is copied.

Example of migration without a simultaneous-write operation

In this example, the simultaneous-write function to two copy storage pools is
enabled for client store sessions. Files in the primary storage pool at the top of the
storage hierarchy are migrated. Simultaneous-write operations during migration
are not enabled.

Primary storage pools FILEPOOL and TAPEPOOL are linked to form a storage
hierarchy. FILEPOOL is at the top of the storage hierarchy. TAPEPOOL is the next
pool in the storage hierarchy. Two copy storage pools, COPYPOOL1 and
COPYPOOL2, are defined to FILEPOOL. The value of the AUTOCOPY parameter for
FILEPOOL is CLIENT. The value of the AUTOCOPY parameter for TAPEPOOL is
NONE.
v Files A, B, and C were written to FILEPOOL during client backup operations.
v File C was simultaneously written to COPYPOOL1.
v The files in FILEPOOL are eligible to be migrated.

Chapter 10. Managing storage pools and volumes 349

 The storage pool hierarchy and the copy storage pools associated with FILEPOOL
are displayed in Figure 35.

COPYPOOL2

FILEPOOL A B
C

COPYPOOL1
C
Next pool

TAPEPOOL

Figure 35. Storage pool hierarchy with files eligible to be migrated

When files A, B and C are migrated, they are written to TAPEPOOL. See Figure 36.

COPYPOOL2
FILEPOOL

COPYPOOL1
C

Next pool
A
B
C

TAPEPOOL

Figure 36. Migration after a simultaneous-write operation

As a precaution, issue the BACKUP STGPOOL and COPY ACTIVEDATA commands after
the migration operation has completed. Data that is simultaneously written to copy
storage pools or active-data pools during migration is not copied when primary
storage pools are backed up or when active data is copied.

350 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Example of a simultaneous-write operation during both migration
and a client store operation
In this example, the storage pool hierarchy consists of a two primary storage pools.
A single copy storage pool is defined to the pools. The simultaneous-write function
during client store operations was enabled. However, one of the backed-up files
was not written to the copy storage pool. The simultaneous-write function during
migration is enabled.

Primary storage pools FILEPOOL and TAPEPOOL are linked to form a storage
hierarchy. FILEPOOL is at the top of the storage hierarchy. TAPEPOOL is the next
pool in the storage hierarchy. One copy storage pool, COPYPOOL, is defined to
both FILEPOOL and TAPEPOOL:
v The simultaneous-write function during client store operations was enabled.
(The setting of the AUTOCOPY parameter for FILEPOOL is CLIENT.)
v During client store operations, files A, B, and C were written to COPYPOOL. A
failure occurred while writing file D to COPYPOOL
v The simultaneous-write function during migration is enabled for TAPEPOOL.
(The setting of the AUTOCOPY parameter for TAPEPOOL is MIGRATION.)

The storage pool hierarchy and the copy storage pool that are associated with
FILEPOOL and TAPEPOOL are displayed in Figure 37.

FILEPOOL A B
C D

A B C
Next pool

COPYPOOL1

TAPEPOOL

Figure 37. Storage pool hierarchy with files eligible to be migrated

When files A, B, C, and D are migrated, the follow events occur:

v File D is migrated to TAPEPOOL and simultaneously written to COPYPOOL.
v Files A, B, and C are migrated to TAPEPOOL. They are not written to
COPYPOOL because copies of those files exist in COPYPOOL.

See Figure 38 on page 352.

Chapter 10. Managing storage pools and volumes 351

 FILEPOOL

A B C
Next pool Storage pool
D backup D
A COPYPOOL1
B
C

TAPEPOOL

Figure 38. A simultaneous-write operation during both migration and client backup operations

As a precaution, issue the BACKUP STGPOOL and COPY ACTIVEDATA commands after
the migration operation has completed. Data that is simultaneously written to copy
storage pools or active-data pools during migration is not copied when primary
storage pools are backed up or when active data is copied.

Planning simultaneous-write operations

Before enabling the simultaneous-write function, you must consider available
resources and configuration settings. As a best practice, you can separate data into
discrete storage hierarchies.

Controlling the number of client mount points for

simultaneous-write operations
During simultaneous-write operations, a client session requires a mount point for
each sequential-access storage pool to which data is written. Transactions fail if the
number of mount points that are required for a client session is insufficient.

About this task

Give careful consideration to the number of mount points that are available for a
simultaneous-write operation. A client session requires a mount point to store data
to a sequential-access storage pool. For example, if a storage pool hierarchy
includes a sequential primary storage pool, the client node requires one mount
point for that pool plus one mount point for each copy storage pool and
active-data pool.

Suppose, for example, you create a storage pool hierarchy like the hierarchy shown
in Figure 24 on page 340. DISKPOOL is a random-access storage pool, and
TAPEPOOL, COPYPOOL1, COPYPOOL2, and ACTIVEDATAPOOL are
sequential-access storage pools. For each client backup session, the client might
have to acquire four mount points if it has to write data to TAPEPOOL. To run
two backup sessions concurrently, the client requires a total of eight mount points.

To indicate the number of mount points a client can have, specify a value for the
MAXNUMMP parameter on the REGISTER NODE or UPDATE NODE commands. Verify the
value of the MAXNUMMP parameter and, if necessary, update it if you want to enable
the simultaneous-write function. A value of 3 for the MAXNUMMP parameter might be
sufficient if, during a client session, all the data is stored in DISKPOOL,
COPYPOOL1, COPYPOOL2, and ACTIVEDATAPOOL.

352 IBM Tivoli Storage Manager for Windows: Administrator's Guide

If the number of mount points that are required for a client session exceeds the
value of the client MAXNUMMP parameter, the transaction fails. If the transaction
involves an active-data pool, all the active-data pools are removed from the
active-data pool list for the duration of the client session, and the client tries the
operation again. If the transaction involves a copy storage pool, the setting of the
COPYCONTINUE parameter in the storage pool definition determines whether the
transaction is tried again:
v If the value of the COPYCONTINUE parameter is NO, the client does not try the
operation again.
v If the value of the COPYCONTINUE parameter is YES, all the copy storage pools are
removed from the copy storage pool list for the duration of the client session.
The client tries the operation again.

Restrictions:
v The setting of the COPYCONTINUE parameter does not affect active-data pools. If a
write failure occurs for any of active-data pools, the server stops writing to the
failing active-data pool for the remainder of the session, but continues storing
files into the primary pool and any remaining active-data pools and copy
storage pools. The active-data pool list is active only for the life of the session
and applies to all the primary storage pools in a particular storage pool
hierarchy.
v The setting of the COPYCONTINUE parameter does not affect the simultaneous-write
function during server import. If data is being written simultaneously and a
write failure occurs to the primary storage pool or any copy storage pool, the
server import process fails.
v The setting of the COPYCONTINUE parameter does not affect the simultaneous-write
function during migration. If data is being written simultaneously and a write
failure occurs to any copy storage pool or active-data pool, the failing storage
pool is removed and the data migration process continues. Write failures to the
primary storage pool cause the migration process to fail.

Controlling the number of mount points for a device class during

simultaneous-write operations
If the number of sequential-access volumes that must be mounted for a
simultaneous-write operation exceeds the maximum number of mount points
specified for a device class, the server is not able to acquire the mount points and
the operation fails.

About this task

To specify the maximum number of sequential-access volumes that can be

simultaneously mounted, use the MOUNTLIMIT parameter in the device class
definition.

If the simultaneous-write operation involves an active-data pool, the IBM Tivoli

Storage Manager server tries to remove the active-data pools that use this device
class until enough mount points can be acquired. The transaction fails, and the
client tries the operation again. If sufficient mount points can be acquired when the
operation is tried again, the data is written into the primary storage pool, any
remaining active-data pools, and any copy storage pools, if they exist.

If the operation involves a copy storage pool, the value of the COPYCONTINUE
parameter in the storage pool definition determines whether the client tries the
operation again:

Chapter 10. Managing storage pools and volumes 353

 v If the value of the COPYCONTINUE parameter is NO, the client does not try the
operation again.
v If the value of the COPYCONTINUE parameter is YES, the server tries to remove the
copy storage pools that use this device class until enough mount points can be
acquired. The transaction fails, and the client tries the operation again. If
sufficient mount points can be acquired when the operation is tried again, the
data is written into the primary storage pool, any remaining copy storage pools,
and any active-data pools, if they exist.

Restrictions:
v The setting of the COPYCONTINUE parameter does not affect active-data pools. If a
write failure occurs for any of active-data pools, the server stops writing to the
failing active-data pool for the remainder of the session, but continues storing
files into the primary pool and any remaining active-data pools and copy
storage pools. The active-data pool list is active only for the life of the session
and applies to all the primary storage pools in a particular storage pool
hierarchy.
v The setting of the COPYCONTINUE parameter does not affect the simultaneous-write
function during server import. If data is being written simultaneously and a
write failure occurs to the primary storage pool or any copy storage pool, the
server import process fails.
v The setting of the COPYCONTINUE parameter does not affect the simultaneous-write
function during migration. If data is being written simultaneously and a write
failure occurs to any copy storage pool or active-data pool, the failing storage
pool is removed and the data migration process continues. Write failures to the
primary storage pool cause the migration process to fail.

Storing data without using the simultaneous-write function

Writing data simultaneously to copy storage pools and active-data pools might not
be an efficient solution for every primary storage pool. When simultaneous-write
operations are not practical, use the BACKUP STGPOOL and COPY ACTIVEDATA
commands to store data in copy storage pools and active-data pools.

About this task

Suppose you use a DISK primary storage pool that is accessed by many clients at
the same time during client data-storage operations. If this storage pool is
associated with copy storage pools, active-data pools, or both, the clients might
have to wait until enough tape drives are available to perform the store operation.
In this scenario, simultaneous-write operations could extend the amount of time
required for client store operations. It might be more efficient to store the data in
the primary storage pool and use the BACKUP STGPOOL command to back up the
DISK storage pool to the copy storage pools and the COPY ACTIVEDATA command to
copy active backup data from the DISK storage pool to the active-data pools.

Reducing the potential for switching storage pools during

simultaneous-write operations
Switching primary storage pools can delay the completion of a simultaneous-write
operation. To reduce the potential for switching, ensure that enough space is
available in the primary storage pools and that the pools can accommodate files of
any size.

354 IBM Tivoli Storage Manager for Windows: Administrator's Guide

About this task

Resources such as disk space, tape drives, and tapes are allocated at the beginning
of a simultaneous-write operation, and typically remain allocated during the entire
operation. If, for any reason, the destination primary pool cannot contain the data
being stored, the IBM Tivoli Storage Manager server attempts to store the data into
a next storage pool in the storage hierarchy. This next storage pool typically uses a
sequential-access device class. If new resources must be acquired for the next
storage pool, or the allocated resources must be released because the server has to
wait to acquire the new resources, the client session must wait until the resources
are available.

To reduce the potential for switching storage pools, follow these guidelines:
v Ensure that enough space is available in the primary storage pools that are
targets for the simultaneous-write operation. For example, to make space
available, run the server migration operation before backing up or archiving
client data and before migration operations by Hierarchical Storage Management
(HSM) clients.
v The MAXSIZE parameter on the DEFINE STGPOOL and UPDATE STGPOOL commands
limits the size of the files that the Tivoli Storage Manager server can store in the
primary storage pools during client operations. Honoring the MAXSIZE parameter
for a storage pool during a store operation causes the server to switch pools. To
prevent switching pools, avoid using this parameter if possible.

Separate storage hierarchies for simultaneous-write operations

When using the simultaneous-write function as part of a backup strategy, separate
data into different storage pool hierarchies.

For example, you can configure production servers to store mission critical data in
one storage pool hierarchy and use the simultaneous-write function to back up the
data to copy storage pools and an active-data pool. See Figure 39 on page 356. In
addition, you can configure the servers to store noncritical, workstation data in
another storage pool hierarchy and back up that data using the BACKUP STGPOOL
command.

Chapter 10. Managing storage pools and volumes 355

 Policy Domain
Policy Set
STANDARD Mission Critical
Management Class Management Class
Backup Backup
Copy Copy
Group Group
Points to

Points to
ACTIVEDATAPOOL B
DISKPOOL A DISKPOOL B

COPYPOOL B2

COPYPOOL B1
TAPEPOOL A TAPEPOOL B

Figure 39. Separate storage pool hierarchies for different types of data

Simultaneous-write function as part of a backup strategy:

Example
The simultaneous-write function is used to create on-site backups of a storage pool
for easy availability. The BACKUP STGPOOL command is used to create storage pool
backups and database backups that are moved off-site to provide data protection
in case a disaster occurs.

About this task

This example also shows how to use the COPY ACTIVEDATA command to copy active
data from primary storage pools to an on-site sequential-access disk (FILE)
active-data pool. When designing a backup strategy, carefully consider your own
system, data storage, and disaster-recovery requirements.

Procedure
1. Define the following storage pools:
v Two copy storage pools, ONSITECOPYPOOL and DRCOPYPOOL
v One active-data pool, ACTIVEDATAPOOL
v Two primary storage pools, DISKPOOL and TAPEPOOL
As part of the storage pool definition for DISKPOOL, specify TAPEPOOL as the
next storage pool, ONSITECOPYPOOL as the copy storage pool, and
ACTIVEDATAPOOL as the active-data pool. Set the copy continue parameter
for copy storage pools to YES. If an error occurs writing to a copy storage pool,
the operation will continue storing data into the primary pool, the remaining
copy storage pool, and the active-data pool.
define stgpool tapepool mytapedevice
define stgpool onnsitepool mytapedevice
define stgpool drcopypoool mytapedevice

356 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 define stgpool activedatapool mydiskdevice
define stgpool diskpool mydiskdevice nextstgpool=tapepool
copystgpool=onsitecopypool copycontinue=yes activedatapools=
activedatapool

This basic configuration is like the configuration that is shown in Figure 24 on

page 340.
2. Schedule or issue the following commands to ensure that all the files are
backed up:
backup stgpool diskpool onsitecopypool
backup stgpool tapepool onsitecopypool
copy activedata diskpool activedatapool
copy activedata tapepool activedatapool
3. To create the storage-pool backup volumes that to be moved off-site, schedule
the following two commands to run every night:
backup stgpool diskpool drcopypool
backup stgpool tapepool drcopypool
4. Every night, after the storage pool backups have completed, back up the
database.
5. To process the database and storage pool backups for off-site storage, issue the
following command every night:
move drmedia copystgpool=drcopypool wherestate=mountable tostate=vault wait=yes
6. Start migration of the files in the DISKPOOL to ensure that sufficient space is
available in DISKPOOL in preparation for the next storage operations:
migrate stgpool diskpool

Keeping client files together using collocation

Collocation reduces the number of volume mounts required when users restore,
retrieve, or recall a large number of files from the storage pool. Collocation thus
reduces the amount of time required for these operations.

About this task

With collocation enabled, the server tries to keep files on a minimal number of
sequential-access storage volumes. The files can belong to a single client node, a
group of client nodes, a client file space, or a group of file spaces. You can set
collocation for each sequential-access storage pool when you define or update the
pool.

Figure 40 on page 358 shows an example of collocation by client node with three
clients, each having a separate volume containing that client's data.

Chapter 10. Managing storage pools and volumes 357

 Figure 40. Example of collocation enabled

Figure 41 shows an example of collocation by group of client nodes. Three groups

have been defined, and the data for each group is stored on separate volumes.

Figure 41. Example of collocation enabled by node collocation group

Figure 42 shows an example of collocation by file space group. Six groups have
been defined. Each group contains data from file spaces that belong to a single
node. The data for each group is stored on a separate volume.

1
3
1
2
4 2
3
5
4
5
6 6

Figure 42. Example of collocation enabled by file space collocation group

When collocation is disabled, the server tries to use all available space on each
volume before selecting a new volume. While this process provides better
utilization of individual volumes, user files can become scattered across many
volumes. Figure 43 on page 359 shows an example of collocation disabled, with

358 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 three clients sharing space on single volume.

Figure 43. Example of collocation disabled

With collocation disabled, more media mount operations might be required to

mount volumes when users restore, retrieve, or recall a large number of files.

Collocation by group is the Tivoli Storage Manager system default for primary
sequential-access storage pools. The default for copy storage pools and active-data
pools is no collocation.

The effects of collocation on operations

The effect of collocation on resources and system performance depends on the type
of operation that is being run.

Table 31 summarizes the effects of collocation on operations.

Table 31. Effect of collocation on operations
Operation Collocation Enabled Collocation Disabled
Backing up, archiving, or migrating More media mounts to collocate files. Fewer media mounts are required.
client files
Restoring, retrieving, or recalling Large numbers of files can be Multiple mounts of media might be
client files restored, retrieved, or recalled more required for a single user because
quickly because files are on fewer files might be spread across multiple
volumes. volumes.

More than one user's files can be

stored on the same sequential-access
storage volume. For example, if two
users try to recover a file that is on
the same volume, the second user is
forced to wait until the first user's
files are recovered.
Storing data on tape The server attempts to use all The server attempts to use all
available tape volumes to separate available space on each tape volume
user files before it uses all available before you use another tape volume.
space on every tape volume.

Chapter 10. Managing storage pools and volumes 359

Table 31. Effect of collocation on operations (continued)
Operation Collocation Enabled Collocation Disabled
Media mount operations More mount operations when user More mount operations are required
files are backed up, archived, or during restore, retrieve, and recall of
migrated from client nodes directly to client files.
sequential-access volumes.

More mount operations during

reclamation and storage pool
migration.

More volumes to manage because

volumes are not fully used.
Generating backup sets Less time is spent searching database More time is spent searching database
entries and fewer mount operations. entries and fewer mount operations.

When collocation is enabled for a group, single client node or file space, all the
data that belongs to the group, the node, or the file space is moved or copied by
one server process. For example, if data is collocated by group, all data for all
nodes that belong to the same collocation group is migrated by the same process.

When collocating data, the Tivoli Storage Manager server tries to keep files
together on a minimal number of sequential-access storage volumes. However,
when the server is backing up data to volumes in a sequential-access storage pool,
the backup process has priority over collocation settings. As a result, the server
completes the backup, but might not be able to collocate the data.

For example, suppose that you are collocating by node and you specify that a node
can use two mount points on the server. Suppose also that the data that is backed
up from the node can easily fit on one tape volume. During backup, the server
might mount two tape volumes, and the node's data might be distributed across
two tapes, rather than one. When you used collocation, the following server
operations use one server process:
v Moving data from random-access and sequential-access volumes
v Moving node data from sequential-access volumes
v Backing up a random-access or sequential-access storage pool
v Restoring a sequential-access storage pool
v Reclamation of a sequential-access storage pool or off-site volumes
v Migration from a random-access storage pool.

When migrating data from a random disk storage pool to a sequential storage
pool, and collocation is by node or file space, nodes or file spaces are automatically
selected for migration based on the amount of data to be migrated. The node or
file space with the most data is migrated first. If collocation is by group, all nodes
in the storage pool are evaluated to determine which node has the most data. The
node with the most data is migrated first along with all the data for all the nodes
that belong to that collocation group. This process takes place, regardless of how
much data is stored in the file spaces of nodes and regardless of whether the low
migration threshold was reached.

However, when migrating collocated data from a sequential storage pool to

another sequential storage pool, the server orders the volumes according to the
date when the volume was last accessed. The volume with the earliest access date
is migrated first, and the volume with the latest access date is migrated last.

360 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 One reason to collocate by group is that individual client nodes often do not have
sufficient data to fill high-capacity tape volumes. Collocating data by groups of
nodes can reduce unused tape capacity by putting more collocated data on
individual tapes. Also, collocating data by groups of file spaces reduces the unused
tape to a finer degree.

The data that belongs to all the nodes in the same collocation group are migrated
by the same process. Therefore, collocation by group can reduce the number of
times a volume that is going to be migrated needs to be mounted. Collocation by
group can also minimize database scanning and reduce tape passes during data
transfer from one sequential-access storage pool to another.

How the server selects volumes with collocation enabled

Volume selection depends on whether collocation is by group, by node, or by file
space.

Table 32 shows how the Tivoli Storage Manager server selects the first volume
when collocation is enabled for a storage pool at the client-node, collocation group,
and file-space level.
Table 32. How the server selects volumes when collocation is enabled
Volume Selection When collocation is by group When collocation is by node When collocation is by file
Order space
1 A volume that already A volume that already A volume that already
contains files from the contains files from the same contains files from the same
collocation group to which the client node file space of that client node
client belongs
2 An empty predefined volume An empty predefined volume An empty predefined volume
3 An empty scratch volume An empty scratch volume An empty scratch volume
4 A volume with the most A volume with the most A volume that contains data
available free space among available free space among from the same client node
volumes that already contain volumes that already contain
data data
5 Not applicable Not applicable A volume with the most
available free space among
volumes that already contain
data

When the server must continue to store data on a second volume, it uses the
following selection order to acquire more space:
1. An empty predefined volume
2. An empty scratch volume
3. A volume with the most available free space among volumes that already
contain data
4. Any available volume in the storage pool

When collocation is by client node or file space, the server tries to provide the best
use of individual volumes and minimizes file mixing from different clients or file
spaces on volumes. This configuration is depicted in Figure 44 on page 362, which
shows that volume selection is horizontal, where all available volumes are used
before all available space on each volume is used. A, B, C, and D represent files
from four different client nodes.

Chapter 10. Managing storage pools and volumes 361

 Remember:
1. If collocation is by node and the node has multiple file spaces, the server does
not attempt to collocate those file spaces.
2. If collocation is by file space and a node has multiple file spaces, the server
attempts to put data for different file spaces on different volumes.

Amount
of space D
used on
each A
B D
volume
C
VOL1 VOL2 VOL3 VOL4 VOL5

Numbers of volumes (1 to n)

Figure 44. Using all available sequential access storage volumes with collocation enabled at
the node or file space level

Collocation can be by filespace group or node group. When collocation is by node

group (node collocation group), the server tries to collocate data from nodes that
belong to the same collocation group. A filespace collocation group uses the same
methods as a node collocation group, but can use up more space because of the
granularity of filespace sizes. As shown in Figure 45, data for the following groups
of nodes was collocated:
v Group 1 consists of nodes A, B, and C
v Group 2 consists of nodes D and E
v Group 3 consists of nodes F, G, H, and I
Whenever possible, the Tivoli Storage Manager server collocates data that belongs
to a group of nodes on a single tape, as represented by Group 2 in the figure. Data
for a single node can also be spread across several tapes that are associated with a
group (Group 1 and 2). If the nodes in the collocation group have multiple file
spaces, the server does not attempt to collocate those file spaces.

H
C
Amount
of space E
used on B G I
each
volume
A D H
C F

Numbers of volumes (1 to n)

Figure 45. Using all available sequential access storage volumes with collocation enabled at
the group level

Remember: Normally, the Tivoli Storage Manager server always writes data to the
current filling volume for the operation that is running. Occasionally, however, you
might notice more than one filling volume in a collocated storage pool. Having
more than one filling volume in a collocated storage pool can occur if different

362 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 server processes or client sessions try to store data into the collocated pool at the
same time. In this situation, Tivoli Storage Manager allocates a volume for each
process or session that needs a volume so that both operations complete as quickly
as possible.

How the server selects volumes with collocation disabled

When collocation is disabled, the server attempts to use all available space in a
storage volume before it accesses another volume.

When storing client files in a sequential-access storage pool where collocation is

disabled, the server selects a volume using the following selection order:
1. A previously used sequential volume with available space (a volume with the
most amount of data is selected first)
2. An empty volume

When the server needs to continue to store data on a second volume, it attempts to
select an empty volume. If none exists, the server attempts to select any remaining
available volume in the storage pool.

Figure 46 shows that volume utilization is vertical when collocation is disabled. In

this example, fewer volumes are used because the server attempts to use all
available space by mixing client files on individual volumes. A, B, C, and D
represent files from four different client nodes.

B
D
Amount C
of space C
used on A A
each
volume B D
D
C A
VOL1 VOL2 VOL3 VOL4 VOL5

Numbers of volumes (1 to n)

Figure 46. Using all available space on sequential volumes with collocation disabled

Collocation on or off settings

After you define a storage pool, you can change the collocation setting by
updating the storage pool. The change in collocation for the pool does not affect
files that are already stored in the pool.

For example, if collocation is off for a storage pool and you turn it on, from then on
client files stored in the pool are collocated. Files that had previously been stored
in the pool are not moved to collocate them. As volumes are reclaimed, however,
the data in the pool tends to become more collocated. You can also use the MOVE
DATA or MOVE NODEDATA commands to move data to new volumes to increase
collocation. However, this causes an increase in the processing time and the
volume mount activity.

Remember: A mount wait can occur or increase when collocation by file space is
enabled and a node has a volume containing multiple file spaces. If a volume is
eligible to receive data, Tivoli Storage Manager will wait for that volume.

Chapter 10. Managing storage pools and volumes 363

 Collocation of copy storage pools and active-data pools
Using collocation on copy storage pools and active-data pools requires special
consideration. Collocation of copy storage pools and active-data pools, especially
by node or file space, results in more partially filled volumes and potentially
unnecessary offsite reclamation activity.

Using collocation on copy storage pools and active-data pools requires special
consideration.

Primary storage pools complete a different recovery role than the recovery role
completed by copy storage pools and active-data pools. Normally you use primary
storage pools (or active-data pools) to recover data to clients directly. In a disaster,
when both clients and the server are lost, you might use offsite active-data pool
volumes to recover data directly to clients and the copy storage pool volumes to
recover the primary storage pools. The types of recovery scenarios that concern
you the most helps you to determine whether to use collocation on your copy
storage pools and active-data pools.

Collocation typically results in partially filled volumes when you collocate by node
or by file space. (Partially filled volumes are less prevalent, however, when you
collocate by group.) Partially filled volumes might be acceptable for primary
storage pools because the volumes remain available and can be filled during the
next migration process. However, partially filled volumes might be unacceptable
for copy storage pools and active-data pools whose storage pool volumes are taken
offsite immediately. If you use collocation for copy storage pools or active-data
pools, you must decide among the following:
v Taking more partially filled volumes offsite, which increases the reclamation
activity when the reclamation threshold is lowered or reached. The rate of
reclamation for volumes in an active-data pool is typically faster than the rate
for volumes in other types of storage pools.
v Leaving these partially filled volumes onsite until they fill and risk not having
an offsite copy of the data on these volumes.
v Whether to collocate by group to use as much tape capacity as possible.

With collocation disabled for a copy storage pool or an active-data pool, typically
there will be only a few partially filled volumes after data is backed up to the copy
storage pool or copied to the active-data pool.

Consider your options carefully before you use collocation for copy storage pools
and active-data pools, and whether to use simultaneous write. When not using
simultaneous write, if you use collocation for your primary storage pools, you
might want to disable collocation for copy storage pools and active-data pools.
Collocation on copy storage pools or active-data pools might be desirable if you
have few clients with each of them having large amounts of incremental backup
data each day. For collocation with simultaneous write, you must ensure that the
collocate settings are identical for the primary storage pools, active-data pools, and
copy storage pools.

364 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 Planning for and enabling collocation
Understanding the effects of collocation can help reduce the number of media
mounts, make better use of space on sequential volumes, and improve the
efficiency of server operations.

About this task

Table 33 lists the four collocation options that you can specify on the DEFINE
STGPOOL and UPDATE STGPOOL commands. The table also shows the effects of
collocation on data that belongs to nodes that are and are not members of
collocation groups.
Table 33. Collocation options and the effects on node data
If a node is not defined as a member of a If a node is defined as a member of a
Collocation option collocation group collocation group
No The data for the node is not collocated. The data for the node is not collocated.
Group The server stores the data for the node on as The server stores the data for the node and for
few volumes in the storage pool as possible. other nodes that belong to the same
collocation group on as few volumes as
possible.
Node The server stores the data for the node on as The server stores the data for the node on as
few volumes as possible. few volumes as possible.
Filespace The server stores the data for the node's file The server stores the data for the node's file
space on as few volumes as possible. If a node space on as few volumes as possible. If a node
has multiple file spaces, the server stores the has multiple file spaces, the server stores the
data for different file spaces on different data for different file spaces on different
volumes in the storage pool. volumes in the storage pool.

Table 34. Collocation group options and effects on file space data
If a file space is not defined as a member of If a file space is defined as a member of a
Collocation option a collocation group collocation group
No The data for the file space is not collocated. The data for the file space is not collocated.
Group The server stores the data for the filespace on The server stores the data for the file space
as few volumes in the storage pool as and other file spaces that belong to the same
possible. collocation group on as few volumes as
possible.
Node The server stores the data for the node on as The server stores the data for the node on as
few volumes as possible. few volumes as possible.
Filespace The server stores the data for the node's file The server stores the data for the file spaces on
space on as few volumes as possible. If a node as few volumes as possible. If a node has
has multiple file spaces, the server stores the multiple file spaces, the server stores the data
data for different file spaces on different for different file spaces on different volumes in
volumes in the storage pool. the storage pool.

When deciding whether and how to collocate data, complete the following steps:

Procedure
1. Familiarize yourself with the potential advantages and disadvantages of
collocation, in general. For a summary of effects of collocation on operations,
see Table 31 on page 359.

Chapter 10. Managing storage pools and volumes 365

 2. If the decision is to collocate, determine how the data is to be organized,
whether by client node, group of client nodes, or file space. If the decision is to
collocate by group, you must decide how to group nodes:
v If the goal is to save space, you might want to group small nodes together to
better use tapes.
v If the goal is potentially faster client restores, group nodes together so that
they fill as many tapes as possible. Doing so increases the probability that
individual node data will be distributed across two or more tapes and that
more tapes can be mounted simultaneously during a multi-session No Query
Restore operation.
v If the goal is to departmentalize data, then you can group nodes by
department.
3. If collocation by group is the wanted result:
a. Define collocation groups with the DEFINE COLLOCGROUP command.
b. Add client nodes to the collocation groups with the DEFINE COLLOCMEMBER
command.
The following query commands are available to help in collocating groups:
QUERY COLLOCGROUP
Displays the collocation groups defined on the server.
QUERY NODE
Displays the collocation group, if any, to which a node belongs.
QUERY NODEDATA
Displays information about the data for one or more nodes in a
sequential-access storage pool.
QUERY STGPOOL
Displays information about the location of client data in a
sequential-access storage pool and the amount of space a node occupies
in a volume.
For more information about these commands, see the Administrator's Reference.
You can also use Tivoli Storage Manager server scripts or PERL scripts to
display information that can be useful in defining collocation groups.
4. Specify how data is to be collocated in a storage pool with the COLLOCATE
parameter on the DEFINE STGPOOL or UPDATE STGPOOL command.
5. If you decide later that you want to delete members of a collocation group, you
can use the DELETE COLLOCMEMBER command. You can also update the
description of a collocation group with the UPDATE COLLOCGROUP command and
delete entire collocation groups by issuing the DELETE COLLOCGROUP command.

Results

Tip: If you use collocation, but want to reduce the number of media mounts and
use space on sequential volumes more efficiently, you can:
v Define a storage pool hierarchy and policy to require that backed-up, archived,
or space-managed files are stored initially in disk storage pools.
When files are migrated from a disk storage pool, the server attempts to migrate
all files that belong to the client node or collocation group that is using the most
disk space in the storage pool. This process works well with the collocation
option because the server tries to place all of the files from a particular client on
the same sequential-access storage volume.

366 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 v Use scratch volumes for sequential-access storage pools to allow the server to
select new volumes for collocation.
v Specify the client option COLLOCATEBYFILESPEC to limit the number of tapes to
which objects associated with one file specification are written. This collocation
option makes collocation by the server more efficient; it does not override
collocation by file space or collocation by node.

For details about the COLLOCATEBYFILESPEC option, see the Backup-Archive Clients
Installation and User's Guide.

When creating collocation groups, keep in mind that the ultimate destination of the
data that belongs to nodes in a collocation group depends on the policy domain to
which nodes belong. For example, suppose that you create a collocation group that
consists of nodes that belong to Policy Domain A. Policy Domain A specifies an
active-data pool as the destination of active data only and has a backup copy
group that specifies a primary storage pool, Primary1, as the destination for active
and inactive data. Other nodes in the same collocation group belong to a domain,
Policy Domain B, that does not specify an active-data pool, but that has a backup
copy group that specifies Primary1 as the destination for active and inactive data.
Primary1 has a designated copy storage pool. The collocation setting on
PRIMARY1, the copy storage pool, and the active-data pool is GROUP.

When the node data is backed up and a simultaneous-write operation occurs,

active and inactive data is stored in Primary1 and the copy storage pool. Note,
however, that although all the nodes belong to a single collocation group, only the
active data which belongs to nodes in Domain A are stored in the active-data pool.
The data in Primary1 and the copy storage pool is collocated by group. The data in
the active-data pool is also collocated by group, but the "group" consists only of
nodes that are members of Policy Domain A.
Related reference:
“Managing client option files” on page 464

Reclaiming space in sequential-access storage pools

Space on a sequential-access storage volume becomes reclaimable as files expire or
are deleted from the volume. Reclamation processing involves consolidating the
remaining data from many sequential-access volumes onto fewer new
sequential-access volumes.

About this task

Files become obsolete because of aging or limits on the number of versions of a

file. Space in volumes in active-data pools also becomes reclaimable as updated
files are added to the pools and as older file versions are deactivated. In
reclamation processing, the server rewrites files on the volume being reclaimed to
other volumes in the storage pool, making the reclaimed volume available for
reuse.

The server reclaims the space in storage pools based on a reclamation threshold
that you can set for each sequential-access storage pool. When the percentage of
space that can be reclaimed on a volume rises above the reclamation threshold, the
server reclaims the volume.

Restrictions:

Chapter 10. Managing storage pools and volumes 367

 v Storage pools that are defined with the NETAPPDUMP, the CELERRADUMP or
the NDMPDUMP data format cannot be reclaimed. However, you can use the
MOVE DATA command to move data out of a volume so that the volume can be
reused. The volumes in the target storage pool must have the same data format
as the volumes in the source storage pool.
v Storage pools that are defined with a CENTERA device class cannot be
reclaimed.

How Tivoli Storage Manager reclamation works

You can set a reclamation threshold for a sequential-access storage pool when you
define or update the pool. When the percentage of reclaimable space on a volume
exceeds the reclamation threshold set for the storage pool, the volume is eligible
for reclamation.

The server checks whether reclamation is needed at least once per hour and begins
space reclamation for eligible volumes. During space reclamation, the server copies
files that remain on eligible volumes to other volumes. For example, Figure 47 on
page 369 shows that the server consolidates the files from tapes 1, 2, and 3 on tape
4. During reclamation, the server copies the files to volumes in the same storage
pool unless you have specified a reclamation storage pool. Use a reclamation
storage pool to allow automatic reclamation for a storage pool with only one drive.

Remember: To prevent contention for the same tapes, the server does not allow a
reclamation process to start if a DELETE FILESPACE process is active. The server
checks every hour for whether the DELETE FILESPACE process has completed so
that the reclamation process can start. After the DELETE FILESPACE process has
completed, reclamation begins within one hour.

The server also reclaims space within an aggregate. An aggregate is a physical file
that contains multiple logical files that are backed up or archived from a client in a
single transaction. Space within the aggregate becomes reclaimable space as logical
files in the aggregate expire, as files are deleted by the client, or as files become
deactivated in active-data pools. The server removes unused space as the server
copies the aggregate to another volume during reclamation processing. However,
reclamation does not aggregate files that were originally stored in non-aggregated
form. Reclamation also does not combine aggregates to make new aggregates. You
can also reclaim space in an aggregate by issuing the MOVE DATA command. See
“Reclaiming space in aggregates by moving data” on page 405 for details.

368 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 4

= valid data

Figure 47. Tape reclamation

After the server moves all readable files to other volumes, one of the following
occurs for the reclaimed volume:
v If you have explicitly defined the volume to the storage pool, the volume
becomes available for reuse by that storage pool.
v If the server acquired the volume as a scratch volume, the server deletes the
volume from the Tivoli Storage Manager database.

Volumes that have a device type of SERVER are reclaimed in the same way as
other sequential-access volumes. However, because the volumes are actually data
stored in the storage of another Tivoli Storage Manager server, the reclamation
process can consume network resources. See “Controlling reclamation of virtual
volumes” on page 374 for details about how the server reclaims these types of
volumes.

Volumes in a copy storage pool and active-data pools are reclaimed in the same
manner as a primary storage pool except for the following:
v Off-site volumes are handled differently.
v The server copies active files from the candidate volume only to other volumes
in the same storage pool.
For details, see “Reclaiming copy storage pools and active-data pools” on page 374.

Chapter 10. Managing storage pools and volumes 369

 Reclamation thresholds
Space is reclaimable because it is occupied by files that have been expired or
deleted from the Tivoli Storage Manager database, or because the space has never
been used. The reclamation threshold indicates how much reclaimable space a
volume must have before the server reclaims the volume.

The server checks whether reclamation is needed at least once per hour. The lower
the reclamation threshold, the more frequently the server tries to reclaim space.
Frequent reclamation optimizes the use of a sequential-access storage pool’s space,
but can interfere with other processes, such as backups from clients.

If the reclamation threshold is high, reclamation occurs less frequently. A high

reclamation threshold is useful if mounting a volume is a manual operation and
the operations staff is at a minimum. Setting the reclamation threshold to 100%
prevents automatic reclamation from occurring. You might want to do this to
control when reclamation occurs, to prevent interfering with other server processes.
When it is convenient for you and your users, you can use the RECLAIM
STGPOOL command to invoke reclamation, or you can lower the reclamation
threshold to cause reclamation to begin.

If you set the reclamation threshold to 50% or greater, the server can combine the
usable files from two or more volumes onto a single new volume.

Reclamation of volumes in an active-data pool usually returns volumes to scratch

status more frequently than reclamation of volumes in non-active-data pools. This
is because the percentage of reclaimable space for sequential volumes in
active-data pools reflects not only the space of deleted files, but also the space of
inactive files. Frequent reclamation requires more resources such as tape drives and
libraries to mount and dismount volumes.

If reclamation is occurring too frequently in your active-data pools, you can

increase the reclamation thresholds until the rate of reclamation is acceptable.
Accelerated reclamation of volumes has more of an effect on active-data pools that
use removable media and, in particular, on removable media that is taken off-site.

Reclaiming volumes with the most reclaimable space

If you have been running with a high reclamation threshold and decide you need
to reclaim volumes, you can lower the threshold in several steps. Lowering the
threshold in steps ensures that volumes with the most reclaimable space are
reclaimed first.

About this task

For example, if you set the reclamation threshold to 100%, first lower the threshold
to 98%. Volumes that have reclaimable space of 98% or greater are reclaimed by
the server. Lower the threshold again to reclaim more volumes.

If you lower the reclamation threshold while a reclamation process is active, the
reclamation process does not immediately stop. If an on-site volume is being
reclaimed, the server uses the new threshold setting when the process begins to
reclaim the next volume. If off-site volumes are being reclaimed, the server does
not use the new threshold setting during the process that is running (because all
eligible off-site volumes are reclaimed at the same time).

Use the CANCEL PROCESS command to stop a reclamation process.

370 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Starting reclamation manually or in a schedule
To gain more control over how and when the reclamation process occurs, you can
use the RECLAIM STGPOOL command. You can also specify the maximum amount of
time a reclamation process takes before it is automatically canceled.

About this task

To perform reclamation when it is least intrusive to normal production needs,

include the RECLAIM STGPOOL command in a schedule. For example, to start
reclamation in a storage pool named ALTPOOL, and to have reclamation end as soon
as possible after one hour, you would issue the following command:
reclaim stgpool altpool duration=60

For copy storage pools and active-data pools, you can also use the RECLAIM
STGPOOL command to specify the maximum number of offsite storage pool volumes
the server should attempt to reclaim:
reclaim stgpool altpool duration=60 offsitereclaimlimit=230

Do not use this command if you are going to use automatic reclamation for the
storage pool. To prevent automatic reclamation from running, set the RECLAIM
parameter of the storage pool definition to 100.

For details about the RECLAIM STGPOOL command, refer to the Administrator's
Reference.

Restriction: Storage pools that are defined with a CENTERA device class cannot
be reclaimed.

Optimizing drive usage using multiple concurrent reclamation

processes
Multiple reclamation processes run concurrently, allowing you to make better use
of your available tape drives or FILE volumes.

About this task

You can specify one or more reclamation processes for each primary
sequential-access storage pool, copy storage pool, or active-data pool using the
RECLAIMPROCESS parameter on the DEFINE STGPOOL and UPDATE STGPOOL
commands.

Each reclamation process requires at least two simultaneous volume mounts (at
least two mount points) and, if the device type is not FILE, at least two drives.
One of the drives is for the input volume in the storage pool being reclaimed. The
other drive is for the output volume in the storage pool to which files are being
moved.

When calculating the number of concurrent processes to run, you must carefully
consider the resources you have available, including the number of storage pools
that will be involved with the reclamation, the number of mount points, the
number of drives that can be dedicated to the operation, and (if appropriate) the
number of mount operators available to manage reclamation requests. The number
of available mount points and drives depends on other Tivoli Storage Manager and
system activity and on the mount limits of the device classes for the storage pools
that are involved in the reclamation.

Chapter 10. Managing storage pools and volumes 371

 For more information about mount limit, see: “Controlling the number of
simultaneously mounted volumes” on page 181

For example, suppose that you want to reclaim the volumes from two sequential
storage pools simultaneously and that all storage pools involved have the same
device class. Each process requires two mount points and, if the device type is not
FILE, two drives. To run four reclamation processes simultaneously (two for each
storage pool), you need a total of at least eight mount points and eight drives. The
device class for each storage pool must have a mount limit of at least eight.

If the device class for the storage pools being reclaimed does not have enough
mount points or drives, you can use the RECLAIMSTGPOOL parameter to direct
the reclamation to a storage pool with a different device class that has the
additional mount points or drives.

If the number of reclamation processes you specify is more than the number of
available mount points or drives, the processes that do not obtain mount points or
drives will wait indefinitely or until the other reclamation processes complete and
mount points or drives become available.

The Tivoli Storage Manager server will start the specified number of reclamation
processes regardless of the number of volumes that are eligible for reclamation. For
example, if you specify ten reclamation processes and only six volumes are eligible
for reclamation, the server will start ten processes and four of them will complete
without processing a volume.

Multiple concurrent reclamation processing does not affect collocation. For

additional information, see “How collocation affects reclamation” on page 378.

Reclaiming volumes in a storage pool with one drive

When a storage pool has only one mount point (that is, just one drive) available to
it through the device class, data cannot be reclaimed from one volume to another
within that same storage pool. To reclaim volumes in a storage pool that has only
one drive, you can define a reclamation storage pool and use it for temporary storage
of reclaimed data.

About this task

When the server reclaims volumes, the server moves the data from volumes in the
original storage pool to volumes in the reclamation storage pool. The server always
uses the reclamation storage pool when one is defined, even when the mount limit
is greater than one.

If the reclamation storage pool does not have enough space to hold all of the data
being reclaimed, the server moves as much of the data as possible into the
reclamation storage pool. Any data that could not be moved to volumes in the
reclamation storage pool still remains on volumes in the original storage pool.

The pool identified as the reclamation storage pool must be a primary sequential
storage pool. The primary purpose of the reclamation storage pool is for temporary
storage of reclaimed data. To ensure that data moved to the reclamation storage
pool eventually moves back into the original storage pool, specify the original
storage pool as the next pool in the storage hierarchy for the reclamation storage

372 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 pool. For example, if you have a tape library with one drive, you can define a
storage pool to be used for reclamation using a device class with a device type of
FILE:
define stgpool reclaimpool fileclass maxscratch=100

Define the storage pool for the tape drive as follows:

define stgpool tapepool1 tapeclass maxscratch=100
reclaimstgpool=reclaimpool

Finally, update the reclamation storage pool so that data migrates back to the tape
storage pool:
update stgpool reclaimpool nextstgpool=tapepool1

Tip:
v In a mixed-media library, reclaiming volumes in a storage pool defined with a
device class with a single mount point (that is, a single drive) requires one of the
following:
– At least one other drive with a compatible read/write format
– Enough disk space to create a storage pool with a device type of FILE

Reducing the time to reclaim tape volumes with high capacity

When a storage pool uses tape volumes with high capacity, reclamation processes
might run for a long time if the drives are relatively slow at positioning tapes.
There are steps that you can take to reduce overall process time.

About this task

To help reduce overall process time:

Procedure
1. Set up the storage pool hierarchy so that the tape storage pool is the next
storage pool for a storage pool that uses either a DISK device type or a FILE
device type.
2. When you need to reclaim volumes, move data from the tape storage pool to
the DISK or FILE storage pool.
3. Allow the data to migrate from the DISK or FILE storage pool back to the tape
storage pool by adjusting the migration thresholds.

Reclamation of write-once, read-many (WORM) media

Reclamation of WORM volumes does not mean that you can reuse this write-once
media. However, reclamation does allow you to make more library space available.

Reclamation of WORM volumes consolidates data from partially filled volumes to

other WORM volumes. You can then eject the empty, used WORM volumes and
add new volumes.

To allow reclamation, you can set the reclamation value to something lower when
you define or update the storage pool.

Chapter 10. Managing storage pools and volumes 373

 Controlling reclamation of virtual volumes
When virtual volumes (volumes with the device type of SERVER) in a primary
storage pool are reclaimed, the client data stored on those volumes is sent across
the network between the source server and the target server. As a result, the
reclamation process can tie up your network resources.

About this task

To control when reclamation starts for these volumes, consider setting the
reclamation threshold to 100% for any primary storage pool that uses virtual
volumes. Lower the reclamation threshold at a time when your network is less
busy, so that the server can reclaim volumes.

For virtual volumes in a copy storage pool or an active-data pool, the server
reclaims a volume as follows:

Procedure
1. The source server determines which files on the volume are still valid.
2. The source server obtains these valid files from volumes in a primary storage
pool, or if necessary, from removable-media volumes in an on-site copy storage
pool or in an on-site active-data pool. The server can also obtain files from
virtual volumes in a copy storage pool or an active-data pool.
3. The source server writes the files to one or more new virtual volumes in the
copy storage pool or active-data pool and updates its database.
4. The server issues a message indicating that the volume was reclaimed.

Results

For information about using the SERVER device type, see “Using virtual volumes
to store data on another server” on page 732.

Reclaiming copy storage pools and active-data pools

On-site and off-site volumes in copy storage pools and active-data pools are
reclaimed when the amount of unused space exceeds the reclamation threshold.
When reclamation occurs and how reclamation processing is done depends on
whether the volumes are marked as off-site.

About this task

Reclamation of volumes in copy storage pools and active-data pools is similar to

reclamation in primary storage pools. For volumes that are on-site, reclamation
usually occurs after the volume is full and then begins to empty because of file
deletion, expiration, or, in the case of active-data pools, deactivation. When the
percentage of reclaimable space on a volume rises above the reclamation threshold,
the server reclaims the volume. Active files on the volume are rewritten to other
volumes in the storage pool, making the original volume available for new files.

For off-site volumes, reclamation can occur when the percentage of unused space
on the volume is greater than the reclaim parameter value. The unused space in
copy storage pool volumes includes both space that has never been used on the
volume and space that has become empty because of file deletion or expiration.
For volumes in active-data pools, reclaimable space also includes inactive versions
of files. Most volumes in copy storage pools and active-data pools might be set to
an access mode of off-site, making them ineligible to be mounted. During

374 IBM Tivoli Storage Manager for Windows: Administrator's Guide

reclamation, the server copies valid files on off-site volumes from the original files
in the primary storage pools. In this way, the server copies valid files on off-site
volumes without having to mount these volumes. For more information, see
“Reclamation of offsite volumes.”

Reclamation of copy storage pool volumes and active-data pool volumes should be
done periodically to allow the reuse of partially filled volumes that are off-site.
Reclamation can be done automatically by setting the reclamation threshold for the
copy storage pool or the active-data pool to less than 100%. However, you need to
consider controlling when reclamation occurs because of how off-site volumes are
treated. For more information, see “Controlling when reclamation occurs for
off-site volumes” on page 376.

Virtual Volumes: Virtual volumes (volumes that are stored on another Tivoli
Storage Manager server through the use of a device type of SERVER) cannot be set
to the off-site access mode.

Using the RECLAIMPROCESS parameter on the DEFINE STGPOOL or UPDATE

STGPOOL command, you can specify multiple concurrent reclamation processes
for a single copy storage pool or active-data pool. Doing so will let you make
better use of your available tape drives or FILE volumes. The principles underlying
multiple concurrent reclamation processes for copy storage pools and active-data
pools are the same principles as those for primary sequential-access storage pools.
In particular, you need to carefully consider available resources (for example, the
number of mount points) when calculating how many processes you can run
concurrently. For details, see “Optimizing drive usage using multiple concurrent
reclamation processes” on page 371.

Reclamation of primary storage pool volumes does not affect copy storage pool
files or files in active-data pools.

Reclamation of offsite volumes

Volumes with the access value of offsite are eligible for reclamation if the amount
of empty space on a volume exceeds the reclamation threshold for the copy storage
pool or active-data pool. The default reclamation threshold for copy storage pools
and active-data pools is 100%, which means that reclamation is not performed.

When an offsite volume is reclaimed, the files on the volume are rewritten to a
read/write volume. Effectively, these files are moved back to the onsite location.
The files might be obtained from the offsite volume after a disaster, if the volume
has not been reused and the database backup that you use for recovery references
the files on the offsite volume.

The server reclaims an offsite volume as follows:

1. The server determines which files on the volume are still valid.
2. The server obtains these valid files from a primary storage pool or, if necessary,
from an onsite volume of a copy storage pool.
3. The server writes the files to one or more volumes in the copy storage pool or
active-data pool and then updates the database. If a file is an aggregate with
unused space, the unused space is removed during this process.
4. A message is issued indicating that the offsite volume was reclaimed.
During offsite reclamation, active-data pools cannot be used to obtain data.
For a single storage pool, the server reclaims all offsite volumes that are eligible
for reclamation at the same time. Reclaiming all the eligible volumes at the
same time minimizes the tape mounts for primary storage pool volumes.
Chapter 10. Managing storage pools and volumes 375
 If you are using the disaster recovery manager, see “Moving copy storage pool and
active-data pool volumes on-site” on page 1028.

Controlling when reclamation occurs for off-site volumes

If you send copy storage pool volumes off-site, you can control reclamation by
adjusting the reclamation threshold.

About this task

Suppose you plan to make daily storage pool backups to a copy storage pool, then
mark all new volumes in the copy storage pool as offsite and send them to the
off-site storage location. This strategy works well with one consideration if you are
using automatic reclamation (the reclamation threshold is less than 100%).

Each day's storage pool backups will create a number of new copy-storage pool
volumes, the last one being only partially filled. If the percentage of empty space
on this partially filled volume is higher than the reclaim percentage, this volume
becomes eligible for reclamation as soon as you mark it off-site. The reclamation
process would cause a new volume to be created with the same files on it. The
volume you take off-site would then be empty according to the Tivoli Storage
Manager database. If you do not recognize what is happening, you could
perpetuate this process by marking the new partially filled volume off-site.

One way to resolve this situation is to keep partially filled volumes on-site until
they fill up. However, this would mean a small amount of your data would be
without an off-site copy for another day.

If you send copy storage pool volumes off-site, it is recommended you control pool
reclamation by using the default value of 100. This turns reclamation off for the
copy storage pool. You can start reclamation processing at desired times by
changing the reclamation threshold for the storage pool. To monitor off-site volume
utilization and help you decide what reclamation threshold to use, enter the
following command:
query volume * access=offsite format=detailed

Depending on your data expiration patterns, you may not need to do reclamation
of off-site volumes each day. You may choose to perform off-site reclamation on a
less frequent basis. For example, suppose you ship copy-storage pool volumes to
and from your off-site storage location once a week. You can run reclamation for
the copy-storage pool weekly, so that as off-site volumes become empty they are
sent back for reuse.

When you do perform reclamation for off-site volumes, the following sequence is
recommended:

Procedure
1. Back up your primary-storage pools to copy-storage pools or copy the active
data in primary-storage pools to active-data pools.
2. Turn on reclamation for copy-storage pools and active-data pools by lowering
the reclamation threshold for copy-storage pools below 100%. The default for
active-data pools is 60.
3. When reclamation processing completes, turn off reclamation by raising the
reclamation thresholds to 100%.
4. Mark any newly created copy-storage pool volumes and active-data pool
volumes as off-site, and then move them to the off-site location.

376 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Results

This sequence ensures that the files on the new copy-storage pool volumes and
active-data pool volumes are sent off-site, and are not inadvertently kept on-site
because of reclamation.

Preventing off-site marking of partially-filled copy storage pool and active-data

pool volumes:

To prevent marking partially-filled copy storage pool or active-data pool volumes

as off-site, you can use storage on another Tivoli Storage Manager server (device
type of SERVER) for storage-pool backups.

About this task

If the other server is at a different site, the copy-storage pool volumes or

active-data pool volumes are already off-site, with no moving of physical volumes
between the sites. See “Using virtual volumes to store data on another server” on
page 732 for more information.

Limiting the number of off-site volumes to be reclaimed

To ensure that reclamation completes within the desired amount of time, you can
use OFFSITERECLAIMLIMIT parameter on the DEFINE STGPOOL or UPDATE STGPOOL
command to limit the number of off-site volumes to be reclaimed.

About this task

When determining the value for the OFFSITERECLAIMLIMIT parameter, consider

using the statistical information in the message issued at the end of the off-site
volume reclamation operation.

Alternatively, you can use the following Tivoli Storage Manager SQL SELECT
command to obtain records from the SUMMARY table for the off-site volume
reclamation operation:
select * from summary where activity=’OFFSITE RECLAMATION’

Two kinds of records are displayed for the off-site reclamation process. One
volume record is displayed for each reclaimed off-site volume. However, the
volume record does not display the following items:
v The number of examined files.
v The number of affected files.
v The total bytes involved in the operation.
This information is summarized in the statistical summary record for the offsite
reclamation. The statistical summary record displays the following items:
v The number of examined files.
v The number of affected files.
v The total bytes involved in the operation.
v The number of off-site volumes that were processed.
v The number of parallel processes that were used.
v The total amount of time required for the processing.

The order in which off-site volumes are reclaimed is based on the amount of
unused space in a volume. (Unused space includes both space that has never been

Chapter 10. Managing storage pools and volumes 377

 used on the volume and space that has become empty because of file deletion.)
Volumes with the largest amount of unused space are reclaimed first.

For example, suppose a copy storage pool contains three volumes: VOL1, VOL2,
and VOL3. VOL1 has the largest amount of unused space, and VOL3 has the least
amount of unused space. Suppose further that the percentage of unused space in
each of the three volumes is greater than the value of the RECLAIM parameter. If
you do not specify a value for the OFFSITERECLAIMLIMIT parameter, all three
volumes will be reclaimed when the reclamation runs. If you specify a value of 2,
only VOL1 and VOL2 will be reclaimed when the reclamation runs. If you specify
a value of 1, only VOL1 will be reclaimed.

Delayed reuse of reclaimed volumes

Delaying reuse may help you to recover data under certain conditions during
recovery from a disaster.

As a best practice, delay the reuse of any reclaimed volumes in copy storage pools
and active-data pools for as long as you keep your oldest database backup. For
more information about delaying volume reuse, see “Delaying reuse of volumes
for recovery purposes” on page 886.

Reclamation of volumes in active-data pools

Inactive files in volumes in an active-data pool are deleted by reclamation
processing. The rate at which reclaimable space accumulates in active-data pool
volumes is typically faster than the rate for volumes in non-active-data pools.

If reclamation of volumes in an active-data pool is occurring too frequently,

requiring extra resources such as tape drives and libraries to mount and dismount
volumes, you can adjust the reclamation threshold until the rate of reclamation is
acceptable. The default reclamation threshold for active-data pools is 60 percent,
which means that reclamation begins when the storage pool reaches 60 percent of
capacity. Accelerated reclamation of volumes has more of an effect on active-data
pools that use removable media and, in particular, on removable media that is
taken off-site.

How collocation affects reclamation

If collocation is enabled and reclamation occurs, the server tries to move the files
for each client node, group of client nodes or client file space onto a minimal
number of volumes.

If the volumes are manually mounted, the mount operators must:

v Be aware that a tape volume may be rewound more than once if the server
completes a separate pass to move the data for each client node or client file
space.
v Mount and dismount multiple volumes to allow the server to select the most
appropriate volume on which to move data for each client node or client file
space. The server tries to select a volume in the following order:
1. A volume that already contains files belonging to the client file space or
client node
2. An empty volume
3. The volume with the most available space
4. Any available volume

378 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 If collocation is disabled and reclamation occurs, the server tries to move usable
data to new volumes by using the following volume selection criteria, in the order
shown:
1. The volume that contains the most data
2. Any partially full volume
3. An empty predefined volume
4. An empty scratch volume

If you specify collocation and multiple concurrent processes, the server attempts to
move the files for each collocation group, client node, or client file space onto as
few volumes as possible. However, if files belonging to a single collocation group
(or node or file space) are on different volumes to begin with and are being moved
at the same time by different processes, the files could be moved to separate
output volumes. For details about multiple concurrent reclamation processing, see
“Optimizing drive usage using multiple concurrent reclamation processes” on page
371.

See also “Reducing the time to reclaim tape volumes with high capacity” on page
373.

Estimating space needs for storage pools

Three default random-access disk storage pools are provided at installation. You
can add space to these storage pools by adding volumes, or you can define
additional storage pools.

About this task

The following default random-access disk storage pools are available at

installation:
v BACKUPPOOL for backed-up files
v ARCHIVEPOOL for archived files
v SPACEMGPOOL for files migrated from client nodes (space-managed files)

As your storage environment grows, you may want to consider how policy and
storage pool definitions affect where workstation files are stored. Then you can
define and maintain multiple storage pools in a hierarchy that allows you to
control storage costs by using sequential-access storage pools in addition to disk
storage pools, and still provide appropriate levels of service to users.

To help you determine how to adjust your policies and storage pools, get
information about how much storage is being used (by client node) and for what
purposes in your existing storage pools. For more information on how to do this,
see “Obtaining information about the use of storage space” on page 396.

Chapter 10. Managing storage pools and volumes 379

 Estimating space requirments in random-access storage
pools
The amount of storage space required for each random-access disk storage pool is
based on your storage needs for backup, archive, and space-management
operations.

About this task

To estimate the amount of storage space required for each random-access disk
storage pool:
v Determine the amount of disk space needed for different purposes:
– For backup storage pools, provide enough disk space to support efficient
daily incremental backups.
– For archive storage pools, provide sufficient space for a user to archive a
moderate size file system without causing migration from the disk storage
pool to occur.
– For storage pools for space-managed files, provide enough disk space to
support the daily space-management load from HSM clients, without causing
migration from the disk storage pool to occur.
v Decide what percentage of this data you want to keep on disk storage space.
Establish migration thresholds to have the server automatically migrate the
remainder of the data to less expensive storage media in sequential-access
storage pools.
See “Migration thresholds” on page 276 for recommendations on setting
migration thresholds.

Estimating space for backed-up files in random-access storage

pools
Space requirements for backed-up files stored in a single random-access storage
pool are based on the total number of workstations, the average data capacity of a
workstation, the fraction of each workstation disk space used, and the number
backup versions you will keep.

About this task

To estimate the total amount of space needed for all backed-up files stored in a
single random-access (disk) storage pool, use the following formula:
Backup space = WkstSize * Utilization * VersionExpansion * NumWkst

where:
Backup Space
The total amount of storage pool disk space needed.
WkstSize
The average data storage capacity of a workstation. For example, if the
typical workstation at your installation has a 4 GB hard drive, then the
average workstation storage capacity is 4 GB.
Utilization
An estimate of the fraction of each workstation disk space used, in the
range 0 to 1. For example, if you expect that disks on workstations are 75%
full, then use 0.75.
VersionExpansion
An expansion factor (greater than 1) that takes into account the additional
380 IBM Tivoli Storage Manager for Windows: Administrator's Guide
 backup versions, as defined in the copy group. A rough estimate allows 5%
additional files for each backup copy. For example, for a version limit of 2,
use 1.05, and for a version limit of 3, use 1.10.
NumWkst
The estimated total number of workstations that the server supports.

If clients use compression, the amount of space required may be less than the
amount calculated, depending on whether the data is compressible.

Estimating space for archived files in random-access storage

pools
The number of archived files generated by users is not necessarily related to the
amount of data stored on their workstations. To estimate the total amount of space
needed for all archived files in a single random-access (disk) storage pool,
determine what percentage of user files are typically archived.

About this task

Work with policy administrators to calculate this percentage based on the number
and type of archive copy groups defined. For example, if policy administrators
have defined archive copy groups for only half of the policy domains in your
enterprise, then estimate that you need less than 50% of the amount of space you
have defined for backed-up files.

Because additional storage space can be added at any time, you can start with a
modest amount of storage space and increase the space by adding storage volumes
to the archive storage pool, as required.

Estimating space needs in sequential-access storage pools

Estimating the space needs in sequential-access storage pools is a relatively
complex calculation based upon multiple considerations.

About this task

To estimate the amount of space required for sequential-access storage pools,

consider:
v The amount of data being migrated from disk storage pools
v The length of time backed-up files are retained, as defined in backup copy
groups
v The length of time archived files are retained, as defined in archive copy groups
v How frequently you reclaim unused space on sequential volumes
See “Reclaiming space in sequential-access storage pools” on page 367 for
information about setting a reclamation threshold.
v Whether or not you use collocation to reduce the number of volume mounts
required when restoring or retrieving large numbers of files from sequential
volumes
If you use collocation, you may need additional tape drives and volumes.
See “Keeping client files together using collocation” on page 357 for information
about using collocation for your storage pools.
v The type of storage devices and sequential volumes supported at your
installation

Chapter 10. Managing storage pools and volumes 381

Monitoring storage-pool and volume usage
Monitor your storage pools and volumes to determine space requirements, the
status of data migration from one to storage pool to the next storage pool in the
storage hierarchy, and the use of disk space by cached copies of files that have
been migrated to the next storage pool.

Monitoring space available in a storage pool

Monitoring the space available in storage pools is important to ensure that client
operations such as backup can complete successfully. To make more space
available, you might need to define more volumes for disk storage pools, or add
more volumes for sequential-access storage pools such as tape.

For more information about maintaining a supply of volumes in libraries, see:

“Managing volumes” on page 128

Obtaining capacity estimates and utilization percentages of

storage pools
Standard reports about storage pools list basic information, such as the estimated
capacity and utilization percentage of all storage pools that are defined to the
system.

About this task

To obtain a standard report, issue the following command:

query stgpool

Figure 48 shows a standard report with all storage pools defined to the system. To
monitor the use of storage pool space, review the Estimated Capacity and Pct Util
columns.

Storage Device Estimated Pct Pct High Low Next

Pool Name Class Name Capacity Util Migr Mig Mig Storage
Pct Pct Pool
----------- ---------- ---------- ----- ----- ---- ---- -----------
ARCHIVEPOOL DISK 0.0 M 0.0 0.0 90 70
BACKTAPE TAPE 180.0 M 85.0 100.0 90 70
BACKUPPOOL DISK 80.0 M 51.6 51.6 50 30 BACKTAPE
COPYPOOL TAPE 300.0 M 42.0
ENGBACK1 DISK 0.0 M 0.0 0.0 85 40 BACKTAPE

Figure 48. Information about storage pools

Estimated Capacity
Specifies the space available in the storage pool in megabytes (M) or
gigabytes (G).
For a disk storage pool, this value reflects the total amount of available
space in the storage pool, including any volumes that are varied offline.
For sequential-access storage pools, estimated capacity is the total
estimated space of all the sequential-access volumes in the storage pool,
regardless of their access mode. At least one volume must be used in a
sequential-access storage pool (either a scratch volume or a private
volume) to calculate estimated capacity.
For tape and FILE, the estimated capacity for the storage pool includes the
following factors:

382 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 v The capacity of all the scratch volumes that the storage pool already
acquired or can acquire. The number of scratch volumes is defined by
the MAXSCRATCH parameter on the DEFINE STGPOOL or UPDATE STGPOOL
command.
v The capacity of all the private volumes that are defined to the storage
pool when using the DEFINE VOLUME command.
The calculations for estimated capacity depend on the available space of
the storage for the device that is assigned to the storage pool. For FILE
storage pools, the capacity for the storage pool is reduced if the available
storage falls below the total estimated space of all the FILE volumes in the
storage pool. The value that is displayed for capacity is reduced by the size
of a FILE volume incrementally as the available space continues to decline.
Tape volumes in a sequential-access storage pool, unlike those in a disk
storage pool, do not contain a precisely known amount of space. Data is
written to a tape volume as necessary until the end of the volume is
reached. For this reason, the estimated capacity is truly an estimate of the
amount of available space in a sequential-access storage pool. This
characteristic does not apply to FILE volumes in sequential-access storage
pools.
Pct Util
Specifies, as a percentage, the space that is used in each storage pool.
For disk storage pools, this value reflects the total number of disk blocks
currently allocated by Tivoli Storage Manager. Space is allocated for
backed-up, archived, or space-managed files that are eligible for server
migration, cached files that are copies of server-migrated files, and files
that reside on any volumes that are varied offline.

Note: The value for Pct Util can be higher than the value for Pct Migr if
you query for storage pool information while a client transaction (such as a
backup) is in progress. The value for Pct Util is determined by the amount
of space actually allocated (while the transaction is in progress). The value
for Pct Migr represents only the space that is occupied by committed files.
At the end of the transaction, Pct Util and Pct Migr become synchronized.
For sequential-access storage pools, this value is the percentage of the total
bytes of storage available that are currently being used to store active data
(data that is not expired). Because the server can only estimate the
available capacity of a sequential-access storage pool, this percentage also
reflects an estimate of the actual utilization of the storage pool.

Figure 48 on page 382 shows that the estimated capacity for a disk storage pool
named BACKUPPOOL is 80 MB, which is the amount of available space on disk
storage. More than half (51.6%) of the available space is occupied by either backup
files or cached copies of backup files.

The estimated capacity for the tape storage pool named BACKTAPE is 180 MB,
which is the total estimated space available on all tape volumes in the storage
pool. This report shows that 85% of the estimated space is currently being used to
store workstation files.

Note: This report also shows that volumes have not yet been defined to the
ARCHIVEPOOL and ENGBACK1 storage pools because the storage pools show an
estimated capacity of 0.0 MB.

Chapter 10. Managing storage pools and volumes 383

 Obtaining statistics about space-trigger and scratch-volume
utilization in storage pools
Detailed reports about a storage pools list not only estimated capacity and
utilization percentage, but also space-trigger and scratch-volume utilization.

About this task

To obtain a detailed report, issue the following command:

query stgpool format=detailed
Space Trigger Utilization
Specifies the utilization of a storage pool, as calculated by the storage pool
space trigger, if any, for the storage pool. You can define space triggers
only for storage pools associated with DISK or FILE device types.
For sequential-access devices, space trigger utilization is expressed as a
percentage of the number of used bytes on each sequential-access volume
relative to the size of the volume, and the estimated capacity of all existing
volumes in the storage pool. It does not include potential scratch volumes.
Unlike the calculation for percent utilization (Pct Util), the calculation for
space trigger utilization favors creation of new private file volumes by the
space trigger over usage of additional scratch volumes.
For disk devices, space trigger utilization is expressed as a percentage of
the estimated capacity, including cached data and deleted data that is
waiting to be shredded. However, it excludes data that resides on any
volumes that are varied offline. If you issue QUERY STGPOOL while a file
creation is in progress, the value for space trigger utilization can be higher
than the value for percent migration (Pct Migr). The value for space trigger
utilization is determined by the amount of space actually allocated while
the transaction is in progress. The value for percent migration represents
only the space occupied by committed files. At the end of the transaction,
these values are synchronized.
The value for space trigger utilization includes cached data on disk
volumes. Therefore, when cache is enabled and migration occurs, the value
remains the same because the migrated data remains on the volume as
cached data. The value decreases only when the cached data expires or
when the space that cached files occupy needs to be used for no-cached
files.
Number of Scratch Volumes Used
Specifies the number of scratch volumes used in a sequential-access storage
pool. You can use this value, along with the value of the field Maximum
Scratch Volumes Allowed to determine the remaining number of scratch
volumes that the server can request for a storage pool.

Monitoring the use of storage pool volumes

Monitoring how storage pool volumes are used lets you make the most efficient
use available storage.

About this task

Task Required Privilege Class
Display information about volumes Any administrator

You can query the server for information about storage pool volumes:

384 IBM Tivoli Storage Manager for Windows: Administrator's Guide

v General information about a volume, for example:
– Current access mode and status of the volume
– Amount of available space on the volume
– Location
v Contents of a storage pool volume (user files on the volume)
v The volumes that are used by a client node

Obtaining information about storage pool volumes

Standard reports provide a quick overview of basic information about storage pool
volumes. More information is available in detailed reports.

About this task

To request general information about all volumes defined to the server, enter:
query volume

Figure 49 shows an example of the output of this standard query. The example
illustrates that data is being stored on the 8 mm tape volume named WREN01, as
well as on several other volumes in various storage pools.

Volume Name Storage Device Estimated Pct Volume

Pool Name Class Name Capacity Util Status
------------------------ ----------- ---------- --------- ----- --------
D:\STOR\AIXVOL.1 AIXPOOL1 DISK 240.0 M 26.3 On-Line
D:\STOR\AIXVOL.2 AIXPOOL2 DISK 240.0 M 36.9 On-Line
D:\STOR\DOSVOL.1 DOSPOOL1 DISK 240.0 M 72.2 On-Line
D:\STOR\DOSVOL.2 DOSPOOL2 DISK 240.0 M 74.1 On-Line
D:\STOR\OS2VOL.1 OS2POOL1 DISK 240.0 M 55.7 On-Line
D:\STOR\OS2VOL.2 OS2POOL2 DISK 240.0 M 51.0 On-Line
WREN00 TAPEPOOL TAPE8MM 2.4 G 0.0 Filling
WREN01 TAPEPOOL TAPE8MM 2.4 G 2.2 Filling

Figure 49. Information about storage pool volumes

To query the server for a detailed report on volume WREN01 in the storage pool
named TAPEPOOL, enter:
query volume wren01 format=detailed

Figure 50 on page 386 shows the output of this detailed query. Table 35 on page
386 gives some suggestions on how you can use the information.

Chapter 10. Managing storage pools and volumes 385

 Volume Name: WREN01
Storage Pool Name: TAPEPOOL
Device Class Name: TAPE8MM
Estimated Capacity: 2.4 G
Pct Util: 26.3
Volume Status: Filling
Access: Read/Write
Pct. Reclaimable Space: 5.3
Scratch Volume?: No
In Error State?: No
Number of Writable Sides: 1
Number of Times Mounted: 4
Write Pass Number: 2
Approx. Date Last Written: 09/04/2002 [Link]
Approx. Date Last Read: 09/03/2002 [Link]
Date Became Pending:
Number of Write Errors: 0
Number of Read Errors: 0
Volume Location:
Last Update by (administrator): TANAGER
Last Update Date/Time: 09/04/2002 [Link]

Figure 50. Detailed information for a storage pool volume

Table 35. Using the detailed report for a volume

Task Fields and Description
Ensure the volume is available.
Volume Status
Access
Check the Volume Status to see if a disk volume has been varied offline, or if a
sequential-access volume is currently being filled with data.

Check the Access to determine whether files can be read from or written to this
volume.
Monitor the use of storage space.
Estimated Capacity
Pct Util
The Estimated Capacity is determined by the device class associated with the
storage pool to which this volume belongs. Based on the estimated capacity, the
system tracks the percentage of space occupied by client files (Pct Util).

In this example, 26.3% of the estimated capacity is currently in use.

Monitor the error status of the
volume. Number of Write Errors
Number of Read Errors
The server reports when the volume is in an error state and automatically
updates the access mode of the volume to read-only. The Number of Write Errors
and Number of Read Errors indicate the type and severity of the problem. Audit a
volume when it is placed in error state. See “Auditing storage pool volumes” on
page 887 for information about auditing a volume.

386 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Table 35. Using the detailed report for a volume (continued)
Task Fields and Description
Monitor the life of
sequential-access volumes that Scratch Volume?
you have defined to the storage Write Pass Number
pool. Number of Times Mounted
Approx. Date Last Written
Approx. Date Last Read
The server maintains usage statistics on volumes that are defined to storage
pools. Statistics on a volume explicitly defined by an administrator remain for as
long as the volume is defined to the storage pool. The server continues to
maintain the statistics on defined volumes even as the volume is reclaimed and
reused. However, the server deletes the statistics on the usage of a scratch
volume when the volume returns to scratch status (after reclamation or after all
files are deleted from the volume).

In this example, WREN01 is a volume defined to the server by an administrator,

not a scratch volume (Scratch Volume? is No).

The Write Pass Number indicates the number of times the volume has been
written to, starting from the beginning of the volume. A value of one indicates
that a volume is being used for the first time.

In this example, WREN01 has a write pass number of two, which indicates space
on this volume may have been reclaimed or deleted once before.

Compare this value to the specifications provided with the media that you are
using. The manufacturer may recommend a maximum number of write passes
for some types of tape media. You may need to retire your tape volumes after
reaching the maximum passes to better ensure the integrity of your data. To
retire a volume, move the data off the volume by using the MOVE DATA
command. See “Moving data from one volume to another volume” on page 400.

Use the Number of Times Mounted, the Approx. Date Last Written, and the Approx.
Date Last Read to help you estimate the life of the volume. For example, if more
than six months have passed since the last time this volume has been written to
or read from, audit the volume to ensure that files can still be accessed. See
“Auditing storage pool volumes” on page 887 for information about auditing a
volume.

The number given in the field, Number of Times Mounted, is a count of the
number of times that the server has opened the volume for use. The number of
times that the server has opened the volume is not always the same as the
number of times that the volume has been physically mounted in a drive. After a
volume is physically mounted, the server can open the same volume multiple
times for different operations, for example for different client backup sessions.
Determine the location of a Location
volume in a sequential-access
When you define or update a sequential-access volume, you can give location
storage pool.
information for the volume. The detailed query displays this location name. The
location information can be useful to help you track volumes (for example,
off-site volumes in copy storage pools or active-data pools).
Determine if a volume in a Date Became Pending
sequential-access storage pool is
A sequential-access volume is placed in the pending state after the last file is
waiting for the reuse delay period
deleted or moved from the volume. All the files that the pending volume had
to expire.
contained were expired or deleted, or were moved from the volume. Volumes
remain in the pending state for as long as specified with the REUSEDELAY
parameter for the storage pool to which the volume belongs.

Chapter 10. Managing storage pools and volumes 387

 Whether or not a volume is full, at times the Pct Util (percent of the volume
utilized) plus the Pct Reclaimable Space (percent of the volume that can be
reclaimed) may add up to more than 100 percent. This can happen when a volume
contains aggregates that have empty space because of files in the aggregates that
have expired or been deleted. The Pct Util field shows all space occupied by both
non-aggregated files and aggregates, including empty space within aggregates. The
Pct Reclaimable Space field includes any space that is reclaimable on the volume,
also including empty space within aggregates. Because both fields include the
empty space within aggregates, these values may add up to more than 100 percent.
For more information about aggregates, see “How the server groups files before
storing” on page 262 and “Obtaining information about the use of storage space”
on page 396.

Obtaining information about the contents of a storage pool

volume
Any administrator can request information about the contents of a storage pool
volume. Viewing the contents of a storage volume is useful when a volume is
damaged or before you request the server to correct inconsistencies in the volume,
move files from one volume to another, or delete a volume from a storage pool.

About this task

Because the server tracks the contents of a storage volume through its database,
the server does not need to access the requested volume to determine its contents.

To produce a report that shows the contents of a volume, issue the QUERY
CONTENT command.

This report can be extremely large and may take a long time to produce. To reduce
the size of this report, narrow your search by selecting one or all of the following
search criteria:
Node name
Name of the node whose files you want to include in the query.
File space name
Names of file spaces to include in the query. File space names are
case-sensitive and must be entered exactly as they are known to the server.
Use the QUERY FILESPACE command to find the correct capitalization.
Number of files to be displayed
Enter a positive integer, such as 10, to list the first ten files stored on the
volume. Enter a negative integer, such as -15, to list the last fifteen files
stored on the volume.
Filetype
Specifies which types of files, that is, backup versions, archive copies, or
space-managed files, or a combination of these. If the volume being
queried is assigned to an active-data pool, the only valid values are ANY
and Backup.
Format of how the information is displayed
Standard or detailed information for the specified volume.
Damaged
Specifies whether to restrict the query output either to files that are known
to be damaged, or to files that are not known to be damaged.
Copied
Specifies whether to restrict the query output to either files that are backed
388 IBM Tivoli Storage Manager for Windows: Administrator's Guide
 up to a copy storage pool, or to files that are not backed up to a copy
storage pool. Whether files are stored in an active-data pool does not affect
the output.

Note: There are several reasons why a file might have no usable copy in a
copy storage pool:
The file was recently added to the volume and has not yet been backed
up to a copy storage pool
The file should be copied the next time the storage pool is backed
up.
The file is damaged
To determine whether the file is damaged, issue the QUERY
CONTENT command, specifying the DAMAGED=YES parameter.
The volume that contains the files is damaged
To determine which volumes contain damaged files, issue the
following command:
select * from contents where damaged=yes
The file is segmented across multiple volumes, and one or more of the
other volumes is damaged
To determine whether the file is segmented, issue the QUERY
CONTENT command, specifying the FORMAT=DETAILED
parameter. If the file is segmented, issue the following command to
determine whether any of the volumes containing the additional
file segments are damaged:
select volume_name from contents where damaged=yes and
file_name like ’%filename%’

For more information about using the SELECT command, see the
Administrator's Reference.

Example: Generating a standard report about the contents of a volume:

A standard report about the contents of a volume displays basic information such
as the names of files.

About this task

To view the first seven backup files on volume WREN01 from file space /usr on
client node TOMC, for example, enter:
query content wren01 node=tomc filespace=/usr count=7 type=backup

Figure 51 on page 390 displays a standard report which shows the first seven files
from file space /usr on TOMC stored in WREN01.

Chapter 10. Managing storage pools and volumes 389

 Node Name Type Filespace Client’s Name for File
Name
------------------------ ---- ---------- --------------------------------------
TOMC Bkup /usr /bin/ acctcom
TOMC Bkup /usr /bin/ acledit
TOMC Bkup /usr /bin/ aclput
TOMC Bkup /usr /bin/ admin
TOMC Bkup /usr /bin/ ar
TOMC Bkup /usr /bin/ arcv
TOMC Bkup /usr /bin/ banner

Figure 51. A standard report on the contents of a volume

The report lists logical files on the volume. If a file on the volume is an aggregate
of logical files (backed-up or archived client files), all logical files that are part of
the aggregate are included in the report. An aggregate can be stored on more than
one volume, and therefore not all of the logical files in the report may actually be
stored on the volume being queried.

Example: Generating a detailed report about the contents of a volume:

A detailed report about volume contents provides basic information as well as

information about whether the file is stored across multiple volumes, whether the
file is part of an aggregate, and whether the file is a cached copy of a file that has
been migrated to the next storage pool in the hierarchy.

About this task

To display detailed information about the files stored on volume VOL1, enter:
query content vol1 format=detailed

Figure 52 on page 391 displays a detailed report that shows the files stored on
VOL1. The report lists logical files and shows whether each file is part of an
aggregate. If a logical file is stored as part of an aggregate, the information in the
Segment Number, Stored Size, and Cached Copy? fields apply to the aggregate,
not to the individual logical file.

If a logical file is part of an aggregate, the Aggregated? field shows the sequence
number of the logical file within the aggregate. For example, the Aggregated? field
contains the value 2/4 for the file [Link], meaning that this file is the
second of four files in the aggregate. All logical files that are part of an aggregate
are included in the report. An aggregate can be stored on more than one volume,
and therefore not all of the logical files in the report may actually be stored on the
volume being queried.

For disk volumes, the Cached Copy? field identifies whether the file is a cached
copy of a file that has been migrated to the next storage pool in the hierarchy.

390 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 Node Name: DWE
Type: Bkup
Filespace Name: OS2
Client’s Name for File: \ README
Aggregated?: No
Stored Size: 27,089
Segment Number: 1/1
Cached Copy?: No

Node Name: DWE

Type: Bkup
Filespace Name: DRIVE_L_K:
Client’s Name for File: \COMMON\DSMCOMMN\ [Link]
Aggregated?: 1/4
Stored Size: 202,927
Segment Number: 1/1
Cached Copy?: No

Node Name: DWE

Type: Bkup
Filespace Name: DRIVE_L_K:
Client’s Name for File: \COMMON\DSMCOMMN\ [Link]
Aggregated?: 2/4
Stored Size: 202,927
Segment Number: 1/1
Cached Copy?: No

Node Name: DWE

Type: Bkup
Filespace Name: DRIVE_L_K:
Client’s Name for File: \COMMON\DSMCOMMN\ [Link]
Aggregated?: 3/4
Stored Size: 202,927
Segment Number: 1/1
Cached Copy?: No

Node Name: DWE

Type: Bkup
Filespace Name: DRIVE_L_K:
Client’s Name for File: \COMMON\DSMCOMMN\ [Link]
Aggregated?: 4/4
Stored Size: 202,927
Segment Number: 1/1
Cached Copy?: No

Figure 52. Viewing a detailed report of the contents of a volume

Identifying the volumes used by a client node

To identify the sequential volumes used by a client node, you can use the server's
SELECT command.

About this task

The SELECT command queries the VOLUMEUSAGE table in the Tivoli Storage
Manager database. For example, to get a list of volumes used by the EXCH1 client
node in the TAPEPOOL storage pool, enter the following command:
select volume_name from volumeusage where node_name=’EXCH1’ and
stgpool_name=’TAPEPOOL’

The results are something like the following:

VOLUME_NAME
------------------
TAPE01
TAPE08
TAPE13
TAPE21

Chapter 10. Managing storage pools and volumes 391

 For more information about using the SELECT command, see the Administrator's
Reference.

Monitoring migration processes

To obtain information about migration processing, you can request a standard
storage-pool report.

About this task

Four fields on the standard storage-pool report provide you with information
about the migration process. They include:
Pct Migr
Specifies the percentage of data in each storage pool that can be migrated.
This value is used to determine when to start or stop migration.
For random-access and sequential-access disk storage pools, this value
represents the amount of disk space occupied by backed-up, archived, or
space-managed files that can be migrated to another storage pool. The
calculation for random-access disk storage pools excludes cached data, but
includes files on volumes that are varied offline.
For sequential-access tape storage pools, this value is the percentage of the
total volumes in the storage pool that actually contain data at the moment.
For example, assume that a storage pool has four explicitly defined
volumes, and a maximum scratch value of six volumes. If only two
volumes actually contain data at the moment, then Pct Migr is 20%.
This field is blank for copy storage pools and active-data pools.
High Mig Pct
Specifies when the server can begin migrating data from this storage pool.
Migration can begin when the percentage of data that can be migrated
reaches this threshold. (This field is blank for copy storage pools and
active-data pools.)
Low Mig Pct
Specifies when the server can stop migrating data from this storage pool.
Migration can end when the percentage of data that can be migrated falls
below this threshold. (This field is blank for copy storage pools and
active-data pools.)
Next Storage Pool
Specifies the primary storage pool destination to which data is migrated.
(This field is blank for copy storage pools and active-data pools.)

Example: Monitoring data migration between storage pools

A storage pool is queried to determine high and low migration thresholds. The
server is queried to monitor the migration process.

About this task

Figure 48 on page 382 shows that the migration thresholds for BACKUPPOOL
storage pool are set to 50% for the high migration threshold and 30% for the low
migration threshold.

When the amount of migratable data stored in the BACKUPPOOL storage pool
reaches 50%, the server can begin to migrate files to BACKTAPE.

392 IBM Tivoli Storage Manager for Windows: Administrator's Guide

To monitor the migration of files from BACKUPPOOL to BACKTAPE, enter:
query stgpool back*

See Figure 53 for an example of the results of this command.

If caching is on for a disk storage pool and files are migrated, the Pct Util value
does not change because the cached files still occupy space in the disk storage
pool. However, the Pct Migr value decreases because the space occupied by cached
files is no longer migratable.

Storage Device Estimated Pct PctHigh Low Next

Pool Name Class Name Capacity Util Migr Mig Mig Storage
Pct Pct Pool
----------- ---------- ---------- ----- ----- ---- ---- -----------
BACKTAPE TAPE 180.0 M 95.2 100.0 90 70
BACKUPPOOL DISK 80.0 M 51.6 28.8 50 30 BACKTAPE

Figure 53. Information on backup storage pools

You can query the server to monitor the migration process by entering:
query process

A message similar to Figure 54 is displayed:

Process Process Description Status

Number
-------- ------------------------ ---------------------------------------------
2 Migration Disk Storage Pool BACKUPPOOL, Moved Files:
1086, Moved Bytes: 25555579, Unreadable
Files: 0, Unreadable Bytes: 0

Figure 54. Information on the migration process

When migration is finished, the server displays the following message:

ANR1101I Migration ended for storage pool BACKUPPOOL.

Managing problems during migration processes

Migration processes can be suspended if a problem occurs. If migration is
suspended, you can retry the process, cancel the process, end the migration process
by changing the attributes of the storage pool from which data is being migrated,
or provide additional space.

Canceling migration processes

To stop server migration when a problem occurs or when you need the resources
the process is using, you can cancel the process.

About this task

First determine the identification number of the migration process by entering:

query process

A message similar to Figure 55 on page 394 is displayed:

Chapter 10. Managing storage pools and volumes 393

 Process Process Description Status
Number
-------- ------------------------ ---------------------------------------------
1 Migration ANR1113W Migration suspended for storage pool
BACKUPPOOL - insufficient space in
subordinate storage pool.

Figure 55. Getting the identification number of the migration process

Then you can cancel the migration process by entering:

cancel process 1

Stopping repeated attempts by the server to restart migration

Some errors cause the server to continue attempting to restart the migration
process after 60 seconds. (If the problem still exists after several minutes, the
migration process ends.) To stop the repeated attempts at restart, you can change
some characteristics of the storage pool from which data is being migrated.

About this task

Depending on your environment, you can:

v Set higher migration thresholds for the storage pool from which data is being
migrated. The higher threshold means the storage pool must have more
migratable data before migration starts. This change delays migration.
In the example in “Example: Monitoring data migration between storage pools”
on page 392, you could update the disk storage pool BACKUPPOOL.
v Add volumes to the pool from which data is being migrated. Adding volumes
decreases the percentage of data that is migratable (Pct Migr).
In the example in “Example: Monitoring data migration between storage pools”
on page 392, you could add volumes to the disk storage pool BACKUPPOOL to
increase its storage capacity.

Tip: Do this only if you received an out-of-space message for the storage pool to
which data is being migrated.

Providing additional space for the migration process

A migration process can be suspended because of insufficient space in the storage
pool to which data is being migrated. To allow the migration process to complete,
you can provide additional storage volumes for that storage pool.

About this task

In the example in “Example: Monitoring data migration between storage pools” on

page 392, you can add volumes to the BACKTAPE storage pool or increase the
maximum number of scratch tapes allowed for it. Either way, you increase the
storage capacity of BACKTAPE.

394 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Monitoring the use of cache space on disk storage
To determine whether cache is being used on disk storage and to monitor how
much space is being used by cached copies, query the server for a detailed storage
pool report.

About this task

The Pct Util value includes cached data on a volume (when cache is enabled) and
the Pct Migr value excludes cached data. Therefore, when cache is enabled and
migration occurs, the Pct Migr value decreases while the Pct Util value remains the
same. The Pct Util value remains the same because the migrated data remains on
the volume as cached data. In this case, the Pct Util value only decreases when the
cached data expires.

If you update a storage pool from CACHE=YES to CACHE=NO, the cached files
will not disappear immediately. The Pct Util value will be unchanged. The cache
space will be reclaimed over time as the server needs the space, and no additional
cached files will be created.

For example, to request a detailed report for BACKUPPOOL, enter:

query stgpool backuppool format=detailed

Figure 56 on page 396 displays a detailed report for the storage pool.

Chapter 10. Managing storage pools and volumes 395

 Storage Pool Name: BACKUPPOOL
Storage Pool Type: PRIMARY
Device Class Name: DISK
Estimated Capacity: 80.0 M
Space Trigger Util: 0.0
Pct Util: 42.0
Pct Migr: 29.6
Pct Logical: 82.1
High Mig Pct: 50
Low Mig Pct: 30
Migration Delay: 0
Migration Continue: Yes
Migration Processes: 1
Reclamation Processes:
Next Storage Pool: BACKTAPE
Reclaim Storage Pool:
Maximum Size Threshold: No Limit
Access: Read/Write
Description:
Overflow Location:
Cache Migrated Files?: Yes
Collocate?:
Reclamation Threshold:
Offsite Reclamation Limit:
Maximum Scratch Volumes Allowed:
Number of Scratch Volumes Used:
Delay Period for Volume Reuse: 0 Day(s)
Migration in Progress?: Yes
Amount Migrated (MB): 0.10
Elapsed Migration Time (seconds): 5
Reclamation in Progress?:
Last Update by (administrator): SERVER_CONSOLE
Last Update Date/Time: 09/04/2002 [Link]
Storage Pool Data Format: Native
Copy Storage Pool(s):
Active-data Pool(s):
Continue Copy on Error?:
CRC Data: No
Reclamation Type:
Overwrite Data when Deleted: 2 Time(s)

Figure 56. Detailed storage pool report

When Cache Migrated Files? is set to Yes, the value for Pct Util should not change
because of migration, because cached copies of files migrated to the next storage
pool remain in disk storage.

This example shows that utilization remains at 42%, even after files have been
migrated to the BACKTAPE storage pool, and the current amount of data eligible
for migration is 29.6%.

When Cache Migrated Files? is set to No, the value for Pct Util more closely
matches the value for Pct Migr because cached copies are not retained in disk
storage.

Obtaining information about the use of storage space

You can generate reports to determine the amount of space used by client nodes
and file spaces, storage pools and device classes, or types of data (backup, archive,
or space-managed). Generating occupancy reports on a regular basis can help you
with capacity planning.

396 IBM Tivoli Storage Manager for Windows: Administrator's Guide

About this task
Task Required Privilege Class
Query the server for information about Any administrator
server storage

To obtain reports with information broken out by node or file space, issue the
QUERY OCCUPANCY command.

Each report gives two measures of the space in use by a storage pool:
v Logical space occupied
The amount of space used for logical files. A logical file is a client file. A logical
file is stored either as a single physical file, or in an aggregate with other logical
files. The logical space occupied in active-data pools includes the space occupied
by inactive logical files. Inactive logical files in active-data pools are removed by
reclamation.
v Physical space occupied
The amount of space used for physical files. A physical file is either a single
logical file, or an aggregate composed of logical files.
An aggregate might contain empty space that was used by logical files that are
now expired or deleted, or that were deactivated in active-data pools. Therefore,
the amount of space used by physical files is equal to or greater than the space
used by logical files. The difference gives you a measure of how much unused
space any aggregates may have. The unused space can be reclaimed in
sequential storage pools.

You can also use this report to evaluate the average size of workstation files stored
in server storage.

Obtaining information about space used by client nodes

You can request information about how much data a client has backed up,
archived, or migrated to server storage. You can also request information about the
amount of storage space used by each client node and file space, as well as the
number of files that are in server storage that were backed up to a copy storage
pool or an active-data pool.

About this task

To determine the amount of server storage space used by the /home file space
belonging to the client node MIKE, for example, enter:
query occupancy mike /home

File space names are case-sensitive and must be entered exactly as they are known
to the server. To determine the correct capitalization, issue the QUERY FILESPACE
command. For more information, see “Managing file spaces” on page 451.

Figure 57 on page 398 shows the results of the query. The report shows the number
of files backed up, archived, or migrated from the /home file space belonging to
MIKE. The report also shows how much space is occupied in each storage pool.

If you back up the ENGBACK1 storage pool to a copy storage pool, the copy
storage pool would also be listed in the report. To determine how many of the
client node's files in the primary storage pool have been backed up to a copy

Chapter 10. Managing storage pools and volumes 397

 storage pool, compare the number of files in each pool type for the client node.

Physical Logical
Node Name Type Filespace Storage Number of Space Space
Name Pool Name Files Occupied Occupied
(MB) (MB)
--------------- ---- ----------- ----------- --------- ---------- --------
MIKE Bkup /home ENGBACK1 513 3.52 3.01

Figure 57. A report of the occupancy of storage pools by client node

You can also use the QUERY NODEDATA command to display information about
the data for one or more nodes in a sequential-access storage pool. (The command
is not supported for random-access storage pools.) The output of the QUERY
NODEDATA command displays the name of the volume on which a node's data is
written, the name of the storage pool in which the volume is located, and the
amount of space occupied by the data on the volume. For example, to display
information about the data for nodes whose names begin with the letter “e,” you
would enter the following command using a wildcard character:
query nodedata e*
Node Name Volume Name Storage Pool Physical
Name Space
Occupied
(MB)
--------- ------------------------------ ------------ --------
EDU_J2 E:\tsm\server\[Link] EDU512 0.01
EDU_J2 E:\tsm\server\[Link] EDU319 0.01
EDU_J3 E:\tsm\server\[Link] EDU512 0.01

For details about the QUERY NODEDATA command, refer to the Administrator's
Reference.

Obtaining information about space utilization of storage pools

You can monitor the amount of space being used by an individual storage pool or
a group of storage pools.

About this task

To query the server for the amount of data stored in backup tape storage pools
belonging to the TAPECLASS device class, for example, enter:
query occupancy devclass=tapeclass

Figure 58 on page 399 displays a report on the occupancy of tape storage pools
assigned to the TAPECLASS device class.

398 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 Node Name Type Filespace Storage Number of Physical Logical
Name Pool Name Files Space Space
Occupied Occupied
(MB) (MB)
--------------- ---- ----------- ----------- --------- ---------- --------
CAROL Arch OS2C ARCHTAPE 5 .92 .89
CAROL Bkup OS2C BACKTAPE 21 1.02 1.02
PEASE Arch /home/peas- ARCHTAPE 492 18.40 18.40
e/dir
PEASE Bkup /home/peas- BACKTAPE 33 7.60 7.38
e/dir
PEASE Bkup /home/peas- BACKTAPE 2 .80 .80
e/dir1
TOMC Arch /home/tomc ARCHTAPE 573 20.85 19.27
/driver5
TOMC Bkup /home BACKTAPE 13 2.02 1.88

Figure 58. A report on the occupancy of storage pools by device class

Tip: For archived data, you might see “(archive)” in the Filespace Name column
instead of a file space name. This means that the data was archived before
collocation by file space was supported by the server.

Requesting information about space used by backed-up,

archived, and space-managed files
You can query the server for the amount of space used by backed-up, archived,
and space-managed files. By determining the average size of workstation files
stored in server storage, you can estimate how much storage capacity you might
need when registering new client nodes to the server.

About this task

For example, to request a report about backup versions stored in the disk storage
pool named BACKUPPOOL, enter:
query occupancy stgpool=backuppool type=backup

Figure 59 displays a report on the amount of server storage used for backed-up
files.

Node Name Type Filespace Storage Number of Physical Logical

Name Pool Name Files Space Space
Occupied Occupied
(MB) (MB)
--------------- ---- ----------- ----------- --------- ---------- --------
CAROL Bkup OS2C BACKUPPOOL 513 23.52 23.52
CAROL Bkup OS2D BACKUPPOOL 573 20.85 20.85
PEASE Bkup /marketing BACKUPPOOL 132 12.90 9.01
PEASE Bkup /business BACKUPPOOL 365 13.68 6.18
TOMC Bkup / BACKUPPOOL 177 21.27 21.27

Figure 59. A report of the occupancy of backed-up files in storage pools

To determine the average size of backup versions stored in BACKUPPOOL,

complete the following steps using the data provided in Figure 59:

Procedure
1. Add the number of megabytes of space occupied by backup versions. In this
example, backup versions occupy 92.22 MB of space in BACKUPPOOL.
2. Add the number of files stored in the storage pool. In this example, 1760
backup versions reside in BACKUPPOOL.

Chapter 10. Managing storage pools and volumes 399

 3. Divide the space occupied by the number of files to determine the average size
of each file backed up to the BACKUPPOOL. In this example, the average size
of each workstation file backed up to BACKUPPOOL is about 0.05 MB, or
approximately 50 KB.

Results

You can use this average to estimate the capacity required for additional storage
pools that are defined to the server.

For information about planning storage space, see “Estimating space needs for
storage pools” on page 379 and “Estimating space for archived files in
random-access storage pools” on page 381.

Obtaining information about free disk space in FILE device

classes
You can monitor the amount of free disk space in directories associated with FILE
device classes. The Tivoli Storage Manager server uses directories as the location
for files that represent storage-pool volumes.

Procedure

To request information about the amount of free disk space in each directory for all
device classes with a device type of FILE, issue QUERY DIRSPACE command.

Example

Figure 60 displays the output for this command.

Device Directory Estimated Estimated

Class Capacity Available
----------- ---------------------------- ---------- ----------
DBBKUP G:\This\is\a\large\directory 13,000 M 5,543 M
DBBKUP G:\This\is\directory2 13,000 M 7,123 M
DBBKUP2 G:\This\is\a\huge\directory 2,256 G 2,200 G

Figure 60. A report of the free disk space for all device classes of device type FILE

To obtain the amount of free space associated with a particular device class, issue
the following command:
query dirspace device_class_name

Moving data from one volume to another volume

You might need to move data in some situations, for example, when you need to
salvage readable data from a damaged volume. To move data (files) from one
volume to another volume in the same or a different storage pool, use the MOVE
DATA command. The volumes can be onsite volumes or off-site volumes.

About this task

Task Required Privilege Class
Move files from a volume in any storage System or unrestricted storage
pool to an available volume in any storage
pool

400 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 Task Required Privilege Class
Move files from one volume to an available Restricted storage
volume in any storage pool to which you
are authorized

During the data movement process, the server:

v Moves any readable files to available volumes in the specified destination
storage pool
v Deletes any cached copies from a disk volume
v Attempts to bypass any files that previously were marked as damaged

During the data movement process, users cannot access the volume to restore or
retrieve files, and no new files can be written to the volume.

Remember:
v Files in a copy storage pool or an active-data pool do not move when primary
files are moved.
v You cannot move data into or out of a storage pool that is defined with a
CENTERA device class.
v In addition to moving data from volumes in storage pools that have NATIVE or
NONBLOCK data formats, you can also move data from volumes in storage
pools that have NDMP data formats (NETAPPDUMP, CELERRADUMP, or
NDMPDUMP). The target storage pool must have the same data format as the
source storage pool. If you are moving data out of a storage pool for the
purpose of upgrading to new tape technology, the target primary storage pool
must be associated with a library that has the new device for the tape drives.
v If you run the MOVE DATA command on an offsite volume that contains collocated
data, it might be necessary to issue the MOVE DATA command multiple times to
move all of the data out of the volume. Each MOVE DATA command moves the
data for a single collocation group.

Data movement within the same storage pool

Moving files from one volume to other volumes in the same storage pool provides
a number of benefits.

Moving files from one volume to other volumes in the same storage pool is useful:
v When you want to free up all space on a volume so that it can be deleted from
the Tivoli Storage Manager server
See “Deleting storage pool volumes” on page 414 for information about deleting
backed-up, archived, or space-managed data before you delete a volume from a
storage pool.
v When you need to salvage readable files from a volume that has been damaged
v When you want to delete cached files from disk volumes
If you want to force the removal of cached files, you can delete them by moving
data from one volume to another volume. During the move process, the server
deletes cached files remaining on disk volumes.

If you move data between volumes within the same storage pool and you run out
of space in the storage pool before all data is moved from the target volume, then
you cannot move all the data from the target volume. In this case, consider moving
data to available space in another storage pool as described in “Data movement to

Chapter 10. Managing storage pools and volumes 401

 a different storage pool.”

Data movement to a different storage pool

You can move all data from a volume in one storage pool to volumes in another
storage pool. When you specify a target storage pool that is different than the
source storage pool, the server uses the storage hierarchy to move data if more
space is required.

Remember: Data cannot be moved from a primary storage pool to a copy storage
pool or to an active-data pool. Data in a copy storage pool or an active-data pool
cannot be moved to another storage pool.

You can move data from random-access storage pools to sequential-access storage
pools. For example, if you have a damaged disk volume and you have a limited
amount of disk storage space, you could move all files from the disk volume to a
tape storage pool. Moving files from a disk volume to a sequential storage pool
may require many volume mount operations if the target storage pool is
collocated. Ensure that you have sufficient personnel and media to move files from
disk to sequential storage.

When a data move from a shred pool is complete, the original data is shredded.
However, if the destination is not another shred pool, you must set the
SHREDTONOSHRED parameter to YES to force the movement to occur. If this
value is not specified, the server issues an error message and does not allow the
data to be moved. See “Securing sensitive client data” on page 539 for more
information about shredding.

Data movement from offsite volumes in copy storage pools or

active-data pools
You can move data from offsite volumes without bringing the volumes onsite.

Processing of the MOVE DATA command for volumes in copy-storage pools and
active-data pools is similar to that of primary-storage pools, with the following
exceptions:
v Volumes in copy-storage pools and active-data pools might be set to an access
mode of offsite, making them ineligible to be mounted. During processing of the
MOVE DATA command, valid files on offsite volumes are copied from the original
files in the primary-storage pools. In this way, valid files on offsite volumes are
copied without having to mount these volumes. These new copies of the files are
written to another volume in the copy-storage pool or active-data pool.
v With the MOVE DATA command, you can move data from any primary-storage
pool volume to any primary-storage pool. However, you can move data from a
copy-storage pool volume only to another volume within the same-copy storage
pool. Similarly, you can move data from an active-data pool volume only to
another volume within the same active-data pool.

When you move files from a volume that is marked as offsite, the server performs
the following actions:
1. Determines which files still remain on the volume from which you are moving
data
2. Obtains the remaining files from a primary-storage pool or from another
copy-storage pool

402 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 3. Copies the files to one or more volumes in the destination copy-storage pool or
active-data pool

During the data movement process, active-data pools cannot be used to obtain
data.

Processing of the MOVE DATA command for primary-storage pool volumes does not
affect copy-storage pool or active-data pool files.

Moving data
You can move data using the MOVE DATA command. Before moving data,
however, take steps to ensure that the move operation succeeds.

Before you begin

Before beginning this procedure:

v If you want to ensure that no new files are written to a volume after you move
data from it, change the volume's access mode to read-only. This prevents the
server from filling the volume with data again as soon as data is moved. You
might want to do this if you want to delete the volume.
See “Updating storage pool volumes” on page 257 for information about
updating the access mode of a storage pool volume.
v Ensure that sufficient space is available on volumes within the specified
destination storage pool by:
1. Querying the source storage volume to determine how much space is
required on other volumes. See “Monitoring the use of storage pool
volumes” on page 384 for information about requesting information about a
storage volume.
2. Querying the specified destination storage pool to ensure that there is
sufficient capacity to store the files that are being moved. See “Monitoring
space available in a storage pool” on page 382 for information about
querying a storage pool.
If you need more storage space, define volumes or increase the maximum
number of scratch volumes in the specified destination storage pool. See
“Defining storage pool volumes” on page 256 for preparing volumes to be used
for server storage.
v If you are moving files from a volume in a sequential storage pool to another
volume in the same storage pool, ensure that the mount limit of the device class
associated with the storage pool is greater than one.
For information about the mount limit value for the device class, see:
“Obtaining information about device classes” on page 198
v If you are moving files from a tape volume to a tape storage pool, ensure that
the two tape drives that are required are available.

Procedure

To move data, issue the MOVE DATA command.

For example, to move the files that are stored in the d:\storage\vol3 volume to
any available volume in the STGTMP1 storage pool, enter:
move data d:\storage\vol3 stgpool=stgtmp1

When you move data from a volume, the server starts a background process and
sends informational messages, such as:

Chapter 10. Managing storage pools and volumes 403

 ANR1140I Move Data process started for volume D:\STORAGE\VOL3
(process ID 32).

To run the MOVE DATA command in the foreground on an administrative client, issue
the command with the WAIT=YES parameter.

Remember:
v A volume might not be totally empty after a move data operation completes. For
example, the server might be unable to relocate one or more files to another
volume because of input/output errors on the device or because errors were
found in the file. You can delete the volume with DISCARDDATA=YES to delete
the volume and any remaining files. The server then deletes the remaining files
that had I/O or other errors.
v In addition to moving data from volumes in storage pools that have NATIVE or
NONBLOCK data formats, you can also move data from volumes in storage
pools that have NDMP data formats (NETAPPDUMP, CELERRADUMP, or
NDMPDUMP). The target storage pool must have the same data format as the
source storage pool. If you are moving data out of a storage pool for the
purpose of upgrading to new tape technology, the target primary storage pool
must be associated with a library that has the new device for the tape drives.
v If you run the MOVE DATA command on an offsite volume that contains collocated
data, it might be necessary to issue the MOVE DATA command multiple times to
move all of the data out of the volume. Each MOVE DATA command moves the
data for a single collocation group.

Requesting information about the data-movement process

You can query the server for statistics about the data-movement process.

About this task

To request information, issue the QUERY PROCESS command.

Figure 61 shows an example of the report that you receive about the data
movement process.

Process Process Description Status

Number
-------- -------------------- -------------------------------------------------
32 Move Data Volume D:\STORAGE\VOL3, (storage pool BACKUPPOOL),
Target Pool STGTMP1, Moved Files: 49, Moved
Bytes: 9,121,792, Unreadable Files: 0,
Unreadable Bytes: 0. Current File (bytes):
3,522,560

Current output volume: VOL1.

Figure 61. Information about the file movement process

404 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Reclaiming space in aggregates by moving data
Empty space accumulates in a file aggregate as logical files in that aggregate are
deleted. During reclamation processing, the aggregate is reconstructed by removing
the empty space left by the deleted files. However, you cannot start reclamation
processing for specific volumes.

About this task

To remove empty space in a specific volume and reconstruct an aggregate, issue

the MOVE DATA command. By default, this command removes the empty space
occupied by deleted files in an aggregate.

Remember:
1. Reclaiming empty space in NDMP-generated images is not an issue because
NDMP-generated images are not aggregated.
2. Reconstruction removes inactive backup files in active-data pools. Specifying
RECONSTRUCT=NO when moving data from volumes in an active-data pool
prevents the inactive backup files from being removed.

Monitoring the movement of data between volumes

You can query the server for volume information to monitor the movement of data
between volumes.

About this task

Procedure

To request information, use the QUERY VOLUME command.

Example

For example, to see how much data has moved from the source volume in the
move operation example, enter:
query volume d:\storage\vol3 stgpool=backuppool

Near the beginning of the move process, querying the volume from which data is
being moved gives the following results:
Volume Name Storage Device Estimated Pct Volume
Pool Name Class Name Capacity Util Status
--------------- ----------- ---------- --------- ----- --------
D:\STORAGE\VOL3 BACKUPPOOL DISK 15.0 M 59.9 On-Line

Querying the volume to which data is being moved (VOL1, according to the
process query output) gives the following results:
Volume Name Storage Device Estimated Pct Volume
Pool Name Class Name Capacity Util Status
---------------- ----------- ---------- --------- ----- --------
VOL1 STGTMP1 8500DEV 4.8 G 0.3 Filling

At the end of the move process, querying the volume from which data was moved
gives the following results:

Chapter 10. Managing storage pools and volumes 405

 Volume Name Storage Device Estimated Pct Volume
Pool Name Class Name Capacity Util Status
---------------- ---------- ---------- --------- ----- --------
D:\STORAGE\VOL3 BACKUPPOOL DISK 15.0 M 0.0 On-Line

Moving data belonging to a client node

You can move data in a sequential-access storage pool for one or more nodes, or
for a single node with selected file spaces, using the MOVE NODEDATA command. The
data can be in either a primary storage pool, copy storage pool, or active-data
pool.

About this task

When the source storage pool is a primary storage pool, you can move data to
other volumes within the same pool or to another primary storage pool. When the
source storage pool is a copy storage pool, data can only be moved to other
volumes within that storage pool. When the source storage pool is an active-data
pool, data can only be moved to other volumes within that same storage pool.

Tips:
v In addition to moving data from volumes in storage pools that have NATIVE or
NONBLOCK data formats, you can also move data from volumes in storage
pools that have NDMP data formats (NETAPPDUMP, CELERRADUMP, or
NDMPDUMP). The target storage pool must have the same data format as the
source storage pool.
v If you are moving files within the same storage pool, there must be volumes
available that do not contain the data that you are moving. That is, the server
cannot use a destination volume that contains data that needs to be moved.
v When you move data from volumes in an active-data pool, you have the option
of reconstructing file aggregates during data movement. Reconstruction removes
inactive backup files in the pool. Specifying no reconstruction prevents the
inactive files from being removed.
v You cannot move node data into or out of a storage pool that is defined with a
CENTERA device class.

Task Required Privilege Class

Move data by node System, unrestricted storage or restricted
storage

Moving data in all file spaces belonging to one or more nodes

Moving data for all file spaces on one or more nodes provides a number of
benefits.

About this task

Moving data is useful:

v When you want to optimize performance by reducing the number of volume
mounts required during a restore operation by consolidating data for a specific
node or nodes within a storage pool
v When you want to move data for specified nodes into a different storage pool

406 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 v When you want to increase performance of client restore processing by first
moving data to a random-access storage pool

Best practice: Avoid movement of data into, out of, or within a storage pool while
MOVE NODEDATA is concurrently processing data on the same storage pool.

To move all file spaces for a single node named ACCOUNTING where the data is
in storage pool ACCTPOOL and the destination storage pool is BACKUPPOOL
enter:
move nodedata accounting fromstgpool=acctpool tostgpool=backuppool

Moving data in selected file spaces belonging to a single node

Moving data for selected file spaces for one node provides a number of benefits.

About this task

Moving data is useful:

v When you want to optimize performance by reducing the number of volume
mounts required during a restore operation by consolidating data for specific file
spaces within a storage pool.
v When you want to consolidate data for critical file spaces allowing restore of
these files to be given higher priority during recovery situations. This would be
advantageous during data recovery when it is essential to first restore only
business-critical data and then restore non-business-critical data.
v When you want to move specific file spaces into a different storage pool.
v When you want to increase performance of client restore processing by first
moving data to a random-access storage pool.

For example, consider moving data for a single node and restricting the data
movement to files in a specific non-Unicode file space (for this example, \\eng\e$)
as well as a specific Unicode file space (for this example, \\eng\d$ ). The node
name owning the data is ENGINEERING and it currently has data stored in the
ENGPOOL storage pool. After the move is complete, the data is located in the
destination storage pool BACKUPPOOL. To move the data enter the following:
move nodedata engineering fromstgpool=engpool
tostgpool=backuppool filespace=\\eng\e$ unifilespace=\\eng\d$

Another example is to move data for a single node named MARKETING from all
primary sequential-access storage pools to a random-access storage pool named
DISKPOOL. First obtain a list of storage pools that contain data for node
MARKETING, issue either:
query occupancy marketing

or
SELECT * from OCCUPANCY where node_name=’MARKETING’;

Chapter 10. Managing storage pools and volumes 407

 For this example the list of resulting storage pool names all begin with the
characters FALLPLAN. To move the data repeat the following command for every
instance of FALLPLAN. The following example displays the command for
FALLPLAN3:
move nodedata marketing fromstgpool=fallplan3
tostgpool=diskpool

A final example shows moving both non-Unicode and Unicode file spaces for a
node. For node NOAH move non-Unicode file space \\servtuc\d$ and Unicode
file space \\tsmserv1\e$ that has a file space ID of 2 from sequential-access
storage pool TAPEPOOL to random-access storage pool DISKPOOL.
move nodedata noah fromstgpool=tapepool tostgpool=diskpool
filespace=\\servtuc\d$ fsid=2

Obtaining information about data-movement processes

You can query the server for statistics about the data movement for a client node.

About this task

To request information on the data movement process, enter:

query process

Figure 62 shows an example of the report that you receive about the data
movement process.

Process Process Description Status

Number
-------- -------------------- -------------------------------------------------
3 Move Node Data Storage Pool 3590FC, Target Pool 3590FC Files
Moved: 0, Bytes Moved: 0, Unreadable Files: 0,
Unreadable Bytes: 0. Current Physical File
(bytes): 268,468,584

Current input volume:

DST308.

Current output volume:

DST279.

Figure 62. Information on the data movement process

Troubleshooting incomplete data-movement operations

There are several reasons why an operation to move node data might not
complete.

About this task

The most common reasons are:

v Files are marked as damaged in the source storage pool. For more information
about how to work with files that are marked as damaged, see “Fixing damaged
files” on page 896.

408 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 v Files in the source storage pool reside on volumes whose access mode is offsite,
destroyed, or unavailable. To complete the move operation, bring the volumes
onsite, restore destroyed volumes from a copy storage pool or an active-data
pool, or make the volumes available.
v Files were moved, added, or deleted during the move operation. To prevent this
situation, avoid the following operations during move processing:
– Migration of any type relating to the storage pool
– Reclamation of volumes within the storage pool
– Simultaneously running MOVE DATA processing for a volume in a storage
pool that contains data to be moved during MOVE NODEDATA processing
– Backup operations into a copy storage pool or an active-data pool while a
MOVE NODEDATA is running for that pool
– Storage of files from a client directly into the storage pool

Note: If you run the MOVE DATA command on an offsite volume that contains
collocated data, it might be necessary to issue the MOVE DATA command multiple
times to move all of the data out of the volume. Each MOVE DATA command moves
the data for a single collocation group.

Renaming storage pools

When distributing policy using enterprise configuration, you might need to rename
a storage pool. Renaming a storage pool can require changing the destination
storage pool in copy groups and management classes.

About this task

To rename a storage pool, issue the UPDATE STGPOOL command.

When you rename a storage pool, any administrators with restricted storage
privilege for the storage pool automatically have restricted storage privilege to the
storage pool under the new name. If the renamed storage pool is in a storage pool
hierarchy, the hierarchy is preserved.

Copy groups and management classes might contain a storage pool name as a
destination. If you rename a storage pool used as a destination, the destination in a
copy group or management class is not changed to the new name of the storage
pool. To continue to use the policy with the renamed storage pool as a destination,
you must change the destination in the copy groups and management classes. You
then activate the policy set with the changed destinations.

For information about setting up a managed server in an enterprise configuration,

see “Setting up a managed server” on page 705.

Defining copy storage pools and active-data pools

Use a copy storage pool or an active-data pool to back up one or more primary
storage pools. When you definecopy storage pools or active-data pools, you can
take advantage of various Tivoli Storage Manager functions by specifying certain
properties.

Chapter 10. Managing storage pools and volumes 409

 About this task

To define a copy storage pool, issue the DEFINE STGPOOL command and specify
POOLTYPE=COPY. To define an active-data pool, issue the DEFINE STGPOOL
command and specify POOLTYPE=ACTIVEDATA. When you define a copy
storage pool or an active-data pool, be prepared to provide some or all of the
information in Table 36.

Remember:
1. To back up a primary storage pool to an active-data pool, the data format must
be NATIVE or NONBLOCK. You can back up a primary storage pool to a copy
storage pool using NATIVE, NONBLOCK, or any of the NDMP formats. The
target storage pool must have the same data format as the source storage pool.
2. You cannot define copy storage pools or active-data pools for a Centera device
class.
Table 36. Information for defining copy storage pools and active-data pools
Information Explanation
Device class Specifies the name of the device class assigned for the storage pool. This
is a required parameter.
Pool type Specifies that you want to define a copy storage pool or an active-data
pool. This is a required parameter. You cannot change the pool type
when updating a storage pool.
Access mode Defines access to volumes in the storage pool for user operations (such
as backup and restore) and system operations (such as reclamation).
Possible values are:
Read/Write
User and system operations can read from or write to the
volumes.
Read-Only
User operations can read from the volumes, but not write.
However, system processes can move files within the volumes
in the storage pool.
Unavailable
Specifies that users cannot access files stored on volumes in the
copy storage pool or an active-data pool. Files can be moved
from volume to volume with the same copy storage pool or
from volume to volume within the same active-data pool, but
no new writes are permitted to the volumes in the storage pool
from volumes outside the storage pool.
Maximum When you specify a value greater than zero, the server dynamically
number of scratch acquires scratch volumes when needed, up to this maximum number.
volumes This is a required parameter.

For automated libraries, set this value equal to the physical capacity of
the library. For details, see:
“Adding scratch volumes to automated library devices” on page 138
Collocation When collocation is enabled, the server tries to keep all files on a
minimal number of sequential-access storage volumes. The files can
belong to a single client node, a group of client nodes, a client file space,
or a group of file spaces. See “Collocation of copy storage pools and
active-data pools” on page 364.

410 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 Table 36. Information for defining copy storage pools and active-data pools (continued)
Information Explanation
Reclamation Specifies when to initiate reclamation of volumes in the copy storage
threshold pool or active-data pool. Reclamation is a process that moves any
remaining files from one volume to another volume, thus making the
original volume available for reuse. A volume is eligible for reclamation
when the percentage of unused space on the volume is greater than the
reclaim parameter value.

Reclamation processing works differently for off-site copy storage pool

volumes, active-data pool volumes, and virtual volumes. When a copy
storage pool volume or an active-data pool volume that is off-site
becomes eligible for reclamation, the reclamation process tries to retrieve
the files on the reclaimable volume from a primary or copy storage pool
volume that is onsite. The process then writes these files to an available
volume in the original copy storage pool or active-data pool. See
“Reclaiming copy storage pools and active-data pools” on page 374 and
“Controlling reclamation of virtual volumes” on page 374 for more
details.
Reuse delay Specifies the number of days that must elapse after all of the files have
period been deleted from a volume before the volume can be rewritten or
returned to the scratch pool. See “Delayed reuse of reclaimed volumes”
on page 378.
Off-site reclaim Specifies the number of off-site volumes to be reclaimed during
limit reclamation for a storage pool. See “Reclamation of offsite volumes” on
page 375.
Reclamation Specifies the number of concurrent processes to use for reclaiming the
processes volumes in a storage pool. See “Reclaiming copy storage pools and
active-data pools” on page 374.

For more information, see “Backing up primary storage pools” on page 883.

Example: Defining a copy storage pool

Copies of the files are stored in the Tivoli Storage Manager default disk storage
pools for disaster recovery purposes. You create a copy storage pool and decide to
use only scratch tapes in the new pool.

About this task

Assume you need to maintain copies of the files stored in BACKUPPOOL,

ARCHIVEPOOL, and SPACEMGPOOL (default disk storage pools) for disaster
recovery purposes. You want to create a copy storage pool named
DISASTER-RECOVERY. You decide to use only scratch tapes in the new pool,
setting the maximum number of scratch volumes to an appropriate value. You
enter the following command:
define stgpool disaster-recovery tapeclass pooltype=copy
maxscratch=100

To store data in the new storage pool, you must back up the primary storage pools
(BACKUPPOOL, ARCHIVEPOOL, and SPACEMGPOOL) to the
DISASTER-RECOVERY pool. See “Backing up primary storage pools” on page 883.

Chapter 10. Managing storage pools and volumes 411

 Properties of primary, copy, and active-data pools
Primary-storage pools, copy-storage pools, and active-data pools have different
properties. Understanding these differences helps you make the most efficient use
of storage space.

Table 37 compares the characteristics of primary, copy-storage, and active-data

pools.
Table 37. Comparing primary-storage pools, copy-storage pools, and active-data pools
Characteristic Primary-storage pool Copy-storage pool Active-data pool
Destination for backed-up Yes No No
or archived files (specified
in backup or archive copy
groups)
Destination for Yes No No
space-managed files
(specified in the
management class)
Off-site access mode for No Yes, except for volumes Yes, except for volumes
volumes with device type SERVER with device type SERVER
Destroyed access mode for Yes No No
volumes
Random-access storage Yes No No
volumes
Sequential-access storage Yes Yes Yes
volumes
Contents Client files (backup Copies of files that are Active-only versions of
versions, archived files, stored in primary storage client backup files that are
space-managed files) pools stored in primary storage
pools. Archive data and
space-managed files are not
permitted.
Moving data allowed Within the same Within the same storage Within the same storage
primary-storage pool, or to pool only. pool only.
any primary-storage pool
v If moving data by volume v If moving data by volume
and volumes are off-site, and volumes are off-site,
data is copied from the data is copied from the
original files in original files in
primary-storage pools. primary-storage pools.

v If volumes are off-site, you v If volumes are off-site, you

cannot move data in those cannot move data in those
volumes by node. volumes by node.
Collocation Yes (sequential-access Yes Yes
storage pools only)

412 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Table 37. Comparing primary-storage pools, copy-storage pools, and active-data pools (continued)
Characteristic Primary-storage pool Copy-storage pool Active-data pool
Reclamation Yes (sequential-access Yes Yes
storage pools only)
Virtual volumes (volumes Virtual volumes (volumes
with device type SERVER) with device type SERVER)
and off-site volumes are and off-site volumes are
handled differently. For handled differently. For
details, see “Controlling details, see “Controlling
reclamation of virtual reclamation of virtual
volumes” on page 374 and volumes” on page 374 and
“Reclamation of offsite “Reclamation of offsite
volumes” on page 375. volumes” on page 375.
File deletion Files are deleted: Files are deleted: Files are deleted:

v During inventory v Whenever the v During reclamation when

expiration processing, if the primary-copy of the file is inactive backup files are
files have expired deleted from the removed
primary-storage pool
v When a file space is (because of expiration, file v Whenever the primary
deleted space deletion, or volume copy of the file is deleted
deletion) from the primary-storage
v When a volume is deleted pool (because of expiration,
with the option to discard v When a volume is deleted file space deletion, or
the data with the option to discard volume deletion)
the data
v When a primary-storage v When a volume is deleted
pool volume is audited v When a copy-storage pool with the option to discard
with the FIX=YES option, if volume is audited with the the data
the files on the volume are FIX=YES option, if the files
damaged and no other on the volume are damaged v When an active-data pool
copies of the file exist volume is audited with the
FIX=YES option, if the files
on the volume are damaged

Deleting storage pools

Before deleting a storage pool, you need to take certain precautions to prevent data
loss and to ensure uninterrupted storage operations.

About this task

Task Required Privilege Class
Delete storage pools System

Before you delete a storage pool, ensure that:

v All volumes within the storage pool have been deleted
Ensure that you have saved any readable data that you want to preserve by
issuing the MOVE DATA command. Moving all of the data that you want to
preserve may require you to issue the MOVE DATA command several times.
Before you begin deleting all volumes that belong to the storage pool, change
the access mode of the storage pool to unavailable so that no files can be written
to or read from volumes in the storage pool.
See “Deleting storage pool volumes that contain data” on page 415 for
information about deleting volumes.

Chapter 10. Managing storage pools and volumes 413

 v The storage pool is not identified as the next storage pool within the storage
hierarchy
To determine whether this storage pool is referenced as the next storage pool
within the storage hierarchy, query for storage pool information as described in
“Monitoring space available in a storage pool” on page 382.
Update any storage pool definitions to remove this storage pool from the storage
hierarchy by performing one of the following:
– Naming another storage pool as the next storage pool in the storage hierarchy
– Entering the value for the NEXTSTGPOOL parameter as "" (double quotation
marks) to remove this storage pool from the storage hierarchy definition
See “Defining storage pools” on page 245 for information about defining and
updating storage pools.
v The storage pool to be deleted is not specified as the destination for any copy
group in any management class within the active policy set of any domain. Also,
a storage pool to be deleted cannot be the destination for space-managed files
(specified in any management class within the active policy set of any domain).
If this pool is a destination and the pool is deleted, operations fail because there
is no storage space to store the data.

To delete a storage pool, issue the DELETE STGPOOL command.

Deleting storage pool volumes

You can delete empty storage pool volumes or volumes that contain data from
primary storage pools, copy storage pools, or active-data pools. You can also delete
the client files that those volumes contain.

About this task

If files that are not cached are deleted from a primary storage pool volume, any
copies of these files in copy storage pools and active-data pools are deleted.

Files in a copy storage pool or an active-data pool are never deleted unless:
v Data retention is off, or the files have met their retention criterion.
v The volume that contains the copy file is deleted by using the
DISCARDDATA=YES option.
v A read error is detected by using AUDIT VOLUME with the FIX=YES option for
a copy storage pool volume or an active-data pool volume.
v The primary file is deleted because of:
– Policy-based file expiration
– File space deletion
– Deletion of the primary storage pool volume

You cannot delete a Centera volume if the data in the volume was stored using a
server with retention protection enabled and if the data has not expired.

Tip: If you are deleting many volumes, delete the volumes one at a time.
Concurrently deleting many volumes can adversely affect server performance.

Task Required Privilege Class

Delete volumes from any storage pool System or unrestricted storage

414 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 Task Required Privilege Class
Delete volumes from storage pools over Restricted storage
which they have authority

Deleting empty storage pool volumes

Use the DELETE VOLUME command to delete empty volumes. Before deleting a
volume, the server issues a confirmation message.

About this task

You can delete empty storage pool volumes. For example, to delete an empty
volume named WREN03, enter:
delete volume wren03

On an administrative client, you will receive the following confirmation messages,

unless the client is running with the NOCONFIRM option:
ANR2200W This command will delete volume WREN03
from its storage pool after verifying that the volume
contains no data.
Do you wish to proceed? (Y/N)

Volumes in a shred pool (DISK pools only) are not deleted until shredding is
completed. See “Securing sensitive client data” on page 539 for more information.

After you respond yes, the server generates a background process to delete the
volume.

The command can be run in the foreground on an administrative client by issuing

the command with the WAIT=YES parameter.

Deleting storage pool volumes that contain data

To prevent accidental deletion of backed-up, archived, or space-managed files, the
server does not allow you to delete a volume that contains user data unless you
specify DISCARDDATA=YES on the DELETE VOLUME command. After all files
have been deleted from the volume, the server deletes the volume from the storage
pool.

About this task

Tips:
1. The Tivoli Storage Manager server will not delete archive files that are on
deletion hold.
2. If archive retention protection is enabled, the Tivoli Storage Manager server
will delete only archive files whose retention period has expired.
3. Volumes in a shred pool (DISK pools only) are note deleted until the data on it
is shredded. See “Securing sensitive client data” on page 539 for more
information.

For example, to discard all data from volume WREN03 and delete the volume
from its storage pool, enter:
delete volume wren03 discarddata=yes

Chapter 10. Managing storage pools and volumes 415

 The server generates a background process and deletes data in a series of batch
database transactions. After all files have been deleted from the volume, the server
deletes the volume from the storage pool. If the volume deletion process is
canceled or if a system failure occurs, the volume might still contain data. Reissue
the DELETE VOLUME command and explicitly request the server to discard the
remaining files on the volume.

To delete a volume but not the files it contains, move the files to another volume.
See “Moving data from one volume to another volume” on page 400 for
information about moving data from one volume to another volume.

Residual data: Even after you move data, residual data may remain on the
volume because of I/O errors or because of files that were previously marked as
damaged. (Tivoli Storage Manager does not move files that are marked as
damaged.) To delete any volume that contains residual data that cannot be moved,
you must explicitly specify that files should be discarded from the volume.

416 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Part 3. Managing client operations
Installations of Tivoli Storage Manager typically include Backup-Archive Clients,
Tivoli Storage Manager for Space Management (HSM clients), and application
clients. You must register these clients with the server. Other priorities include
managing their access to the server, managing client data, and scheduling
operations such as backing up and archiving.

© Copyright IBM Corp. 1993, 2013 417

418 IBM Tivoli Storage Manager for Windows: Administrator's Guide
Chapter 11. Adding client nodes
When adding client nodes, the server views its registered clients as nodes that
require services and resources from the server.

When the Tivoli Storage Manager server is installed, the Tivoli Storage Manager
backup-archive client and the administrative client are installed on the same server
by default. However, many installations of Tivoli Storage Manager include remote
clients, and application clients on other servers, often running on different
operating systems.

The term “nodes” indicate the following type of clients and servers that you can
register as client nodes:
v Tivoli Storage Manager backup-archive clients
v Tivoli Storage Manager application clients, such as Tivoli Storage Manager for
Mail clients
v Tivoli Storage Manager for Space Management (HSM client)
v Tivoli Storage Manager source server registered as a node on a target server
v Network-attached storage (NAS) file server using NDMP support

Each node must be registered with the server and requires an option file with a
pointer to the server.

For details on many of the topics in this chapter, refer to the Backup-Archive Clients
Installation and User's Guide.
Related concepts:
“Accepting default closed registration or enabling open registration” on page 421
“Overview of clients and servers as nodes”
Related tasks:
“Installing client node software” on page 420
“Registering nodes with the server” on page 420
Related reference:
“Connecting nodes with the server” on page 425
“Comparing network-attached nodes to local nodes” on page 427

Overview of clients and servers as nodes

Each backup-archive client, Tivoli Storage Manager for Space Management (HSM
client), application client, and source server is given a node name when it is
registered as a node with the Tivoli Storage Manager server. The server considers
each as a node that requires services and resources from the server.

Typically a node is equivalent to a server, as in the case of a backup-archive client

that is installed on a user's computer for file system backups. However, multiple
nodes can exist on a single server. For example, a Structured Query Language
(SQL) server can contain both a Tivoli Storage Manager for SQL server application
client for database and transaction log backups, and a Tivoli Storage Manager
backup-archive client for file system backups.

© Copyright IBM Corp. 1993, 2013 419

Installing client node software
Administrators can install backup-archive clients, application clients, or Tivoli
Storage Manager for Space Management clients using three different methods.

About this task

The following are the methods for installing client node software:
v Install directly from the CD
v Transfer installable files from the CD to a target server
v Create client software images and install the images

You can also install using the silent installation technique.

For more information about installing:

v Client software, refer to Backup-Archive Clients Installation and User's Guide.
v System Storage Archive Manager application client software, refer to the
application client documentation for your particular client.
v Tivoli Storage Manager application client software, refer to the application client
documentation for your particular client.

Use the related procedures to configure a node after it is installed.

Registering nodes with the server

Administrators can register Tivoli Storage Manager clients, application clients, and
Tivoli Storage Manager for Space Management (HSM clients) as client nodes.

About this task

When a node is registered, Tivoli Storage Manager automatically creates an

administrative user ID with client owner authority over the node. You can use this
administrative user ID to access the Web backup-archive client from remote
locations through a Web browser. If an administrative user ID already exists with
the same name, an administrative user ID is not automatically defined.

Tip: You can connect to a Web backup-archive client directly from a supported
Web browser or from a hyperlink in the Web administrative Enterprise Console. To
do so, specify the node's URL and port number during the registration process or
update the node later with this information.
Related concepts:
“Overview of remote access to web backup-archive clients” on page 445

420 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Accepting default closed registration or enabling open
registration
Before a user can request Tivoli Storage Manager services, the node must be
registered with the server.

The administrator must register client nodes with the server when registration is
set to closed. Closed registration is the default.

Windows users can also use the Minimal Configuration option in the Initial
Configuration Task List.

Open registration allows the client nodes to register their node names, passwords,
and compression options. On UNIX and Linux systems, only the root user can
register a client node with the server.

With either registration mode, by default, an administrative user ID with client

owner authority is created over the node.

Important: Changes to the registration process do not affect existing registered

client nodes.

Adding nodes with closed registration

To add a node with closed registration, an administrator can issue the REGISTER
NODE command to register the node and specify the initial password.

About this task

The administrator can also specify the following optional parameters:

v Contact information.
v The name of the policy domain to which the node is assigned.
v Whether the node compresses its files before sending them to the server for
backup and archive.
v Whether the node can delete backups and archives from server storage.
v The name of a client option set to be used by the node.
v Whether to force a node to change or reset the password.
v The type of node being registered.
v The URL address used to administer the client node.
v The maximum number of mount points the node can use.
v Whether the client node keeps a mount point for an entire session.
v The transfer path used when the node sends data.
v The transfer path used when data is read for a client.
v Whether the server or client node initiates sessions.
v The IP address of the node.
v The low level address of the node.

Chapter 11. Adding client nodes 421

 Adding nodes with open registration
The server prompts the user, on their first attempt to connect to the server, for a
node name, password, and contact information. That information is added to a
node during open registration.

About this task

With open registration, the server automatically assigns the node to the
STANDARD policy domain. The server, by default, allows users to delete archive
copies, but not backups in server storage. Nodes are created with the default
authentication method that is defined on the server. Nodes are registered with the
default authentication method if it is defined on the server with the SET
DEFAULTAUTHENTICATION command. The default is LOCAL.

Procedure
1. Enable open registration by entering the following command from an
administrative client command line:
set registration open

For examples and a list of open registration defaults, see the Administrator's
Reference.
2. To change the defaults for a registered node, issue the UPDATE NODE command.

Node compression considerations

When you enable compression, it reduces network utilization and saves server
storage, but causes additional central processing unit (CPU) overhead to the node.
Data compression is recommended only when there is insufficient network
capacity.

About this task

To optimize performance or to ease memory constraints at the workstation, an

administrator can restrict file compression. You can select one of three options:
v Compress files.
v Do not compress files.
v Use the value set in the COMPRESSION option.
Set the COMPRESSION option in the client system options file or in the API
configuration file.
On a UNIX or a Linux system, a root user can define the COMPRESSION option in
the [Link] client options file.

Remember: Use either client compression or drive compression, but not both.
Related concepts:
“Data compression” on page 200

422 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Registering nodes with client options sets
Administrators can use client options sets in conjunction with the client options file
to register nodes with the server.

About this task

Client option sets are considered advanced implementation.

Procedure

Specify an option set for a node when you register or update the node. Issue the
following example command:
register node mike pass2eng cloptset=engbackup

The client node MIKE is registered with the password pass2eng. When the client
node MIKE performs a scheduling operation, the schedule log entries are kept for
5 days.
Related reference:
“Managing client option files” on page 464

Registering a network-attached storage file server as a node

To include a network-attached storage (NAS) file server as a node that Tivoli
Storage Manager can back up and restore with NDMP (network data management
protocol) operations, you can register the file server as a NAS node. Data that is
backed up from the NAS file server will be associated with the NAS node name.

About this task

The REGISTER NODE and UPDATE NODE commands have a default parameter of
TYPE=CLIENT.

Procedure

To register a NAS file server as a node, specify the TYPE=NAS parameter. Issue
the following command, which is an example, to register a NAS file server with a
node name of NASXYZ and a password of PW4PW:
register node nasxyz pw4pw type=nas

You must use this same node name when you later define the corresponding data
mover name.
Related reference:
Chapter 9, “Using NDMP for operations with NAS file servers,” on page 203

Chapter 11. Adding client nodes 423

 Registering a source server as a node on a target server
A virtual volume is a volume that appears to be a sequential media volume on a
source server. The volume is actually stored as an archive file on a target server.

About this task

To use virtual volumes, register the source server as a client node on the target
server.

The REGISTER NODE and UPDATE NODE commands have a default parameter of
TYPE=CLIENT.

Procedure

Register a source server as a node. Specify the TYPE=SERVER parameter.

Related tasks:
“Using virtual volumes to store data on another server” on page 732

Registering an API to the server

Workstation users can request IBM Tivoli Storage Manager services by using an
application that uses the Tivoli Storage Manager application programming
interface (API).

An administrator can issue the REGISTER NODE command to register the workstation
as a node.

Setting the compression option

There are several ways to determine the compression for applications that use the
IBM Tivoli Storage Manager application programming interface (API).

You can determine the compression by using one of the following methods:
v An administrator during registration who can:
– Require that files are compressed
– Restrict the client from compressing files
– Allow the application user or the client user to determine the compression
status
v The client options file. If an administrator does not set compression on or off,
Tivoli Storage Manager checks the compression status that is set in the client
options file. The client options file is required, but the API user configuration file
is optional.
v One of the object attributes. When an application sends an object to the server,
some object attributes can be specified. One of the object attributes is a flag that
indicates whether or not the data has already been compressed. If the
application turns this flag on during either a backup or an archive operation,
then Tivoli Storage Manager does not compress the data a second time. This
process overrides what the administrator sets during registration.

For more information on setting options for the API and on controlling
compression, see IBM Tivoli Storage Manager Using the Application Program Interface

424 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 Setting the file deletion option
An administrator can set the file deletion option for applications that use the Tivoli
Storage Manager application programming interface (API).

The administrator who sets the file deletion option can use the following methods:
v An administrator during registration
If an administrator does not allow file deletion, then an administrator must
delete objects or file spaces that are associated with the workstation from server
storage.
If an administrator allows file deletion, then Tivoli Storage Manager checks the
client options file.
v An application using the Tivoli Storage Manager API deletion program calls
If the application uses the dsmDeleteObj or dsmDeleteFS program call, then
objects or files are marked for deletion when the application is executed.

Connecting nodes with the server

The client options file connects each node to the server. Administrators and users
on all platforms can modify their client options file ([Link]) with a text editor.
Client options files can be updated differently across platforms.

On the Windows platform, you can use a wizard to work with the client options
file.

Important: If any changes are made to the [Link] file, the client must be restarted
for changes in the options file to have any affect.

The client options file [Link] is located in the client, application client, or host
server directory. If the file does not exist, copy the [Link] file. Users and
administrators can edit the client options file to specify:
v The network address of the server
v The communication protocol
v Backup and archive options
v Space management options
v Scheduling options
Related concepts:
“Creating or updating a client options file” on page 426

Required client options

Each node requires a client options file. Each client options file must contain the
network address of the IBM Tivoli Storage Manager server and other
communication options that allow the node to communicate with the server.

Figure 63 on page 426 shows the contents of a client options file that is configured
to connect to the server using TCP/IP. The communication options specified in the
client options file satisfy the minimum requirements for the node to connect to the
server.

Chapter 11. Adding client nodes 425

 Figure 63. Client options file

Many non-required options are available that can be set at any time. These options
control the behavior of Tivoli Storage Manager processing.

Refer to Backup-Archive Clients Installation and User's Guide for more information
about non-required client options.
Related concepts:
“Creating or updating a client options file”

UNIX and Linux client options

For UNIX and Linux, client options are located in the client systems options file,
client user options file, and the include-exclude options file.

Clients on other platforms use a single options file.

Updating the password for scheduling operations

Administrators can determine passwords when they register the nodes.

For nodes and host servers that run Windows, one more step is required.
Administrators must update passwords by using the IBM Tivoli Storage Manager
Scheduler Configuration Utility (DSMCUTIL). This utility allows administrators to
store Tivoli Storage Manager passwords in the Windows registry. After a password
is stored in the registry, the scheduler can run as a protected account under its own
authority. If the password expires, Tivoli Storage Manager automatically generates
a new password. To specify that the server provide a new password if the current
password expires, remove the asterisk from the following line in the client options
file: * passwordaccess generate.

Creating or updating a client options file

Creating or updating client options files depends on the client platform. You might
use a text editor, the client configuration wizard, or the client options file wizard.

426 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 Using a text editor to create or configure a client options file
All options files ([Link]) can be edited with a text editor. Anyone can edit the
client options file if they have access to the directory where the node software is
installed.

Editing individual options files is the most direct method, but may not be suitable
for sites with many client nodes.

Using the client configuration wizard to create or update a

client options file
When a local backup-archive client GUI starts initially and Tivoli Storage Manager
does not find an options file, a setup wizard guides the user through the
configuration process.

From the backup-archive client GUI, the client can also display the setup wizard
by selecting Utilities > Setup Wizard. The user can follow the panels in the setup
wizard to browse Tivoli Storage Manager server information in the Active
Directory. The user can determine which server to connect to and what
communication protocol to use.

Restriction: This wizard is not available for the Web client.

Comparing network-attached nodes to local nodes

A Tivoli Storage Manager environment can be either a server and client on the
same server (stand-alone environment) or a server and network-attached clients
(network environment).

The stand-alone environment of Tivoli Storage Manager consists of a

backup-archive client and an administrative client on the same computer as the
server. There is nothing more to do to connect the client. This is shown in
Figure 64.

Backup-archive client

Server

Administrative client

Figure 64. Stand-alone environment

Figure 65 on page 428 shows that a network environment Tivoli Storage Manager
consists of a backup-archive client and an administrative client on the same
computer as the server. However, network-attached client nodes can also connect
to the server.

Chapter 11. Adding client nodes 427

 Backup-archive client
Backup-archive client

Server

Administrative client

Application client

Figure 65. Network environment

Each client requires a client options file. A user can edit the client options file at
the client node. The options file contains a default set of processing options that
identify the server, communication method, backup and archive options, space
management options, and scheduling options.

Adding clients through the administrative command line client

The administrator can register nodes by issuing the REGISTER NODE command. For
more information, refer to the Administrator's Reference.

Enabling open registration

The default registration mode at installation is closed.

Procedure

To change the default to open so users can register their own client nodes, issue
the following command:
set registration open

This is true unless you use the Minimal Configuration option.

Example: registering three client nodes using the

administrative command line
For this example, the goal is to register three workstations from the engineering
department and assign them to the ENGPOLDOM policy domain.

About this task

Before you can assign client nodes to a policy domain, the policy domain must
exist.

You want to let users delete backed up or archived files from storage pools. From
an administrative client, you can use the macro facility to register more than one
client node at a time.

428 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Procedure
1. Create a macro file named [Link], that contains the following REGISTER
NODE commands:
register node ssteiner choir contact=’department 21’
domain=engpoldom archdelete=yes backdelete=yes
register node carolh skiing contact=’department 21, second shift’
domain=engpoldom archdelete=yes backdelete=yes
register node mab guitar contact=’department 21, third shift’
domain=engpoldom archdelete=yes backdelete=yes
2. Issue the MACRO command.
macro [Link]

For information on the MACRO command, see the Administrator's Reference.

Related concepts:
Chapter 13, “Implementing policies for client data,” on page 475

Chapter 11. Adding client nodes 429

430 IBM Tivoli Storage Manager for Windows: Administrator's Guide
Chapter 12. Managing client nodes
If you have already installed and configured your client nodes, you must be able
to manage and control their access to the server.

The Tivoli Storage Manager server views its registered clients, application clients,
and source servers as nodes. The term “client node” refers to the following type of
clients and servers:
v Tivoli Storage Manager backup-archive clients
v Tivoli Storage Manager application clients, such as Tivoli Storage Manager for
Mail clients
v Tivoli Storage Manager source servers registered as nodes on a target server
v Network-attached storage (NAS) file servers using network data management
protocol (NDMP) support
Related concepts:
“Accepting default closed registration or enabling open registration” on page 421
“Overview of clients and servers as nodes” on page 419
Related tasks:
“Installing client node software” on page 420
“Registering nodes with the server” on page 420
Related reference:
“Connecting nodes with the server” on page 425
“Comparing network-attached nodes to local nodes” on page 427

Managing nodes
From the perspective of the server, each client and application client is a node
requiring IBM Tivoli Storage Manager services.

Client nodes can be local or remote to the server.

Administrators can perform the following activities when managing client nodes.

Task Required Privilege Class

Updating, renaming, locking, or unlocking System or unrestricted policy
any client nodes
Updating, renaming, locking, or unlocking System, unrestricted policy, or restricted
client nodes assigned to specific policy policy for those domains
domains
Displaying information about client nodes or Any administrator
file spaces
Deleting any client nodes System or unrestricted policy
Removing client nodes assigned to specific System, unrestricted policy, or restricted
policy domains policy for those domains
Managing client access authority levels System

Related reference:
“Defining client nodes and file spaces” on page 451

© Copyright IBM Corp. 1993, 2013 431

 “Comparing network-attached nodes to local nodes” on page 427

Managing client nodes across a firewall

In most cases, the IBM Tivoli Storage Manager server and clients can work across a
firewall or the server can securely manage client backup and restore operations
and administrative functions across a firewall. Because every firewall is different,
the firewall administrator might need to consult the instructions for the firewall
software or hardware in use.

IBM Tivoli Storage Manager has two methods for enabling communication
between the client and the server across a firewall: client-initiated communication
and server-initiated communication. To allow either client-initiated or
server-initiated communication across a firewall, client options must be set in
concurrence with server parameters on the REGISTER NODE or UPDATE NODE
commands. Enabling server-initiated communication overrides client-initiated
communication, including client address information that the server may have
previously gathered in server-prompted sessions.

Attention: A firewall should not be configured in a manner that causes

terminations of sessions in use by either the server or the storage agent. When a
firewall terminates a valid session, unpredictable problems can occur which make
processes and sessions appear to stop due to communication input/output (I/O).
Configuring IBM Tivoli Storage Manager components with known ports helps
when you want to exclude IBM Tivoli Storage Manager sessions from timeout
restrictions.

Client-initiated sessions
You can enable clients to communicate with a server across a firewall by opening
the TCP/IP port for the server and modifying the [Link] file.

Procedure
1. To enable clients to communicate with a server across a firewall, open the
TCP/IP port for the server on the TCPPORT option in the [Link] file. The
default TCP/IP port is 1500. When authentication is turned on, the information
that is sent over the wire is encrypted.
2. To enable administrative clients to communicate with a server across a firewall,
open the TCP/IP port for the server on the TCPADMINPORT option in the
[Link] file. The default TCP/IP port is the TCPPORT value. When
authentication is turned on, the information that is sent over the wire is
encrypted. See the Backup-Archive Clients Installation and User's Guide for more
information.

Results
1. If the TCPADMINPORT option is specified, sessions from clients without
administration authority can be started on the TCPPORT port only. If the server
[Link] specifies TCPADMINPORT that is different from the TCPPORT and sets
ADMINONCLIENTPORT to NO, then administrative client sessions can be started on
the TCPADMINPORT port only.
2. You can specify either IPv4 or IPv4/IPv6 in the COMMMETHOD option when you
start the server, storage agent, client, or API application. The same port
numbers are used by the server, storage agent, client, or API application for
both IPv4 and IPv6.

432 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 IPv6 address formats are acceptable for all functions that support IPv6.
However, if you use IPv6 addresses for functions that do not support IPv6,
communications fail. The following functions do not support IPv6:
v Network data management protocol (NDMP): backing up and restoring
storage pools, copying and moving data
v Automated cartridge system library software (ACSLS)
v Simple network management protocol (SNMP)
v Centera device support
v Shared memory protocol
v Windows Microsoft Management Console functions
v Tivoli Enterprise Console® support

Remember: You can continue to use IPv4 address formats for the following
functions:
v NDMP: backing up and restoring storage pools, copying and moving data
v ACSLS
v SNMP
v Centera device support
v Shared memory protocol
v Windows Microsoft Management Console functions

If you select the CLIENTORSERVER option of the SESSIONINITIATION server parameter,

the client may start sessions with the server. Or, server-prompted scheduling may
be used to prompt the client to connect to the server.

Server-initiated sessions
To limit the start of backup-archive client sessions to the IBM Tivoli Storage
Manager server, specify the SESSIONINITIATION parameter on the server. You must
also synchronize the information in the client option file.

About this task

In either the REGISTER NODE or UPDATE NODE command, select the SERVERONLY option
of the SESSIONINITIATION parameter. Provide the HLADDRESS and LLADDRESS
client node addresses. For example,
register node fran secretpw hladdress=[Link] lladdress=1501
sessioninitiation=serveronly

The HLADDRESS specifies the IP address of the client node, and is used whenever the
server contacts the client. The LLADDRESS specifies the low level address of the
client node and is used whenever the server contacts the client. The client node
listens for sessions from the server on the LLADDRESS port number.

If SESSIONINITIATION=SERVERONLY for a node defined on the IBM Tivoli

Storage Manager server, the client must have SESSIONINITIATION=SERVERONLY
in its option file. In addition, the TCP/IP address of the client must correspond to
the information supplied with the HLADDRESS server parameter. Finally,
TCPCLIENTPORT in the client option file must correspond to the information
supplied with the LLADDRESS server parameter, or the server will not know how to
contact the client.

Restriction: Server-initiated sessions cannot utilize SSL (Secure Sockets Layer) or

LDAP authenticated passwords.

Chapter 12. Managing client nodes 433

 Note:
1. If you switch from server-prompted to server-initiated sessions, the server will
discard any addressing information it had and will use only the information
from the HLADDRESS and LLADDRESS parameters of the REGISTER NODE and UPDATE
NODE commands in contacting the client.
2. The server uses Domain Name System (DNS) to determine the name of client
nodes. If your DNS is not correctly configured, there may be delays or failures
in looking up names. The DNSLOOKUP option is available to restrict usage of DNS
services. It can be changed dynamically through the SETOPT DNSLOOKUP
command.
3. Server-initiated sessions cannot utilize SSL (Secure Sockets Layer) or LDAP
authenticated passwords.
Table 38. Server-Initiated sessions
Setting or parameter on the Location on the IBM
IBM Tivoli Storage Tivoli Storage Manager Must match this on the
Manager server: server client: Location on the client
SESSIONINITIATION= REGISTER NODE or UPDATE SESSIONINITIATION= client option file
SERVERONLY NODE command SERVERONLY
HLADDRESS REGISTER NODE or UPDATE TCP/IP address TCP/IP address
NODE command
LLADDRESS REGISTER NODE or UPDATE TCPCLIENTPORT client option file
NODE command

Updating client node information

You can issue the UPDATE NODE command to update information such as the client's
assigned policy domain, the user's password or contact information, and the client
option set used by the node.

Procedure

Update client node TOMC to prevent it from deleting archived files from storage
pools by entering the following example command:
update node tomc archdelete=no

Command-line backup-archive client automatic deployment

feature: overview
The backup-archive client automatic deployment feature helps you schedule
updates to one or more backup-archive clients with deployment packages. The
deployment packages can be backup-archive client updates or full releases.

The backup-archive client deployment feature can simultaneously update multiple

backup-archive clients, but it does not install new backup-archive clients. The
deployment is configured and scheduled from the administrative command line.
For more information about the client deployment feature, see Technote 1653433 at:
[Link]

The following list shows the deployment restrictions:

v Windows IBM Tivoli Storage Manager backup-archive clients that are at levels
earlier than 5.4 cannot be updated with the available deployment packages.
v Operating systems that are not supported, such as Windows 2000, cannot be
upgraded through the deployment process.

434 IBM Tivoli Storage Manager for Windows: Administrator's Guide

v Do not schedule automatic client deployments to workstations that have an IBM
Tivoli Storage Manager for Enterprise Resource Planning application installed on
them.
At deployment time, a dependency check is performed to ensure that the API
library does not conflict with the client package that is currently installed. IBM
Tivoli Storage Manager for Enterprise Resource Planning applications do not use
the same installation technology that the client installation program uses. The
API library that is installed might not be compatible with the API library that
was installed by the IBM Tivoli Storage Manager for Enterprise Resource
Planning application. The newly deployed API library can cause the IBM Tivoli
Storage Manager for Enterprise Resource Planning applications to fail.

To use the feature, the backup-archive client must meet these requirements:
v The PASSWORDACCESS option must be set to generate.
v The client acceptor (CAD) or backup-archive client schedule must be running at
the time of the deployment. The backup-archive client is deployed from the
server as a scheduled task.
v The backup-archive client must meet the following disk space requirements:
Table 39. Disk space required to deploy a backup-archive client package
Operating system Total required disk space
AIX 1500 MB
Solaris 1200 MB
HP-UX 900 MB
Mac OS X 200 MB
Linux x86/x86 64 950 MB
Windows 2 GB

Important: You must use the SET SERVERHLADDRESS command for all automatic
client deployments.

You can find the deployment packages in the maintenance directory on the FTP
site: [Link]
client.

Using the command-line interface to configure the server for a

backup-archive client deployment
You can access the backup-archive client automatic deployment feature from the
command-line interface and schedule client deployments for packages that you
acquire from the FTP site.

The following list outlines the deployment requirements:

v Before using the backup-archive client deployment feature, you must configure
the server.
v You must be an administrator on the server to issue backup-archive deployment
commands related to configuring the server and importing the deployment
packages.
v To schedule a backup-archive client deployment for a node, you must have
system privileges or domain privileges on the domain to which the node
belongs.

Chapter 12. Managing client nodes 435

 The following example command can be used to configure the server to deploy
backup-archive client packages with the command-line interface:

define devclass ibm_client_deploy_import devtype=file directory=import_directory

define stgpool stgpool_name storage_dc_name maxscratch=20

define domain ibm_client_deploy

define policyset ibm_client_deploy ibm_client_deploy

define mgmtclass ibm_client_deploy ibm_client_deploy ibm_client_deploy

define copygroup ibm_client_deploy ibm_client_deploy ibm_client_deploy

standard type=archive destination=stgpool_name retver=retention_value

assign defmgmtclass ibm_client_deploy ibm_client_deploy ibm_client_deploy

activate policyset ibm_client_deploy ibm_client_deploy

set serverhladdress=[Link]

where:
v ibm_client_deploy_import is the temporary location from where the deployment packages
are imported. This parameter is defined by the deployment manager.
v import_directory is a previously defined directory that is accessible from the server.
v stgpool_name is the name of a storage pool of your choosing where the deployment
packages are stored on the server. The storage pool name is based on a previously
defined device class. That device class is different from the one which is used to perform
IMPORT operations.
v storage_dc_name represents the device class where the deployment packages are stored on
the server.
v retention_value (RETVER) of the DEFINE COPYGROUP command sets the retention time for the
package. You can set it to NOLimit or to a number of days. The default for the
Administration Center is five years.

Important: The retention value must be set to a value that includes the amount of time
that the package was on the FTP site. For example, if a deployment package is on the FTP
site for 30 days, the retention value for the copy group must be greater than 30 days. If
not, the package expires when the next EXPIRE INVENTORY command is issued.
v [Link] is the server IP address or host name from which you scheduled
the client automatic deployment.

Importing the target level to the server

The backup-archive client deployment packages must be imported to the servers
after you acquire them from the FTP site.

Before you begin

Ensure that you configure the server for backup-archive client automatic
deployments before you import the packages.

Procedure

Complete the following steps to import deployment packages to the server:

1. Download the backup-archive client deployment packages. If you are working
outside of the Administration Center, you can download the packages from the
FTP site: [Link]

436 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 maintenance/client. Download the client deployment packages from the FTP
site to the Tivoli Storage Manager server or to an internal FTP site.
2. Issue the IMPORT command. For example:
import node IBM_CLIENT_DEPLOY_WIN devclass=upgradedev
filedata=archive mergefilespace=yes
volumenames=[Link]

For Macintosh, issue the following command:

import node IBM_CLIENT_DEPLOY_MAC devclass=upgradedev
filedata=archive mergefilespace=yes replace=all
volumenames=[Link]

where:
upgradedev is the file device class name.
[Link] is the deployment package name. You can also use a
comma-separated list of package names.
If you want to view the progress, issue the QUERY PROCESS command.
3. Verify that the packages are in a location that the server can reach. Enter the
following command:
select * from ARCHIVES where node_name=’IBM_CLIENT_DEPLOY_WIN’

For Macintosh, issue the following command:

select * from ARCHIVES where node_name=’IBM_CLIEN_DEPLOY_MAC’

where ARCHIVES is the type of file that is imported through the IMPORT NODE
command.
Related reference:
“Using the command-line interface to configure the server for a backup-archive
client deployment” on page 435

Defining a schedule for an automatic deployment

You can schedule a backup-archive client deployment after you import the
packages to the server.

About this task

Complete the following steps to create a schedule for an automatic client

deployment:

Procedure
1. Issue the DEFINE SCHEDULE command. For example:
define schedule domain_name schedule_name action=deploy
objects="deployment_package_location destination_for_package"
options="-fromnode=IBM_CLIENT_DEPLOY_WIN
-postnschedulecmd=’relative_path\[Link] schedule=schedule_name
autodeploy=noreboot nodeinfo=TBD’ -subdir=yes -replace=all"
startdate=4/20/2013 perunits=onetime

where
deployment_package_location is the path to the deployment package
destination_for_package is the path to where you want to store the
deployment package
IBM_CLIENT_DEPLOY_WIN is the predefined name (for a Windows
deployment) for the -fromnode option

Chapter 12. Managing client nodes 437

 AUTODEPLOY can be YES, NO, or NOREBOOT. The default is YES.
nodeinfo=TBD must be entered exactly as shown.

Note: The -fromnode option can have the following values:

IBM_CLIENT_DEPLOY_WIN
IBM_CLIENT_DEPLOY_UNX
IBM_CLIENT_DEPLOY_MAC
If your current backup-archive client is AIX, Linux, Solaris, or HP-UX and is at
version 6.1 or later, use nodeinfo2=TBD in the POSTNSCHEDULECMD command.
Macintosh Backup-Archive Clients at version 5.5 and later also must use
nodeinfo2=TBD.
The target location for the backup-archive client packages is in relation to the
location of the working client scheduler directory. If the directory is C:\program
files\tivoli\tsm\baclient, the client deployment package target location is
C:\program files\tivoli\tsm\IBM_ANR_WIN\v6300\. This location is also the
deployment manager’s working directory. Here is an example post-schedule
command for Windows servers:
-postnschedulecmd=’..\relative_path\[Link]
schedule=schedule_name domain=domain_name autodeploy=noreboot nodeinfo=TBD’
The target location for backup-archive client packages on Macintosh operating
systems is in relation to the location of the working client scheduler directory.
This location is also the deployment manager’s working directory. Here is an
example post-schedule command for Macintosh:
-postnschedulecmd=’../relative_path/[Link]>
./IBM_ANR_MAC/V6303/upmin#schedule=schedule_name#domain=standard#nodeinfo2=TBD’

Important: If you use AUTODEPLOY=NOREBOOT in your command, the client

operating system does not restart after the deployment completes. Restarting
can affect any critical applications that are running on the client operating
system. Applications that are not Tivoli Storage Manager applications that are
running on the client operating system might not be restarted automatically. In
most cases, the installation can complete without restarting the client operating
system. There are rare cases where the deployment manager cannot detect the
restart.
For example, if client processes are started from a script. In these cases, the new
backup-archive client installation continues, but a manual restart of the client
computer is required.
2. Issue the DEFINE ASSOCIATION command to associate nodes with the
deployment schedule.

Verifying the server for backup-archive client automatic

deployments
You can verify that the server is configured to automatically update the
backup-archive clients.

About this task

You can verify that the server is configured to manage deployment packages for
the backup-archive client.

Procedure

Complete the following steps to verify that the server is configured to manage
deployment packages:

438 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 1. Verify that the IBM_CLIENT_DEPLOY_IMPORT device class is defined by
issuing QUERY DEVCLASS IBM_CLIENT_DEPLOY_IMPORT F=D.
2. Verify that a storage pool is defined to store the deployment packages on the
server by issuing QUERY STGPOOL stgpool_name F=D.
3. Verify that an archive copy group is defined by issuing QUERY COPYGROUP
IBM_CLIENT_DEPLOY IBM_CLIENT_DEPLOY IBM_CLIENT_DEPLOY TYPE=ARCHIVE.
4. Verify that the default management class is set to the active policy set by
issuing QUERY POLICYSET IBM_CLIENT_DEPLOY F=D.
5. Verify that the schedule is created for your client nodes by issuing QUERY EVENT
domain_name sched_name F=D.
6. Verify that the targetlevel is correctly updated to match the level of the
deployment package by issuing QUERY NODE node_name F=D.

Renaming client nodes

You can rename a client node by issuing the RENAME NODE command. You might
need to rename a client node if the workstation network name or host name
changes. For example, with UNIX and Linux clients, users define their node name
based on the value returned by the HOSTNAME command.

About this task

When users access the server, their IBM Tivoli Storage Manager user IDs match the
host name of their workstations. If the host name changes, you can update a client
node user ID to match the new host name.

Procedure

To rename CAROLH to ENGNODE, issue the following example command:

rename node carolh engnode

ENGNODE retains the contact information and access to back up and archive data
that belonged to CAROLH. All files backed up or archived by CAROLH now
belong to ENGNODE.

Results

If you rename a node that authenticates with an LDAP directory server, names for
same-named nodes on other servers that share namespace are not renamed. You
must issue a RENAME command for each node. If you want to keep the nodes in
sync, change their name to match the new name. If you do not, the node on the
other server can no longer authenticate with the LDAP directory server if you
specify SYNCLDAPDELETE=YES.

Rename all nodes starship that authenticate to LDAP directory

servers

If you have a node that shares namespace on an LDAP directory server with other
nodes, you can rename each node. The renaming must, however, be done on each
server. For example, you can issue the following command on each server:
rename node starship moonship syncldapdelete=yes

The node starship, that authenticates to an LDAP directory server, changes their
name to moonship. With SYNCLDAPDELETE=YES, the entry on the LDAP directory
server changes to moonship and removes node starship from the LDAP server.

Chapter 12. Managing client nodes 439

 Therefore, other servers cannot authenticate node starship with the LDAP server.
You can register node starship with the LDAP server, or rename node starship to
moonship.

Locking and unlocking client nodes

You can prevent client nodes from accessing the server with the LOCK NODE
command. This prevents client nodes from performing functions such as either
backup and restore or archive and retrieve.

About this task

You can restore a locked node’s access to the server with the UNLOCK NODE
command.

Procedure
1. To prevent client node MAB from accessing the server, issue the following
example command:
lock node mab
2. To let client node MAB access the server again, issue the following example
command:
unlock node mab
Related tasks:
“Disabling or enabling access to the server” on page 471

Deleting client nodes

You can delete a client node from the server with the REMOVE NODE command. All
file spaces that belong to the client node must first be deleted from server storage.
After all of the client node's file spaces are deleted, you can delete the node.

Before you begin

Before you can delete a network-attached storage (NAS) node, you must first
delete any file spaces, then delete any defined paths for the data mover with the
DELETE PATH command. Delete the corresponding data mover with the DELETE
DATAMOVER command. Then you can issue the REMOVE NODE command to delete the
NAS node.

Procedure

Remove client node DEBBYG by completing the following example commands:

1. Issue the following command:
delete filespace debbyg * type=any
2. To delete the DEBBYG node, issue the following command:
remove node debbyg
Related tasks:
“Deleting file spaces” on page 463

440 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Consolidating multiple clients under a single client node name
Backups of multiple nodes which share storage can be consolidated to a common
target node name on the Tivoli Storage Manager server. For example, several nodes
in an IBM General Parallel File System (GPFS™) cluster, NODE_1, NODE_2, and
NODE_3, can back up to the same node (NODE_OLIV) on the server.

This is useful when the server responsible for performing the backup may change
over time, such as with a cluster. Consolidating shared data from multiple servers
under a single name space on the Tivoli Storage Manager server means that the
directories and files can be easily found when restore operations are required.
Backup time can be reduced and clustered configurations can store data with
proxy node support. Client nodes can also be configured with proxy node
authority to support many of the systems which support clustering failover.

By granting client nodes proxy node authority to another node, you gain the
ability to backup, archive, migrate, restore, recall, and retrieve shared data on
multiple clients under a single node name on the Tivoli Storage Manager server.
When authorized as agent nodes, Tivoli Storage Manager nodes and Tivoli Storage
Manager for Space Management (HSM) clients can be directed to backup or restore
data on behalf of another node (the target node).

Distributing workloads to reduce backup and restore time

Concurrent operations can reduce backup and restore times in environments such
as IBM General Parallel File System (GPFS). Conventional cluster backups are
unable to do this with very large file systems because if a password is changed by
the Tivoli Storage Manager server, it takes some time to communicate that change
to all the nodes.

Administrators must then create scripts that change the passwords manually before
they expire. Using proxy node support, it is possible to break up a large GPFS into
smaller units for backup purposes and not have password coordination issues.

The following example shows how scheduling would work where workload is
distributed, for example in the DB2 Universal Database Enterprise Extended
Edition (EEE) environment. In this example, NODE_A, NODE_B and NODE_C all
work together to back up this distributed environment, all acting on behalf of
NODE-Z. NODE_A directs the backup for all three physical servers. NODE_A
either has ASNODENAME=NODE_Z in its local options file or the server (through
the DEFINE SCHEDULE command) has indicated that NODE_A needs to request
proxy authority to NODE_Z. See the Backup-Archive Clients Installation and User's
Guide for more information on the ASNODENAME client option.

An administrator can define the schedule that does a DB2 UDB EEE backup on
behalf of NODE_Z by issuing the following command:
DEFINE SCHEDULE STANDARD BACKUP-SCHED ACTION=INCREMENTAL
OPTIONS=’-ASNODENAME=NODE_Z’

Define the association of the schedule to the three nodes:

DEFINE ASSOCIATION STANDARD BACKUP-SCHED NODE_A

Chapter 12. Managing client nodes 441

 Defining agent and target nodes
Target client nodes own data and agent nodes act on behalf of the target nodes.
When granted proxy node authority to a target node, an agent node can perform
backup and restore operations for the target node. Data that the agent node stores
on behalf of the target node is stored under the target node's name in server
storage.

Agent nodes are considered traditional nodes in that there is usually a one-to-one
relationship between a traditional node and a physical server. A target node can be
a logical entity, meaning no physical server corresponds to the node. Or, it can be a
predefined node which corresponds to a physical server.

By using the GRANT PROXYNODE command, you can grant proxy node authority to all
nodes sharing data in the cluster environment to access the target node on the
Tivoli Storage Manager server. QUERY PROXYNODE displays the nodes to which a
proxy node relationship was authorized. See the Administrator's Reference for more
information about these commands.

Relationships between agent nodes and target nodes:

v A client node can be both an agent and a target at the same time. For example,
NODE_A is an agent node for NODE_Z. NODE_A can also act as a target node
for agent node NODE_D.
v Proxy node relationships can work conversely. For example, NODE_A can be an
agent node for NODE_Z, and NODE_Z can be an agent node for NODE_A.
v Proxy node relationships cannot be inherited. For example, if NODE_A is an
agent for NODE_X, and NODE_X is an agent for NODE_Z, NODE_A is not
automatically an agent for NODE_Z.

Proxy node relationships will not be imported by default; however, the associations
can be preserved by specifying the PROXYNODEASSOC option on the IMPORT NODE and
IMPORT SERVER commands. Exporting to sequential media maintains proxy node
relationships, but exporting to a server requires specifying the PROXYNODEASSOC
option on EXPORT NODE and EXPORT SERVER.

Important:
v If a proxy node relationship is authorized for incompatible file spaces, there is a
possibility of data loss or other corruption.
v Central command routing or importing of the GRANT PROXYNODE and REVOKE
PROXYNODE commands can create access issues.
v The maximum number of mount points for agent nodes should be increased to
allow parallel backup operations across the target nodes.

Configuring shared access example

Shared access must be configured according to the nodes on the server and the
relationships between the nodes.

About this task

The following example shows how to set up proxy node authority for shared
access. In the example, client agent nodes NODE_1, NODE_2, and NODE_3 all
share the same General Parallel File System (GPFS). Because the file space is so
large, it is neither practical nor cost effective to back up this file system from a
single client node. By using Tivoli Storage Manager proxy node support, the very
large file system can be backed up by the three agent nodes for the target

442 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 NODE_GPFS. The backup effort is divided among the three nodes. The end result
is that NODE_GPFS has a backup from a given point in time.

All settings used in the proxy node session are determined by the definitions of the
target node, in this case NODE_GPFS. For example, any settings for
DATAWRITEPATH or DATAREADPATH are determined by the target node, not
the agent nodes (NODE_1, NODE_2, NODE_3).

Assume that NODE_1, NODE_2 and NODE_3 each need to execute an incremental
backup and store all the information under NODE_GPFS on the server.

Procedure

Perform the following steps to set up a proxy node authority for shared access:
1. Define four nodes on the server: NODE_1, NODE_2, NODE_3, and
NODE_GPFS. Issue the following commands:
register node node_1 mysecretpa5s
register node node_2 mysecret9pas
register node node_3 mypass1secret
register node node_gpfs myhiddp3as
2. Define a proxy node relationship among the nodes by issuing the following
commands:
grant proxynode target=node_gpfs agent=node_1,node_2,node_3
3. Define the node name and asnode name for each of the servers in the
respective [Link] files. See the Backup-Archive Clients Installation and User's
Guide for more information on the NODENAME and ASNODENAME client options.
Issue the following commands:
nodename node_1
asnodename node_gpfs
4. Optionally, define a schedule:
define schedule standard gpfs-sched action=macro options="gpfs_script"
5. Assign a schedule to each client node by issuing the following commands:
define association standard gpfs-sched node_1
define association standard gpfs-sched node_2
define association standard gpfs-sched node_3
6. Execute the schedules by issuing the following command:
dsmc schedule

Displaying information about client nodes

You can display information about client nodes in different aspects.

For example, as a policy administrator, you might query the server about all client
nodes assigned to the policy domains for which you have authority. Or you might
query the server for detailed information about one client node.

Chapter 12. Managing client nodes 443

 Displaying information about client nodes assigned to specific
policy domains
You can display information about client nodes that are assigned to specific policy
domains.

Procedure

Issue the following command to view information about client nodes that are
assigned to the STANDARD and ENGPOLDOM policy domains:
query node * domain=standard,engpoldom

The data from that command might display similar to the following output:
Node Name Platform Policy Domain Days Since Days Since Locked?
Name Last Password
Access Set
---------- -------- -------------- ---------- ---------- -------
JOE WinNT STANDARD 6 6 No
ENGNODE AIX ENGPOLDOM <1 1 No
HTANG Mac STANDARD 4 11 No
MAB AIX ENGPOLDOM <1 1 No
PEASE Linux86 STANDARD 3 12 No
SSTEINER SOLARIS ENGPOLDOM <1 1 No

Displaying information about a specific client node

You can view information about specific client nodes.

Procedure

For example, to review the registration parameters defined for client node JOE,
issue the following command:
query node joe format=detailed

The resulting report may appear similar to the following output:

444 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 Node Name: JOE
Platform: WinNT
Client OS Level: 5.00
Client Version: Version 5, Release 1, Level 5.0
Policy Domain Name: STANDARD
Last Access Date/Time: 05/19/2002 [Link]
Days Since Last Access: 6
Password Set Date/Time: 05/19/2002 [Link]
Days Since Password Set: 6
Invalid Sign-on Count: 0
Locked?: No
Contact:
Compression: Client’s Choice
Archive Delete Allowed?: Yes
Backup Delete Allowed?: No
Registration Date/Time: 03/19/2002 [Link]
Registering Administrator: SERVER_CONSOLE
Last Communication Method Used: Tcp/Ip
Bytes Received Last Session: 108,731
Bytes Sent Last Session: 698
Duration of Last Session (sec): 0.00
Pct. Idle Wait Last Session: 0.00
Pct. Comm. Wait Last Session: 0.00
Pct. Media Wait Last Session: 0.00
Optionset:
URL: [Link]
Node Type: Client
Password Expiration Period: 60
Keep Mount Point?: No
Maximum Mount Points Allowed: 1
Auto Filespace Rename: No
Validate Protocol: No
TCP/IP Name: JOE
TCP/IP Address: [Link]
Globally Unique ID: 11.9c.54.e0.8a.b5.11.d6.b3.c3.[Link].c1.5b
Transaction Group Max: 0
Data Write Path: ANY
Data Read Path: ANY
Session Initiation: ClientOrServer
High-level Address: [Link]
Low-level Address: 1501
Collocation Group Name: minster
Proxynode Target: node_gpfs
Proxynode Agent: node_1
Node Groups:

Overview of remote access to web backup-archive clients

With the introduction of the web backup-archive client, when a client node is
registered with an IBM Tivoli Storage Manager server, an identical administrative
user ID is created at the same time. This user ID has client owner authority over
the node by default.

Restriction: Administrators who authenticate their password with an LDAP

directory server cannot access the web client.

A web backup-archive client can be accessed from a web browser or opened from
the Operations Center interface. This allows an administrator with the proper
authority to perform backup, archive, restore, and retrieve operations on any
server that is running the web backup-archive client.

You can establish access to a web backup-archive client for help desk personnel
that do not have system or policy privileges by granting those users client-access
authority to the nodes that they must manage. Help desk personnel can then
perform activities on behalf of the client node such as backup and restore
operations.

Chapter 12. Managing client nodes 445

 A native backup-archive client can log on to Tivoli Storage Manager using their
node name and password, or administrative user ID and password. The
administrative user ID password is managed independently from the password
that is generated with the passwordaccess generate client option. The client must
have the option passwordaccess generate specified in their client option file to
enable use of the web backup-archive client.

To use the web backup-archive client from your web browser, specify the URL and
port number of the Tivoli Storage Manager backup-archive client computer that is
running the web client. The browser that you use to connect to a web
backup-archive client must be Microsoft Internet Explorer 5.0 or Netscape 4.7 or
later. The browser must have the Java Runtime Environment (JRE) 1.3.1, which
includes the Java Plug-in software. The JRE is available at http://
[Link]/.

During node registration, you have the option of granting client owner or client
access authority to an existing administrative user ID. You can also prevent the
server from creating an administrative user ID at registration. If an administrative
user ID exists with the same name as the node that is being registered, the server
registers the node but does not automatically create an administrative user ID. This
process also applies if your site uses open registration.

For more information about installing and configuring the web backup-archive
client, refer to Backup-Archive Clients Installation and User's Guide.

Defining node privilege class and client access authorities

Access to a Web backup-archive client requires either client owner authority or
client access authority.

Administrators with system or policy privileges over the client node's domain,
have client owner authority by default. The administrative user ID created
automatically at registration has client owner authority by default. This
administrative user ID is displayed when an administrator issues a QUERY ADMIN
command.

The following definitions describe the difference between client owner and client
access authority when defined for a user that has the node privilege class:
Client owner
You can access the client through the Web backup-archive client or
native backup-archive client.
You own the data and have a right to physically gain access to the data
remotely. You can backup and restore files on the same or different
servers, you can delete file spaces or archive data.
The user ID with client owner authority can also access the data from
another server using the –NODENAME or -VIRTUALNODENAME parameter.
The administrator can change the client node's password for which they
have authority.
This is the default authority level for the client at registration. An
administrator with system or policy privileges to a client's domain has
client owner authority by default.
Client access
You can only access the client through the Web backup-archive client.
You can restore data only to the original client.

446 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 A user ID with client access authority cannot access the client from
another server using the –NODENAME or –VIRTUALNODENAME \parameter.
This privilege class authority is useful for help desk personnel so they
can assist users in backing up or restoring data without having system
or policy privileges. The client data can only be restored to none other
than the original client. A user ID with client access privilege cannot
directly access client's data from a native backup-archive client.

Managing client access authority levels

By default, an administrator with system or policy privilege over a client's domain
can remotely access clients and perform backup and restore operations.

You can grant client access or client owner authority to other administrators by
specifying CLASS=NODE and AUTHORITY=ACCESS or AUTHORITY=OWNER parameters on
the GRANT AUTHORITY command. You must have one of the following privileges to
grant or revoke client access or client owner authority:
v System privilege
v Policy privilege in the client's domain
v Client owner privilege over the node
v Client access privilege over the node
You can grant an administrator client access authority to individual clients or to all
clients in a specified policy domain. For example, you may want to grant client
access privileges to users that staff help desk environments.
Related tasks:
“Example: setting up help desk access to client computers in a specific policy
domain” on page 449

Granting client authority

You might have to grant client authority to a user at some time.

Procedure
1. Issue the following command to grant client access authority to administrator
FRED for the LABCLIENT node:
grant authority fred class=node node=labclient

The administrator FRED can now access the LABCLIENT client, and perform
backup and restore. The administrator can only restore data to the LABCLIENT
node.
2. Issue the following command to grant client owner authority to ADMIN1 for
the STUDENT1 node:
grant authority admin1 class=node authority=owner node=student1

The user ID ADMIN1 can now perform backup and restore operations for the
STUDENT1 client node. The user ID ADMIN1 can also restore files from the
STUDENT1 client node to a different client node.

Chapter 12. Managing client nodes 447

 Automatically creating an administrative user ID with client
owner authority
When you issue the REGISTER NODE command, by default, the server creates an
administrative user ID in addition to the client node. The administrative user ID
has client owner authority to the node when the node is defined to the server.

About this task

When the node is created, the authentication method and Secure Sockets Layer
(SSL) settings are inherited by the administrator.

Procedure

To register client node DESK2, issue the following example command:

register node desk2 pass2dsk

The following output is an example of this command:

ANR2060I Node DESK2 registered in policy domain STANDARD.

ANR2099I Administrative userid DESK2 defined for OWNER access to node DESK2.

The DESK2 client node is registered, in addition to an administrative user ID with

the same ID. The administrative user ID DESK2 has a password of pass2dsk with
client owner authority to the DESK2 node. When the PASSWORDACCESS=GENERATE
option is used by the client to change the password, the administrative DESK2 ID
can still access the client from a remote location.

Preventing automatic creation of an administrative user ID with

client owner authority
You can prevent automatic creation of an administrative user ID with client owner
authority by specifying USERID=NONE on the REGISTER NODE command.

Procedure

To register DESK2 without creating an administrative user ID with client owner

authority by default, issue the following example command:
register node desk2 pass2dsk userid=none

Registering a node and granting an existing administrative ID

client owner authority
You can grant client owner authority to an existing administrative user ID.

Procedure

To give client owner authority to the HELPADMIN user ID when registering the
NEWCLIENT node, issue the following command:
register node newclient pass2new userid=helpadmin

This command results in the NEWCLIENT node being registered with a password
of pass2new, and also grants HELPADMIN client owner authority. This command
would not create an administrator ID. The HELPADMIN client user ID is now able
to access the NEWCLIENT node from a remote location.

448 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Example: setting up help desk access to client computers in a
specific policy domain
The example is for setting up help desk access for user HELP1 to the client nodes
in the FINANCE domain.

About this task

You are also granting HELP1 client access authority to the FINANCE domain
without having to grant system or policy privileges.

The client nodes have been previously set up as follows:

v Installed and configured. The URL and port numbers were specified during the
REGISTER NODE process.
v Assigned to the FINANCE policy domain.
v Started the Client Acceptor service.
v Specified passwordaccess generate option in their client option files.

The help desk person, using HELP1 user ID, has a Web browser with Java
Runtime Environment (JRE) 1.3.1.

Procedure
1. Register an administrative user ID of HELP1.
register admin help1 05x23 contact="M. Smith, Help Desk x0001"
2. Grant the HELP1 administrative user ID client access authority to all clients in
the FINANCE domain. With client access authority, HELP1 can perform backup
and restore operations for clients in the FINANCE domain. Client nodes in the
FINANCE domain are Dave, Sara, and Joe.
grant authority help1 class=node authority=access domains=finance
The following output is generated by this command:

ANR2126I GRANT AUTHORITY: Administrator HELP1 was granted ACCESS authority for client
DAVE.
ANR2126I GRANT AUTHORITY: Administrator HELP1 was granted ACCESS authority for client
JOE.
ANR2126I GRANT AUTHORITY: Administrator HELP1 was granted ACCESS authority for client
SARA.

3. The help desk person, HELP1, opens the Web browser and specifies the URL
and port number for client computer Sara:
[Link]
A Java applet is started, and the client hub window is displayed in the main
window of the Web browser. When HELP1 accesses the backup function from
the client hub, the IBM Tivoli Storage Manager login screen is displayed in a
separate Java applet window. HELP1 authenticates with the administrative user
ID and password. HELP1 can perform a backup for Sara.

Results

For information about what functions are not supported on the Web
backup-archive client, refer to Backup-Archive Clients Installation and User's Guide.

Chapter 12. Managing client nodes 449

 Collecting processor value unit information in a VMware host
environment
If you are installing Tivoli Storage Manager on a system that uses VMware
virtualization software, ensure that PVU information is collected in the virtual
environment. To collect PVU information in a virtual environment, you must have
VMware ESX 4 software installed on your system, and you must set up and run
hardware scans.

Procedure

Complete the following steps:

1. Obtain the Tivoli Storage Manager backup-archive client installation package
for the Linux x86 operating system.
For information about obtaining installation packages, see the Passport
Advantage® website at [Link]
passportadvantage/[Link].
2. Install the Tivoli Storage Manager backup-archive client on a computer in your
system.
3. From the ../CIT/VMware/esx-4 directory, download the dispatcher,
[Link], [Link], and cpuid files.
4. Copy the downloaded files to the ESX 4 host operating system.

Tip: You can copy the files to any location on the host operating system, but
ensure that all files are copied to the same directory.
5. Ensure that guest virtual machines are running. This step is necessary to ensure
that the guest virtual machines are detected during the hardware scan.
6. To collect PVU information, issue the following command:
retrieve -v

What to do next

If you restart the host machine or change the configuration, run the retrieve
command again to ensure that current information is retrieved.

Tip: When the IBM Tivoli Storage Manager for Virtual Environments license file is
installed on a VMware vStorage backup server, the platform string that is stored
on the Tivoli Storage Manager server is set to TDP VMware for any node name
that is used on the server. The reason is that the server is licensed for Tivoli
Storage Manager for Virtual Environments. The TDP VMware platform string can
be used for PVU calculations. If a node is used to back up the server with standard
backup-archive client functions, such as file-level and image backup, interpret the
TDP VMware platform string as a backup-archive client for PVU calculations.

450 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Managing file spaces
A file space name identifies a group of files that are stored as a logical unit in
server storage. Administrators manage file spaces in which IBM Tivoli Storage
Manager stores each client node's data.

Administrators can perform the following activities when managing file spaces:

Task Required Privilege Class

Determine when existing file spaces are System, unrestricted policy privilege, or
renamed to allow for the creation of new restricted policy privilege for the policy
Unicode-enabled file spaces domain to which the client node is assigned.
Displaying information about file spaces Any administrator
Move selected file spaces for a single node, System, unrestricted storage, or restricted
as well as move a node's data located in a storage privilege for the source storage pool.
sequential access storage pool If your authorization is restricted storage
privilege and you intend to move data to
another storage pool, you must also have
the appropriate authority for the destination
storage pool.
Deleting file spaces System or unrestricted policy
Deleting file spaces assigned to specific System, unrestricted policy, or restricted
policy domains policy for those domains

Related reference:
“Defining client nodes and file spaces”

Defining client nodes and file spaces

Each client is given a node name when it is registered with the server. The server
views its registered nodes as clients that require services and resources from the
server.

Typically, a node is equivalent to a computer as in the case of a backup-archive

client installed on a user's computer for file system backups. However, multiple
nodes can exist on a single computer as in the case of a SQL server containing both
an application client for SQL database and transaction log backups, and a
backup-archive client for file system backups.

Typically, each client file system is represented on the server as a unique file space
that belongs to each client node. Therefore, the number of file spaces a node has
depends on the number of file systems on the client computer. For example, a
Windows desktop system may have multiple drives (file systems), such as C: and
D:. In this case, the client's node has two file spaces on the server; one for the C:
drive and a second for the D: drive. The file spaces can grow as a client stores
more data on the server. The file spaces decrease as backup and archive file
versions expire and the server reclaims the space.

IBM Tivoli Storage Manager does not allow an administrator to delete a node
unless the node's file spaces have been deleted.

Chapter 12. Managing client nodes 451

 Mapping file spaces for clients
For client nodes running on Windows, file spaces map to logical partitions and
shares. Each file space is named with the UNC name of the respective client
partition or share.

For client nodes running on NetWare, file spaces map to NetWare volumes. Each
file space is named with the corresponding NetWare volume name.

For clients running on Macintosh, file spaces map to Macintosh volumes. Each file
space is named with the corresponding Macintosh volume name.

For clients running on UNIX or Linux, a file space name maps to a file space in
storage that has the same name as the file system or virtual mount point from
which the files originated. The VIRTUALMOINTPOINT option allows users to define a
virtual mount point for a file system to back up or archive files beginning with a
specific directory or subdirectory. For information on the VIRTUALMOUNTPOINT
option, refer to the appropriate Backup-Archive Clients Installation and User's Guide.

Defining object names for clients

A fully qualified name for a client object can refer to a full file name and path that
is combined, or to the full directory structure for the object.

About this task

For client nodes that are running on Windows, it is possible to create objects with
long fully qualified names. The IBM Tivoli Storage Manager clients for Windows
are able to support fully qualified names of up to 8704 bytes in length for backup
and restore functions. These long names are often generated with an automatic
naming function or are assigned by an application.

Important: The Unicode representation of a character can occupy several bytes, so

the maximum number of characters that a fully qualified name might contain can
vary. See Setting options in the client options file in the Backup-Archive Clients
Installation and User's Guide for Windows for a full explanation of file path names
and limits.

Long object names can be difficult to display and use through normal operating
system facilities, such as a command prompt window or Windows Explorer. To
manage them, Tivoli Storage Manager assigns an identifying token to the name
and abbreviates the length. The token ID is then used to display the full object
name. For example, an error message might display as follows, where
[TSMOBJ:9.1.2084] is the assigned token ID:
ANR9999D file.c(1999) Error handling file [TSMOBJ:9.1.2084] because of
lack of server resources.

The token ID can then be used to display the fully qualified object name by
specifying it in the DISPLAY OBJNAME command.

Procedure

Issue the DISPLAY OBJNAME command with the token ID [TSMOBJ:9.1.2084]:

display objname 9.1.2084
ANR0160I Displaying segment 1 of 2 for client object 9.1.2084 with text:
\L08000-100\SUBDIRECTORY000000000000000000000000000000000000000000000000000000000
000000000000000000000000000002\SUBDIRECTORY00000000000000000000000000000000000000
0000000000000000000000000000000000000000000000003\SUBDIRECTORY0000000000000000000
00000000000000000000000000000000000000000000000000000000000000000004\SUBDIRECTORY

452 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 000000000000000000000000000000000000000000000000000000000000000000000000000000000
000005\SUBDIRECTORY00000000000000000000000000000000000000000000000000000000000000
0000000000000000000000006\SUBDIRECTORY0000000000000000000000000000000000000000000
00000000000000000000000000000000000000000007\SUBDIRECTORY000000000000000000000000
000000000000000000000000000000000000000000000000000000000000008\SUBDIRECTORY00000
000000000000000000000000000000000000000000000000000000000000000000000000000000000
9\SUBDIRECTORY0000000000000000000000000000000000000000000000000000000000000000000
00000000000000000010\SUBDIRECTORY000000000000000000000000000000000000000000000000
000000000000000000000000000000000000011\SUB.
ANR0160I Displaying segment 2 of 2 for client object 9.1.2084 with text:
DIRECTORY000000000000000000000000000000000000000000000000000000000000000000000000
000000000000012\SUBDIRECTORY00000000000000000000000000000000000000000000000000000
0000000000000000000000000000000013\SUBDIRECTORY0000000000000000000000000000000000
00000000000000000000000000000000000000000000000000014\SUBDIRECTORY000000000000000
000000000000000000000000000000000000000000000000000000000000000000000015\SUBDIREC
TORY00000000000000000000000000000000000000000000000000000000000000000000000000000
0000000016\SUBDIRECTORY0000000000000000000000000000000000000000000000000000000000
00000000000000000000000000017\SUBDIRECTORY000000000000000000000000000000000000000
000000000000000000000000000000000000000000000018\SUBDIRECTORY00000000000000000000
0000000000000000000000000000000000000000000000000000000000000000019\FILE001.

The fully qualified object name is displayed. If you are displaying long object
names that are included in backup sets, a token ID might not be included if the
entries for the path no longer exist in the Tivoli Storage Manager server database.
You cannot obtain a token ID by issuing QUERY BACKUPSETCONTENTS. To view the
fully qualified name, you can load the backup set table of contents from the client.

For more information about fully qualified object names and issuing the DISPLAY
OBJNAME command, see the Administrator's Reference.

Supporting Unicode-enabled clients

Unicode is a universal character encoding standard that supports the interchange,
processing, and display of text that is written in any of the languages of the
modern world.

For Windows, Macintosh OS X, and NetWare systems with the Unicode-enabled

client, the server supports storing file spaces with Unicode file space names,
directory names, and file names in server storage. The file spaces in server storage
that have Unicode names are called Unicode-enabled file spaces. Support for
Unicode names enables a client to successfully process an IBM Tivoli Storage
Manager operation, even when the file spaces contain directory names or files in
multiple languages, or when the client uses a different code page than the server.

New clients storing data on the server for the first time require no special setup. If
the client has the latest IBM Tivoli Storage Manager client software installed, the
server automatically stores Unicode-enabled file spaces for that client.

However, if you have clients that already have data stored on the server and the
clients install the Unicode-enabled IBM Tivoli Storage Manager client software, you
need to plan for the migration to Unicode-enabled file spaces. To allow clients with
existing data to begin to store data in Unicode-enabled file spaces, IBM Tivoli
Storage Manager provides a function for automatic renaming of existing file
spaces. The file data itself is not affected; only the file space name is changed.
After the existing file space is renamed, the operation creates a new file space that
is Unicode-enabled. The creation of the new Unicode-enabled file space for clients
can greatly increase the amount of space required for storage pools and the
amount of space required for the server database. It can also increase the amount

Chapter 12. Managing client nodes 453

 of time required for a client to run a full incremental backup, because the first
incremental backup after the creation of the Unicode-enabled file space is a full
backup.

When clients with existing file spaces migrate to Unicode-enabled file spaces, you
need to ensure that sufficient storage space for the server database and storage
pools is available. You also need to allow for potentially longer backup windows
for the complete backups.

Attention: After the server is at the latest level of software that includes support
for Unicode-enabled file spaces, you can only go back to a previous level of the
server by restoring an earlier version of IBM Tivoli Storage Manager and the
database.

A Unicode-enabled IBM Tivoli Storage Manager client is currently available for

Windows, Macintosh OS X, and NetWare operating systems. Data in a Unicode
code page from any other source, including down-level clients and API clients, will
not be identified or treated as Unicode-enabled.

It is strongly recommended that users of Windows, Macintosh, and NetWare

operating systems migrate their non-Unicode file spaces to Unicode-enabled file
spaces. For more information see Backup-Archive Clients Installation and User's Guide.
Related concepts:
“Unicode-enabled clients and existing backup sets” on page 462
Related tasks:
“Migrating clients to Unicode-enabled file spaces” on page 455
Related reference:
“Reasons for migrating clients to Unicode-enabled file spaces”
“Querying Unicode-enabled file spaces” on page 462

Reasons for migrating clients to Unicode-enabled file spaces

Without IBM Tivoli Storage Manager support for storing Unicode-enabled file
spaces, some clients experience backup failures when file spaces contain names of
directories or files in multiple languages, or have names that cannot be converted
to the server's code page.

When IBM Tivoli Storage Manager cannot convert the code page, the client may
receive one or all of the following messages if they were using the command line:
ANS1228E, ANS4042E, and ANS1803E. Clients that are using the GUI may see a
“Path not found” message. If you have clients that are experiencing such backup
failures, then you need to migrate the file spaces for these clients to ensure that
these systems are completely protected with backups. If you have a large number
of clients, set the priority for migrating the clients based on how critical each
client's data is to your business.

Any new file spaces that are backed up from client systems with the
Unicode-enabled IBM Tivoli Storage Manager client are automatically stored as
Unicode-enabled file spaces in server storage.

Objects backed up or archived with a Unicode-enabled IBM Tivoli Storage

Manager client in any supported language environment can be restored or
retrieved with a Unicode-enabled client in the same or any other supported
language environment. This means, for example, that files backed up by a Japanese
Unicode-enabled client can be restored by a German Unicode-enabled client.

454 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Important: Objects backed up or archived by a Unicode-enabled IBM Tivoli
Storage Manager client cannot be restored or retrieved by a client that is not
Unicode-enabled.
Related tasks:
“Migrating clients to Unicode-enabled file spaces”

Migrating clients to Unicode-enabled file spaces

To allow clients with existing data to migrate to Unicode-enabled file spaces, IBM
Tivoli Storage Manager provides an automatic rename function for file spaces.

About this task

When enabled, IBM Tivoli Storage Manager uses the rename function when it
recognizes that a file space that is not Unicode-enabled in server storage matches
the name of a file space on a client. The existing file space in server storage is
renamed, so that the file space in the current operation is then treated as a new,
Unicode-enabled file space. For example, if the operation is an incremental backup
at the file space level, the entire file space is then backed up to the server as a
Unicode-enabled file space.

The following example shows how this process works when automatic renaming is
enabled from the server, for an existing client node that has file spaces in server
storage.

Procedure
1. The administrator updates a client node definition by issuing an UPDATE NODE
command with the parameter, AUTOFSRENAME YES.
2. The client processes an incremental back up.
3. The IBM Tivoli Storage Manager processes the backup as follows:
a. Renames the existing file space (_OLD)
b. Creates a new Unicode-enabled file space
c. Processes the backup in the current operation to the new Unicode-enabled
file space

Results

If you force the file space renaming for all clients at the same time, backups can
contend for network and storage resources, and storage pools can run out of
storage space.
Related tasks:
“Planning for Unicode versions of existing client file spaces” on page 458
“Examining issues when migrating to Unicode” on page 459
“Example of a migration process” on page 460
Related reference:
“Defining options for automatically renaming file spaces” on page 456
“Defining the rules for automatically renaming file spaces” on page 457

Chapter 12. Managing client nodes 455

 Defining options for automatically renaming file spaces:

As an administrator, you can control whether the file spaces of any existing clients
are renamed to force the creation of new Unicode-enabled file spaces. By default,
no automatic renaming occurs.

To control the automatic renaming, use the parameter AUTOFSRENAME when you
register or update a node. You can also allow clients to make the choice. Clients
can use the client option AUTOFSRENAME.

Restriction: The setting for AUTOFSRENAME affects only clients that are
Unicode-enabled.

You have the following options:

v Do not allow existing file spaces to be renamed, so that Unicode-enabled file
spaces are not created (AUTOFSRENAME=NO, the default).
IBM Tivoli Storage Manager does not automatically rename client file spaces
when the client system upgrades to the Unicode-enabled Tivoli Storage Manager
client. This setting can help an administrator control how many clients' file
spaces can be renamed at one time. The administrator can determine how many
Unicode-enabled clients exist by issuing the QUERY NODE FORMAT=DETAILED
command. The output displays the client level.
v Automatically rename existing file spaces, forcing the creation of
Unicode-enabled file spaces in place of the renamed file spaces
(AUTOFSRENAME=YES).
Tivoli Storage Manager automatically renames client file spaces in server storage
when the client upgrades to the Unicode-enabled client and runs one of the
following operations: archive, selective backup, full incremental backup, or
partial incremental backup. Tivoli Storage Manager automatically renames the
file spaces that are specified in the current operation and creates new,
Unicode-enabled file spaces where files and directories are stored to complete
the operation. Other file spaces that are not specified in the current operation are
not affected by the rename. Therefore, a client can have mixed file spaces.
Attention: If you force the renaming of the file space for all clients at the same
time, client operations can contend for network and storage resources. Also,
storage pools can run out of storage space.
v Allow clients to choose whether to rename files spaces, in effect, choosing
whether new Unicode-enabled file spaces are created
(AUTOFSRENAME=CLIENT).
If you use this value for a client node, the client can set its AUTOFSRENAME option
in its options file. The client option determines whether file spaces are renamed
or whether the user is prompted for renaming during a Tivoli Storage Manager
operation (PROMPT).
The default value for the client option is PROMPT. When the option is set for
prompting, the client is presented with a choice about renaming file spaces.
When a client that has existing file spaces on server storage upgrades to the
Unicode-enabled client, the file spaces might be renamed. If the client runs a
Tivoli Storage Manager operation with the server, the user is asked to choose
whether to rename the file spaces that are involved in the current operation.
The client is prompted only one time about renaming a particular file space.
If the client does not choose to rename the file space, the administrator can later
rename the file space to create a new Unicode-enabled file space. The file space

456 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 is created the next time that the client processes an archive, selective backup, full
incremental backup, or partial incremental backup.
Attention: There is no prompt for operations that run with the client scheduler.
If the client is running the scheduler and the client AUTOFSRENAME option is set to
PROMPT, there is no prompt and the file space is not renamed. A client session
can then run unattended. The prompt is displayed during the next interactive
session on the client.

The following table summarizes what occurs with different parameter and option
settings.
Table 40. The effects of the AUTOFSRENAME option settings
Parameter on the Option on the client Result for file spaces Is the file space
server (for each renamed?
client)
Yes Yes, No, Prompt Renamed Yes
No Yes, No, Prompt Not renamed No
Client Yes Renamed Yes
Client No Not renamed Yes
Client Prompt Command-line or GUI: The user receives a Depends on the
one-time-only prompt about renaming response from the user
(yes or no)
Client Prompt Client Scheduler: Not renamed (prompt is No
displayed during the next command-line
or GUI session)

Related reference:
“Defining the rules for automatically renaming file spaces”

Defining the rules for automatically renaming file spaces:

With its automatic renaming function, IBM Tivoli Storage Manager renames a file
space by adding the suffix _OLD.

For example:

Original file space name New file space name

\\maria\c$ \\maria\c$_OLD

If the new name would conflict with the name of another file space, a number is
added to the suffix. For example:

Original file space name New file space name Other existing file spaces
\\maria\c$ \\maria\c$_OLD \\maria\c$_OLD1
\\maria\c$_OLD2

If the new name for the file space exceeds the limit of 64 characters, the file space
name is truncated before the suffix _OLD is added.

Chapter 12. Managing client nodes 457

 Planning for Unicode versions of existing client file spaces:

Several factors must be considered before you plan for Unicode versions of
existing client file spaces.

About this task

Consider the following items when planning:

v After clients with existing file spaces start to create Unicode-enabled file spaces,
they will still need to have access to the renamed file spaces that are not
Unicode-enabled for some period of time.
v Your storage pool and database space requirements can double if you allow all
clients to create Unicode-enabled file spaces in addition to their existing file
spaces that are not Unicode-enabled.
v Because the initial backups after migration are complete backups, it can also
greatly increase the time required to finish backup operations.

To minimize problems, you need to plan the storage of Unicode-enabled file spaces
for clients that already have existing file spaces in server storage.

Procedure
1. Determine which clients need to migrate.
Clients that have had problems with backing up files because their file spaces
contain names of directories or files that cannot be converted to the server's
code page should have the highest priority. Balance that with clients that are
most critical to your operations. If you have a large number of clients that need
to become Unicode-enabled, you can control the migration of the clients.
Change the rename option for a few clients at a time to keep control of storage
space usage and processing time. Also consider staging migration for clients
that have a large amount of data backed up.
2. Allow for increased backup time and network resource usage when the
Unicode-enabled file spaces are first created in server storage.
Based on the number of clients and the amount of data those clients have,
consider whether you need to stage the migration. Staging the migration means
setting the AUTOFSRENAME parameter to YES or CLIENT for only a small number
of clients every day.

Note: If you set the AUTOFSRENAME parameter to CLIENT, be sure to have the
clients (that run the client scheduler) set their option to AUTOFSRENAME YES. This
ensures the file spaces are renamed.
3. Check the current storage usage for the clients that need to become
Unicode-enabled.
You can use the QUERY OCCUPANCY command to display information on how
much space each client is currently using. Initially, clients will need only the
amount of space used by active files. Therefore, you need to estimate how
much of the current space is used by copies (different versions of the same file).
Migration will result in a complete backup at the next incremental backup, so
clients will need space for that backup, plus for any other extra versions that
they will keep. Therefore, the amount of storage required also depends on
policy (see the next step). Your IBM Tivoli Storage Manager policy specifies
how files are backed up, archived, migrated from client node storage, and
managed in server storage.

458 IBM Tivoli Storage Manager for Windows: Administrator's Guide

4. Understand how your IBM Tivoli Storage Manager policies affect the storage
that will be needed.
If your policies expire files based only on the number of versions (Versions
Data Exists), storage space required for each client will eventually double, until
you delete the old file spaces.
If your policies expire files based only on age (Retain Extra Versions), storage
space required for each client will increase initially, but will not double.
If your policies use both the number of versions and their age, each client will
need less than double their current usage.
5. Estimate the effect on the database size.
The database size depends on the number of files in server storage, as well as
the number of versions of those files. As Unicode-enabled file spaces are
backed up, the original file spaces that were renamed remain. Therefore, the
server requires additional space in the database to store information about the
increased number of file spaces and files.
6. Arrange for the additional storage pool space, including space in copy storage
pools and active-data pools, based on your estimate from step 3 on page 458
and 4.
7. Check the server database space that is available and compare with your
estimate from step 5.
8. Ensure that you have a full database backup before you proceed with migration
of Unicode-enabled file spaces.
9. Consider how you will manage the renamed file spaces as they age. The
administrator can delete them, or the clients can be allowed to delete their own
file spaces.
Related tasks:
“Estimating database space requirements” on page 652

Examining issues when migrating to Unicode:

When you migrate to Unicode, there are several issues that you must consider.

About this task

The server manages a Unicode-enabled client and its file spaces as follows:
v When a client upgrades to a Unicode-enabled client and logs in to the server, the
server identifies the client as Unicode-enabled.

Remember: That same client (same node name) cannot log in to the server with
a previous version of IBM Tivoli Storage Manager or a client that is not
Unicode-enabled.
v The original file space that was renamed (_OLD) remains with both its active
and inactive file versions that the client can restore if needed. The original file
space will no longer be updated. The server will not mark existing active files
inactive when the same files are backed up in the corresponding
Unicode-enabled file space.

Important: Before the Unicode-enabled client is installed, the client can back up
files in a code page other than the current locale, but cannot restore those files.
After the Unicode-enabled client is installed, if the same client continues to use
file spaces that are not Unicode-enabled, the client skips files that are not in the
same code page as the current locale during a backup. Because the files are
skipped, they appear to have been deleted from the client. Active versions of the

Chapter 12. Managing client nodes 459

 files in server storage are made inactive on the server. When a client in this
situation is updated to a Unicode-enabled client, you should migrate the file
spaces for that client to Unicode-enabled file spaces.
v The server does not allow a Unicode-enabled file space to be sent to a client that
is not Unicode-enabled during a restore or retrieve process.
v Clients should be aware that they will not see all their data on the
Unicode-enabled file space until a full incremental backup has been processed.
When a client performs a selective backup of a file or directory and the original
file space is renamed, the new Unicode-enabled file space will contain only the
file or directory specified for that backup operation. All other directories and
files are backed up on the next full incremental backup.
If a client needs to restore a file before the next full incremental backup, the
client can perform a restore from the renamed file space instead of the new
Unicode-enabled file space. For example:
– Sue had been backing up her file space, \\sue-node\d$.
– Sue upgrades the IBM Tivoli Storage Manager client on her system to the
Unicode-enabled IBM Tivoli Storage Manager client.
– Sue performs a selective backup of the [Link] file.
– The automatic file space renaming function is in effect and IBM Tivoli Storage
Manager renames\\sue-node\d$ to \\sue-node\d$_OLD. IBM Tivoli Storage
Manager then creates a new Unicode-enabled file space on the server with the
name \\sue-node\d$. This new Unicode-enabled file space contains only the
[Link] file.
– All other directories and files in Sue's file system will be backed up on the
next full incremental backup. If Sue needs to restore a file before the next full
incremental backup, she can restore the file from the \\sue-node\d$_OLD file
space.
Refer to the Backup-Archive Clients Installation and User's Guide for more
information.

Example of a migration process:

The example of a migration process includes one possible sequence for migrating
clients.

About this task

Assumptions for this scenario are:

v The IBM Tivoli Storage Manager server database has been backed up.
v The latest server software has been installed. This installation has also
performed an upgrade to the server database.
v Clients have installed the latest software.
v A few clients are file servers. Most clients are workstations used by individuals.
v Clients generally run scheduled incremental backups every night.

The following migration process is possible to perform:

Procedure
1. Have all clients install the Unicode-enabled IBM Tivoli Storage Manager client
software.

460 IBM Tivoli Storage Manager for Windows: Administrator's Guide

2. Migrate the file servers first. For clients that are file servers, update the
AUTOFSRENAME parameter to enable automatic renaming for the file spaces. For
example, if the client node names for all file servers begin with FILE, issue the
following command:
update node file* autofsrename=yes

This forces the file spaces to be renamed at the time of the next backup or
archive operation on the file servers. If the file servers are large, consider
changing the renaming parameter for one file server each day.
3. Allow backup and archive schedules to run as usual. Monitor the results.
a. Check for the renamed file spaces for the file server clients. Renamed file
spaces have the suffix _OLD or _OLDn, where n is a number.
b. Check the capacity of the storage pools. Add tape or disk volumes to
storage pools as needed.
c. Check database usage statistics to ensure you have enough space.

Note: If you are using the client acceptor to start the scheduler, you must first
modify the default scheduling mode.
4. Migrate the workstation clients. For example, migrate all clients with names
that start with the letter a.
update node a* autofsrename=yes
5. Allow backup and archive schedules to run as usual that night. Monitor the
results.
6. After sufficient time passes, consider deleting the old, renamed file spaces.
Related tasks:
“Modifying the default scheduling mode” on page 576
Related reference:
“Managing the renamed file spaces”
“Defining the rules for automatically renaming file spaces” on page 457

Managing the renamed file spaces:

The file spaces that were automatically renamed (_OLD) to allow the creation of
Unicode-enabled file spaces continue to exist on the server. Users can still access
the file versions in these file spaces.

Because a renamed file space is not backed up again with its new name, the files
that are active (the most recent backup version) in the renamed file space remain
active and never expire. The inactive files in the file space expire according to the
policy settings for how long versions are retained. To determine how long the files
are retained, check the values for the parameters, Retain Extra Versions and
Retain Only Versions, in the backup copy group of the management class to
which the files are bound.

When users no longer have a need for their old, renamed file spaces, you can
delete them. If possible, wait for the longest retention time for the only version
(Retain Only Version) that any management class allows. If your system has
storage constraints, you may need to delete these file spaces before that.

Chapter 12. Managing client nodes 461

 Querying Unicode-enabled file spaces
You can determine which file spaces are Unicode-enabled by querying all of the
file spaces.

Issue the following command:

query filespace

The result is similar to the following output:

Node Name Filespace FSID Platform Filespace Is Capacity Pct
Name Type Filespace (MB) Util
Unicode?
---------- ----------- ---- ------- --------- --------- -------- -----
SUE \\sue\c$ 1 WinNT NTFS Yes 2,502.3 75.2
SUE \\sue\d$ 2 WinNT NTFS Yes 6,173.4 59.6
JOE \\joe\c$ 1 WinNT NTFS No 12,299.7 31.7

To query a specific Unicode-enabled file space, it may be more convenient to use

the file space identifier (FSID) than the file space name. File space names for
Unicode-enabled file spaces may not be readable when displayed in the server's
code page. Attempting to enter the name of a Unicode-enabled file space may not
work because it depends on the server's code page and conversion routines that
attempt to convert from the server's code page to Unicode.
Related tasks:
“Displaying information about file spaces”

Unicode-enabled clients and existing backup sets

A client can have a backup set that contains both file spaces that are
Unicode-enabled and file spaces that are not Unicode-enabled. The client must
have the same level of IBM Tivoli Storage Manager or higher to restore the data in
the backup set.

For example, a Version 5.1.0 client backs up file spaces, and then upgrades to
Version 5.2.0 with support for Unicode-enabled file spaces. That same client can
still restore the non-Unicode file spaces from the backup set.

Unicode-enabled file spaces in a backup set can only be accessed by a

Unicode-enabled client, and not by an earlier version of the client. The server
allows only Unicode-enabled clients to restore data from Unicode-enabled file
spaces.
Related reference:
“Restoring backup sets from a backup-archive client” on page 548

Displaying information about file spaces

You can display file space information by identifying the client node name and file
space name.

About this task

You can display file space information for the following reasons:
v To identify file spaces that are defined to each client node, so that you can delete
each file space from the server before removing the client node from the server
v To identify file spaces that are Unicode-enabled and identify their file space ID
(FSID)
v To monitor the space that is used on workstation's disks

462 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 v To monitor whether backups are completing successfully for the file space
v To determine the date and time of the last backup

Note: File space names are case-sensitive and must be entered exactly as known to
the server.

Procedure

To view information about file spaces that are defined for client node JOE, issue
the following command:
query filespace joe *

The following figure shows the output from this command:

When you display file space information in detailed format, the Filespace Name
Node Name Filespace FSID Platform Filespace Is Capacity Pct
Name Type Filespace (MB) Util
Unicode?
---------- ----------- ---- ------- --------- --------- -------- -----
JOE \\joe\c$ 1 WinNT NTFS Yes 2,502.3 75.2
JOE \\joe\d$ 2 WinNT NTFS Yes 6,173.4 59.6

field might display file space names as “...”. This indicates to the administrator that
a file space does exist but could not be converted to the server's code page.
Conversion can fail if the string includes characters that are not available in the
server code page, or if the server has a problem accessing system conversion
routines.
File space names and file names that can be in a different code page or locale than
the server do not display correctly in the Operations Center or the administrative
command-line interface. The data itself is backed up and can be restored properly,
but the file space name or file name may display with a combination of invalid
characters or blank spaces.
Refer to the Administrator's Reference for details.

Moving data for a client node

You can move a client node's data in a sequential-access storage pool or move
selected file spaces for a single node.
Related tasks:
“Moving data belonging to a client node” on page 406

Deleting file spaces

You can delete a client node from a server, but first you must delete all of that
client's data from server storage by deleting any file spaces that belong to the
node.

About this task

Administrators might want to delete a file space in the following cases:

v Users are not authorized to delete backed-up or archived files in storage pools.
v You want to remove a client node from the server.
v You want to delete a specific user's files.

Procedure
1. For users who are not authorized to delete backed-up or archived files in
storage pools, as an example, client node PEASE no longer needs archived files

Chapter 12. Managing client nodes 463

 in file space /home/pease/dir2. However, the user does not have the authority
to delete those files. To delete the files in /home/pease/dir2, issue the following
command:
delete filespace pease /home/pease/dir2 type=archive
The authority to delete backed-up or archived files from server storage is set
when a client node is registered.
2. You must delete a user's files from storage pools before you can remove a client
node. For example, to delete all file spaces that belong to client node DEBBYG,
issue the following command:
delete filespace debbyg * type=any

After you delete all of a client node's file spaces, you can delete the node with
the REMOVE NODE command.

Results

For client nodes that support multiple users, such as UNIX or Linux, a file owner
name is associated with each file on the server. The owner name is the user ID of
the operating system, such as the UNIX Linux user ID. When you delete a file
space that belongs to a specific owner, only files that have the specified owner
name in the file space are deleted.

When a node has more than one file space and you issue a DELETE FILESPACE
command for only one file space, a QUERY FILESPACE command for the node during
the delete process shows no file spaces. When the delete process ends, you can
view the remaining file spaces with the QUERY FILESPACE command. If data
retention protection is enabled, the only files that are deleted from the file space
are those that meet the retention criterion. The file space is not deleted if one or
more files within the file space cannot be deleted.

Note: Data stored using the System Storage Archive Manager product cannot be
deleted using the DELETE FILESPACE command if the retention period for the data
has not expired. If this data is stored in a Centera storage pool, then it is
additionally protected from deletion by the retention protection feature of the
Centera storage device.
Related concepts:
“Accepting default closed registration or enabling open registration” on page 421
Related tasks:
“Deleting client nodes” on page 440

Managing client option files

A client node connects with the server by using the information in a client options
file ([Link]). This file, located in the client directory, contains client options that
control processing and connections with the server.

The most important option is the network address of the server, but you can add
many other client options at any time. Administrators can also control client
options by creating client option sets on the server that are used in conjunction
with client option files on client nodes.
Related tasks:
“Creating client option sets on the server” on page 465
“Managing client option sets” on page 467

464 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 Related reference:
“Connecting nodes with the server” on page 425

Creating client option sets on the server

An administrator can create a set of client options to be used by a client node that
is at IBM Tivoli Storage Manager Version 3 or later. The client options specified in
the set are used in conjunction with the client options [Link] file.

About this task

Client option sets allow the administrator to specify additional options that may
not be included in the client's option file ([Link]). You can specify which clients
use the option set with the REGISTER NODE or UPDATE NODE commands. The client
can use these defined options during a backup, archive, restore, or retrieve process.
See the Backup-Archive Clients Installation and User's Guide for detailed information
about individual client options.

Procedure

To create a client option set and have the clients use the option set, perform the
following steps:
1. Create the client option set with the DEFINE CLOPTSET command.
2. Add client options to the option set with the DEFINE CLIENTOPT command.
3. Specify which clients should use the option set with the REGISTER NODE or
UPDATE NODE command.
Related reference:
“Connecting nodes with the server” on page 425

Creating a client option set

When you create a client option set, you define a name for the option set and can,
optionally, provide a description of the option set.

Procedure

To provide a description of the option set, issue the following example command:
define cloptset engbackup description=’Backup options for eng. dept.’

Tip: The option set is empty when it is first defined.

Adding client options in an option set

You can add client options in a defined client option set.

Procedure

Issue the following example command to add a client option (MAXCMDRETRIES) in

the ENGBACKUP option set:
define clientopt engbackup maxcmdretries 5

For a list of client options that you can specify, refer to Administrative client options
in the Administrator's Reference.
The server automatically assigns sequence numbers to the specified options, or you
can choose to specify the sequence number for order of processing. This is helpful
if you have defined more than one of the same option as in the following example:

Chapter 12. Managing client nodes 465

 define clientopt engbackup inclexcl "include d:\admin"
define clientopt engbackup inclexcl "include d:\payroll"

A sequence number of 0 is assigned to the option include d:\admin. A sequence

number of 1 is assigned to the option include d:\payroll. If you want to
specifically process one option before another, include the sequence parameter as
follows:
define clientopt engbackup inclexcl "include d:\admin" seqnumber=2"
define clientopt engbackup inclexcl "include d:\payroll" seqnumber=1"

The options are processed starting with the highest sequence number.

Any include-exclude statements in the server client option set have priority over
the include-exclude statements in the local client options file. The server
include-exclude statements are always enforced and placed last in the
include-exclude list and evaluated before the client include-exclude statements. If
the server option set has several include-exclude statements, the statements are
processed starting with the first sequence number. The client can issue the QUERY
INCLEXCL command to show the include-exclude statements in the order that they
are processed. QUERY INCLEXCL also displays the source of each include-exclude
statement. For more information on the processing of the include-exclude
statements see the Backup-Archive Clients Installation and User's Guide.

The FORCE parameter allows an administrator to specify whether the server forces
the client to use an option value. This parameter has no affect on additive options
such as INCLEXCL and DOMAIN. The default value is NO. If FORCE=YES, the server
forces the client to use the value, and the client cannot override the value. The
following example shows how you can prevent a client from using subfile backup:
define clientopt engbackup subfilebackup no force=yes
Related reference:
“The include-exclude list” on page 488

Registering client nodes and assigning them to an option set

You can register or update a client node and specify an option set for the client to
use.

Procedure

Issue the following command to register or update a client node:

register node mike pass2eng cloptset=engbackup

The client node MIKE is registered with the password pass2eng. When the client
node MIKE performs a scheduling operation, his schedule log entries are kept for
five days.

466 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 Managing client option sets
Administrators can perform several activities to manage client option sets.

Procedure

Perform the following steps:

1. Update the sequence number for a client option.
You can update the sequence number for a client option to change its
processing order. This is helpful if you have more than one of the same option,
for example several INCLUDE options.
The following command shows you how to change the sequence number for
the DATEFORMAT option from 0 to 9:
update clientopt engbackup dateformat 0 9
2. Delete an option from a client option set.
You can remove an option that is defined in a client option set.
The following example shows you how to remove the SCHEDMODE polling option
from the financeschd option set:
delete clientopt financeschd schedmode
3. Copy a client option set. You can copy an existing client option to another
option set.
The following example shows you how to copy the engbackup option set to
financeschd option set:
copy cloptset engbackup financeschd
4. Request information about a client option set. To display information about the
contents of a client option set, issue the following command:
query cloptset financeschd
5. Update the description for a client option set. You can update the description
for a client option set. The following example shows you how to update the
description for the engbackup option set:
update clopset engbackup description=’Scheduling information’
6. Delete a client option set. When you delete a client option set, client node
references to the option set are null. The clients continue to use their existing
client options file. The following example shows you how to delete the
engbackup client option set:
delete cloptset engbackup

Managing IBM Tivoli Storage Manager sessions

Each time an administrator or client node connects with the server, an
administrative or client session is established. IBM Tivoli Storage Manager tracks
its sessions in the server database.

Backup-archive clients are eligible for client restartable restore sessions; however,
application clients are not.

Tivoli Storage Manager can hold a client restore session in DSMC loop mode until
one of these conditions is met:
v The device class MOUNTRETENTION limit is satisfied.
v The client IDLETIMEOUT period is satisfied.
v The loop session ends.

Chapter 12. Managing client nodes 467

 Administrators can perform the following activities when managing IBM Tivoli
Storage Manager sessions:

Task Required Privilege Class

Displaying information about client sessions Any administrator
Canceling a client session System or operator
Disabling or enabling a client session System or operator

Related concepts:
“Managing client restartable restore sessions” on page 471

Displaying information about IBM Tivoli Storage Manager

sessions
Each client session is assigned a unique session number.

Procedure

To display information about client sessions, issue the following command:

query session

Figure 66 shows a sample client session report.

You can determine the state of the server by examining the session state and wait
Sess Comm. Sess Wait Bytes Bytes Sess Platform Client Name
Number Method State Time Sent Recvd Type
------ ------ ------ ------ ------- ------- ----- -------- -----------
471 Tcp/Ip IdleW 36 S 592 186 Node WinNT JOEUSER
472 Tcp/Ip RecvW 0 S 730 638 Node WinNT STATION1

Figure 66. Information about client sessions

time to determine how long (in seconds, minutes, or hours) the session has been in
the current state.

Server session states

The server session state can be Start, Run, End, RecvW, SendW, MediaW, or IdleW.

See the following definitions for the server session states:

Start Connecting with a client session.
Run Executing a client request.
End Ending a client session.
RecvW
Waiting to receive an expected message from the client while a database
transaction is in progress. A session in this state is subject to the
COMMTIMEOUT limit.
SendW
Waiting for acknowledgment that the client has received a message sent by
the server.
MediaW
Waiting for removable media to become available.
Aggregation can cause multiple media waits within a transaction and is
indicated by one client message.

468 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 Important: If QUERY SESSION FORMAT=DETAILED is specified, the Media
Access Status field displays the type of media wait state.
IdleW
Waiting for communication from the client, and a database transaction is
NOT in progress. A session in this state is subject to the limit as specified
in the server options file.
If a client does not initiate communication within the specified time limit
set by the IDLETIMEOUT option in the server options file, then the server
cancels the client session.
For example, if the IDLETIMEOUT option is set to 30 minutes, and a user
does not initiate any operations within those 30 minutes, then the server
cancels the client session. The client session is automatically reconnected to
the server when it starts to send data again.
Related tasks:
“Reclaiming space in sequential-access storage pools” on page 367

Canceling an IBM Tivoli Storage Manager session

You can cancel a client session with the CANCEL SESSION command and the
associated session number. Canceling sessions may be necessary when a user's
computer is not responding or as a prerequisite to halting the server.

About this task

Administrators can display a session number with the QUERY SESSION command.

Users and administrators whose sessions have been canceled must reissue their last
command to access the server again.

If an operation, such as a backup or an archive process, is interrupted when you

cancel the session, the server rolls back the results of the current transaction. That
is, any changes made by the operation that are not yet committed to the database
are undone. If necessary, the cancellation process may be delayed.

If the session is in the Run state when it is canceled, the cancel process does not
take place until the session enters the SendW, RecvW, or IdleW state.

If the session you cancel is currently waiting for a media mount, the mount request
is automatically canceled. If a volume associated with the client session is currently
being mounted by an automated library, the cancel may not take effect until the
mount is complete.

Procedure

For example, to cancel a session for client MARIE:

1. Query client sessions to determine the session number. The example report
displays MARIE's session number 6.
2. Cancel node MARIE's session by entering:
cancel session 6

Chapter 12. Managing client nodes 469

 Results

If you want to cancel all backup and archive sessions, enter:

cancel session all
Related tasks:
“Displaying information about IBM Tivoli Storage Manager sessions” on page 468
Related reference:
“Server session states” on page 468

When a client session is automatically canceled

Client sessions can be automatically canceled.

The reasons are based on the settings of the following server options:
COMMTIMEOUT
Specifies how many seconds the server waits for an expected client
message during a transaction that causes a database update. If the length
of time exceeds this time-out, the server rolls back the transaction that was
in progress and ends the client session. The amount of time it takes for a
client to respond depends on the speed and processor load for the client
and the network load.
IDLETIMEOUT
Specifies how many minutes the server waits for a client to initiate
communication. If the client does not initiate communication with the
server within the time specified, the server ends the client session. For
example, the server prompts the client for a scheduled backup operation
but the client node is not started. Another example can be that the client
program is idle while waiting for the user to choose an action to perform
(for example, backup archive, restore, or retrieve files). If a user starts the
client session and does not choose an action to perform, the session will
time out. The client program automatically reconnects to the server when
the user chooses an action that requires server processing. A large number
of idle sessions can inadvertently prevent other users from connecting to
the server.
THROUGHPUTDATATHRESHOLD
Specifies a throughput threshold, in kilobytes per second, a client session
must achieve to prevent being cancelled after the time threshold is reached.
Throughput is computed by adding send and receive byte counts and
dividing by the length of the session. The length does not include time
spent waiting for media mounts and starts at the time a client sends data
to the server for storage. This option is used in conjunction with the
THROUGHPUTTIMETHRESHOLD server option.
THROUGHPUTTIMETHRESHOLD
Specifies the time threshold, in minutes, for a session after which it may be
canceled for low throughput. The server ends a client session when it has
been active for more minutes than specified and the data transfer rate is
less than the amount specified in the THROUGHPUTDATATHRESHOLD server
option.

Refer to the Administrator's Reference for more information.

470 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Disabling or enabling access to the server
You can prevent clients from establishing sessions with the server by issuing the
DISABLE SESSIONS command.

About this task

This command does not cancel sessions currently in progress or system processes
such as migration and reclamation.

Task Required Privilege Class

Disabling and enabling client node access to System or operator
the server
Displaying server status Any administrator

Procedure

To disable client node access to the server, issue the following example command:
disable sessions

You continue to access the server and current client activities complete unless a
user logs off or an administrator cancels a client session. After the client sessions
have been disabled, you can enable client sessions and resume normal operations
by issuing the following command:
enable sessions

You can issue the QUERY STATUS command to determine if the server is enabled or
disabled.
Related tasks:
“Locking and unlocking client nodes” on page 440

Managing client restartable restore sessions

Some large restore operations may invoke a special type of restore operation called
client restartable restore sessions. These special sessions allow users to restart the
restore session from where it stopped, if the session was interrupted.

IBM Tivoli Storage Manager identifies client restartable restore sessions by

displaying message ANS1247I on the client computer when the session starts.
These restore sessions can be restarted as long as the restore interval has not
expired.

After a restore operation that comes directly from tape, the Tivoli Storage Manager
server does not release the mount point to IDLE status from INUSE status. The
server does not close the volume to allow additional restore requests to be made to
that volume. However, if there is a request to perform a backup in the same
session, and that mount point is the only one available, then the backup operation
will stop and the server will issue message ANS1114I. You can avoid this by
closing the DSMC restore session after the restore operation completes. This
releases the mount point for subsequent sessions.

When a restartable restore session is saved in the server database the file space is
locked in server storage. The following rules are in effect during the file space lock:
v Files residing on sequential volumes associated with the file space cannot be
moved.

Chapter 12. Managing client nodes 471

 v Files associated with the restore cannot be backed up. However, files not
associated with the restartable restore session that are in the same file space are
eligible for backup. For example, if you are restoring all files in directory A, you
can still backup files in directory B from the same file space.

The RESTOREINTERVAL server option allows administrators to specify how long

client restartable restore sessions are saved in the server database. Consider
scheduled backup operations when setting this option. For more information, refer
to the RESTOREINTERVAL server option in the Administrator's Reference.

Administrators can perform the following activities when managing client

restartable restore sessions:

Task Required Privilege Class

Displaying information about client Any administrator
restartable restore sessions
Canceling client restartable restore sessions System or operator
Interrupting client restartable restore System or operator
sessions

Displaying information about a client restartable restore session

You can display information about client restartable restore sessions with the QUERY
RESTORE command.

Procedure

To determine which client nodes have eligible restartable restore sessions, issue the
following example command:
query restore

Restartable restore sessions have a negative session number.

Canceling a client restartable restore session

When a client restore session is in a restartable state, the file space is locked in
server storage and no files can be moved from sequential volumes. This prevents
the data from being migrated, moved, reclaimed, or backed up by another
operation.

About this task

These sessions will automatically expire when the specified restore interval has
passed.

An administrator can cancel a restartable restore session that is in an active or

restartable state. If the restore session is active, any outstanding mount requests
related to the active session are automatically canceled. When a restartable restore
session is canceled with the CANCEL RESTORE command, it cannot be restarted from
the point of interruption. A restartable restore session always has a negative
session number.

Procedure

To cancel a restartable restore session, you must specify the session number. For
example:

472 IBM Tivoli Storage Manager for Windows: Administrator's Guide

cancel restore -1

Interrupting an active client restartable restore session

An administrator can interrupt an active restartable restore session by canceling the
session, but the session can not then be restarted.

About this task

A session that ends prematurely through an error or ends by an administrator

using CTRL-C on the Tivoli Storage Manager client might be restartable. Issue the
QUERY RESTORE command to show the restartable restore sessions. A session with a
negative number can be restarted.

Procedure

Issue the following command to cancel a session:

cancel session -2

Session -2 cannot be restarted after you issue this command.

Chapter 12. Managing client nodes 473

474 IBM Tivoli Storage Manager for Windows: Administrator's Guide
Chapter 13. Implementing policies for client data
Policies are rules that you set at the IBM Tivoli Storage Manager server to help you
manage client data. Policies control how and when client data is stored.

For example:
v How and when files are backed up and archived to server storage
v How space-managed files are migrated to server storage
v The number of copies of a file and the length of time copies are kept in server
storage

IBM Tivoli Storage Manager provides a standard policy that sets rules to provide a
basic amount of protection for data on workstations. If this standard policy meets
your needs, you can begin using Tivoli Storage Manager immediately.

The server process of expiration is one way that the server enforces policies that
you define. Expiration processing determines when files are no longer needed, that
is, when the files are expired. For example, if you have a policy that requires only
four copies of a file be kept, the fifth and oldest copy is expired. During expiration
processing, the server removes entries for expired files from the database,
effectively deleting the files from server storage.

You might need more flexibility in your policies than the standard policy provides.
To accommodate individual user's needs, you may fine-tune the STANDARD
policy, or create your own policies. Some types of clients or situations require
special policy. For example, you may want to enable clients to restore backed-up
files to a specific point-in-time.

Policy can be distributed from a configuration manager to managed servers.

Related concepts:
“More on management classes” on page 486
“Protection and expiration of archive data” on page 514
“Distributing policy using enterprise configuration” on page 528
Related tasks:
“Client operations controlled by policy” on page 480
“Getting users started” on page 478
“Changing policy” on page 478
“Assigning client nodes to a policy domain” on page 512
“Policy configuration scenarios” on page 522
“Configuring policy for direct-to-tape backups” on page 522
Related reference:
“Basic policy planning” on page 476
“Reviewing the standard policy” on page 477
“File expiration and expiration processing” on page 479
“The parts of a policy” on page 482
“How Tivoli Storage Manager selects files for policy operations” on page 492
“Creating your own policies” on page 496

© Copyright IBM Corp. 1993, 2013 475

 “Defining and updating a policy domain” on page 498
“Defining and updating a management class” on page 501
“Validating and activating a policy set” on page 510
“Running expiration processing to delete expired files” on page 512
“Configuring policy for Tivoli Storage Manager application clients” on page 523
“Policy for logical volume backups” on page 523
“Configuring policy for NDMP operations” on page 524
“Configuring policy for LAN-free data movement” on page 525
“Policy for Tivoli Storage Manager servers as clients” on page 527
“Setting policy to enable point-in-time restore for clients” on page 527
“Querying policy” on page 529
“Deleting policy” on page 531

Basic policy planning

Start out simply to plan your policy. You may be able to use the default policy that
comes with the server.

Ask the following questions:

v How many backup versions do clients need?
v How long do clients need the backup versions?

Examine the default policy to see if it meets your needs:

v Up to two backup versions of a file on the client’s system are retained in server
storage.
v The most recent backup version is retained for as long as the original file is on
the client file system. All other versions are retained for up to 30 days after they
become inactive.
v One backup version of a file that has been deleted from the client’s system is
retained in server storage for 60 days.
v An archive copy is kept for up to 365 days.

The server manages files based on whether the files are active or inactive. The
most current backup or archived copy of a file is the active version. All other
versions are called inactive versions. An active version of a file becomes inactive
when:
v A new backup is made
v A user deletes that file on the client node and then runs an incremental backup
Policy determines how many inactive versions of files the server keeps, and for
how long. When files exceed the criteria, the files expire. Expiration processing can
then remove the files from the server database.
Related reference:
“File expiration and expiration processing” on page 479
“Running expiration processing to delete expired files” on page 512
“Reviewing the standard policy” on page 477

476 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Reviewing the standard policy
The standard policy consists of a standard policy domain, policy set, management
class, backup copy group, and archive copy group. Each of these parts is named
STANDARD.

The attributes of the default policy are as follows:

Table 41. Summary of default policy
Policy Object where the policy is set
Backup Policies
Files are backed up to the default disk storage pool, STANDARD backup copy group,
BACKUPPOOL. DESTINATION parameter
An incremental backup is performed only if the file STANDARD backup copy group,
has changed since the last backup. MODE parameter
Files cannot be backed up while they are being STANDARD backup copy group,
modified. SERIALIZATION parameter
Up to two backup versions of a file on the client's STANDARD backup copy group,
system are retained in server storage. The most recent the following parameters:
backup version is retained for as long as the original VEREXISTS
file is on the client file system. All other versions are
RETEXTRA
retained for up to 30 days after they become inactive.
RETONLY
One backup version of a file that has been deleted STANDARD backup copy group,
from the client's system is retained in server storage VERDELETED parameter
for 60 days.
When a backed up file is no longer associated with a STANDARD policy domain,
backup copy group, it remains in server storage for 30 BACKRETENTION parameter
days (backup retention grace period).
Archive Policies
Files are archived in the default disk storage pool, STANDARD archive copy group,
ARCHIVEPOOL. DESTINATION parameter
Files cannot be archived while they are being STANDARD archive copy group,
modified. SERIALIZATION parameter
An archive copy is kept for up to 365 days. STANDARD archive copy group,
RETVER parameter
When an archived file is no longer associated with an STANDARD policy domain,
archive copy group, it remains in server storage for ARCHRETENTION parameter
365 days (archive retention grace period).
General
The default management class is STANDARD. STANDARD policy set (ACTIVE),
ASSIGN DEFMGMTCLASS command
Tivoli Storage Manager for Space Management (HSM) Policy
Client files are not space-managed (there are no HSM STANDARD management class,
clients). SPACEMGTECHNIQUE parameter

Related reference:
“The parts of a policy” on page 482

Chapter 13. Implementing policies for client data 477

 Getting users started
When you register a client node, the default is to assign the node to the
STANDARD policy domain. If users register their own workstations during open
registration, they are also assigned to the STANDARD policy domain.

About this task

To help users take advantage of IBM Tivoli Storage Manager, you can further tune
the policy environment by performing the following tasks:
v Define sets of client options for the different groups of users.
v Help users with creating the include-exclude list. For example:
– Create include-exclude lists to help inexperienced users who have simple file
management needs. One way to do this is to define a basic include-exclude
list as part of a client option set. This also gives the administrator some
control over client usage.
– Provide a sample include-exclude list to users who want to specify how the
server manages their files. You can show users who prefer to manage their
own files how to:
- Request information about management classes
- Select a management class that meets backup and archive requirements
- Use include-exclude options to select management classes for their files
For information on the include-exclude list, see the user’s guide for the
appropriate client.
v Automate incremental backup procedures by defining schedules for each policy
domain. Then associate schedules with client nodes in each policy domain.
Related tasks:
“Creating client option sets on the server” on page 465
Chapter 15, “Scheduling operations for client nodes,” on page 559
Related reference:
“The include-exclude list” on page 488

Changing policy
Some types of clients and situations require policy changes. For example, if you
need to direct client data to storage pools different from the default storage pools,
you need to change policy.

About this task

Other situations may also require policy changes. See “Policy configuration
scenarios” on page 522 for details.

To change policy that you have established in a policy domain, you must replace
the ACTIVE policy set. You replace the ACTIVE policy set by activating another
policy set.

Procedure

Perform the following steps:

1. Create or modify a policy set so that it contains the policy that you want to
implement.

478 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 v Create a new policy set either by defining a new policy set or by copying a
policy set.
v Modify an existing policy set (it cannot be the ACTIVE policy set).

Note: You cannot directly modify the ACTIVE policy set. If you want to make
a small change to the ACTIVE policy set, copy the policy to modify it and
follow the steps here.
2. Make any changes that you need to make to the management classes, backup
copy groups, and archive copy groups in the new policy set.
3. Validate the policy set.
4. Activate the policy set. The contents of your new policy set becomes the
ACTIVE policy set.
Related tasks:
“Defining and updating an archive copy group” on page 508
“Policy configuration scenarios” on page 522
Related reference:
“Validating a policy set” on page 510
“Activating a policy set” on page 511
“Defining and updating a management class” on page 501
“Defining and updating a backup copy group” on page 502

File expiration and expiration processing

An expired file is a file that the server no longer needs to keep, according to policy.

Files expire under the following conditions:

v Users delete file spaces from client nodes
v Users expire files by using the EXPIRE command on the client
v A file that is a backup version exceeds the criteria in the backup copy group
(how long a file is kept and how many inactive versions of a file are kept)
v An archived file exceeds the time criteria in the archive copy group (how long
archived copies are kept)
v A backup set exceeds the retention time that is specified for it

Important:
1. A base file is not eligible for expiration until all of its dependent subfiles have
been expired.
2. An archive file is not eligible for expiration if there is a deletion hold on it. If a
file is not held, it will be handled according to existing expiration processing.

The server deletes expired files from the server database only during expiration
processing. After expired files are deleted from the database, the server can reuse
the space in the storage pools that was occupied by expired files. You should
ensure that expiration processing runs periodically to allow the server to reuse
space.

Expiration processing also removes from the database any restartable restore
sessions that exceed the time limit set for such sessions by the RESTOREINTERVAL
server option.
Related concepts:
“Managing client restartable restore sessions” on page 471

Chapter 13. Implementing policies for client data 479

 “Deletion hold” on page 515
“Expiration processing of base files and subfiles” on page 554
Related tasks:
“Reclaiming space in sequential-access storage pools” on page 367
Related reference:
“Running expiration processing to delete expired files” on page 512

Client operations controlled by policy

IBM Tivoli Storage Manager policies govern the backup and restore, archive and
retrieve, and client migration and recall client operations.
Related concepts:
“Backup and restore”
“Archive and retrieve” on page 481
“Client migration and recall” on page 481

Backup and restore

Backup-archive clients can back up and restore files and directories.
Backup-archive clients on UNIX, Linux, and Windows systems can also back up
and restore logical volumes.

Backups allow users to preserve different versions of files as they change.

Backup
To guard against the loss of information, the backup-archive client can copy files,
subdirectories, and directories to media controlled by the server. Backups can be
controlled by administrator-defined policies and schedules, or users can request
backups of their own data.

The backup-archive client provides two types of backup:

Incremental backup
The backup of files according to policy defined in the backup copy group
of the management class for the files. An incremental backup typically
backs up all files that are new or that have changed since the last
incremental backup.
Selective backup
Backs up only files that the user specifies. The files must also meet some of
the policy requirements defined in the backup copy group.

See Backup-Archive Clients Installation and User's Guide for details on backup-archive
clients that can also back up logical volumes. The logical volume must meet some
of the policy requirements that are defined in the backup copy group.
Related reference:
“Policy for logical volume backups” on page 523

480 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 Restore
When a user restores a backup version of a file, the server sends a copy of the file
to the client node. The backup version remains in server storage. Restoring a
logical volume backup works the same way.

If more than one backup version exists, a user can restore the active backup
version or any inactive backup versions.

If policy is properly set up, a user can restore backed-up files to a specific time.

Restriction: If you back up or archive data with a Tivoli Storage Manager V6.3
client, you cannot restore or retrieve that data with a V6.2 or earlier client.
Related reference:
“Setting policy to enable point-in-time restore for clients” on page 527

Archive and retrieve

To preserve files for later use or for records retention, a user with a backup-archive
client can archive files, subdirectories, and directories on media controlled by the
server. When users archive files, they can choose to have the backup-archive client
erase the original files from their workstation after the client archives the files.

When a user retrieves a file, the server sends a copy of the file to the client node.
The archived file remains in server storage.

Client migration and recall

With aTivoli Storage Manager HSM product, a user can migrate files from
workstation storage to server storage and recall those files as needed.

The HSM client frees space for new data and makes more efficient use of your
storage resources.

For details about using Tivoli Storage Manager for Space Management, see Space
Management for UNIX and Linux User's Guide and Tivoli Storage Manager HSM for
Windows User's Guide.

Migration
When a file is migrated to the server, it is replaced on the client node with a small
stub file of the same name as the original file. The stub file contains data needed to
locate the migrated file on server storage.

Tivoli Storage Manager for Space Management provides selective and automatic
migration. Selective migration lets users migrate files by name. The two types of
automatic migration are:
Threshold
If space usage exceeds a high threshold set at the client node, migration
begins and continues until usage drops to the low threshold also set at the
client node.
Demand
If an out-of-space condition occurs for a client node, migration begins and
continues until usage drops to the low threshold.

To prepare for efficient automatic migration, Tivoli Storage Manager for Space
Management copies a percentage of user files from the client node to the IBM

Chapter 13. Implementing policies for client data 481

 Tivoli Storage Manager server. The premigration process occurs whenever Tivoli
Storage Manager for Space Management completes an automatic migration. The
next time free space is needed at the client node, the files that have been
pre-migrated to the server can quickly be changed to stub files on the client. The
default premigration percentage is the difference between the high and low
thresholds.

Files are selected for automatic migration and premigration based on the number
of days since the file was last accessed and also on other factors set at the client
node.

Recall
Tivoli Storage Manager for Space Management provides selective and transparent
recall. Selective recall lets users recall files by name. Transparent recall occurs
automatically when a user accesses a migrated file.

When recalling active file versions, the server searches in an active-data storage
pool associated with a FILE device class, if such a pool exists.
Related concepts:
“Selection for active file versions” on page 243

Reconciliation
Migration and premigration can create inconsistencies between stub files on the
client node and space-managed files in server storage.

For example, if a user deletes a migrated file from the client node, the copy
remains at the server. At regular intervals set at the client node, IBM Tivoli Storage
Manager compares client node and server storage and reconciles the two by
deleting from the server any outdated files or files that do not exist at the client
node.

The parts of a policy

Policy administrators use IBM Tivoli Storage Manager policy to specify how files
are backed up, archived, migrated from client node storage, and managed in server
storage.

Figure 67 on page 483 shows the parts of a policy and the relationships among the
parts.

482 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 Policy domain

Policy sets

Active policy set

Management classes

Default management class

Backup Archive
copy copy
group group Additional Additional
policy policy
set set
Additional
management class

Additional
management class

Figure 67. IBM Tivoli Storage Manager Policy

Backup copy group

Controls the backup processing of files associated with the management
class. A backup copy group determines the following items:
v How frequently a file can be backed up
v How to handle files that are in use during a backup
v Where the server initially stores backup versions of files and directories
v How many backup versions the server keeps of files and directories
v How long the server keeps backup versions of files and directories
Archive copy group
Controls the archive processing of files associated with the management
class. An archive copy group determines the following items:
v How to handle files that are in use during archive
v Where the server stores archived copies of files
v How long the server keeps archived copies of files
Management class
Associates backup and archive groups with files, and specifies if and how
client node files are migrated to storage pools. A management class can
contain one backup or archive copy group, both a backup and archive
copy group, or no copy groups. Users can bind (that is, associate) their
files to a management class through the include-exclude list.
Policy set
Specifies the management classes that are available to groups of users.
Policy sets contain one or more management classes. You must identify one
management class as the default management class. Only one policy set,
the ACTIVE policy set, controls policy operations.
Policy domain
Lets an administrator group client nodes by the policies that govern their

Chapter 13. Implementing policies for client data 483

 files and by the administrators who manage their policies. A policy domain
contains one or more policy sets, but only one policy set (named ACTIVE)
can be active at a time. The server uses only the ACTIVE policy set to
manage files for client nodes assigned to a policy domain.
You can use policy domains to:
v Group client nodes with similar file management requirements
v Provide different default policies for different groups of clients
v Direct files from different groups of clients to different storage
hierarchies based on need (different file destinations with different
storage characteristics)
v Restrict the number of management classes to which clients have access
Related concepts:
“More on management classes” on page 486
Related reference:
“Example: sample policy objects” on page 497
“Running expiration processing to delete expired files” on page 512

Relationships among clients, storage, and policy

The relationship among the physical device environment, storage and policy
objects, and clients are represented in a figure.

The numbers in the following list correspond to the numbers in the figure.

484 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 Clients

Disk
Policy domain
4
Policy set Volume Volume

Management class

Storage Represents
Copy Points to
2 group pool

DISK
device class

Migrate
6
Tape
5
Volume Volume
Management class
Represents

Copy Points to Storage Media

3 group pool

Device class

Library
Represents

Drives
Device
Drive Drive

Figure 68. How clients, server storage, and policy work together

1 When clients are registered, they are associated with a policy domain. In a
policy domain, there is a policy set, management class, and copy groups.

2,
3
When a client backs up, archives, or migrates a file, it is bound to a
management class. A management class and its backup and archive copy
groups, specifies where files are stored, and how they are managed.

4,
5
Storage pools are a collection of volumes where backed-up, archived, or

Chapter 13. Implementing policies for client data 485

 space-managed files are stored. The copy groups of a management class
specify storage pools for backed-up or archived files. For space-managed
files, the management classes specify the storage pools.
Storage pools are mapped to device classes, which represent devices. A
storage pool contains volumes of the media-type indicated by the
associated device class. The following example illustrates this concept:
v A storage pool that is mapped to a device class with a device type of 8
MM contains only 8 mm tapes.

6
Files that are initially stored on disk storage pools can migrate to the
following storage pool if the pools are set up in a storage hierarchy:
v Tape or other types of storage pools

Figure 68 on page 485 summarizes the relationships among the physical device
environment, IBM Tivoli Storage Manager storage and policy objects, and clients.

More on management classes

Management classes are the key connection between client files and policy. Each
client node is assigned to a single policy domain, and the client node has access
only to the management classes contained in the active policy set.

The management classes specify whether client files are migrated to storage pools
(hierarchical storage management). The copy groups in these management classes
specify the number of backup versions retained in server storage and the length of
time to retain backup versions and archive copies.

For example, if a group of users needs only one backup version of their files, you
can create a policy domain that contains only one management class whose backup
copy group allows only one backup version. Then you can assign the client nodes
for these users to the policy domain.
Related tasks:
“Registering nodes with the server” on page 420
Related reference:
“Contents of a management class”
“Default management classes” on page 487
“The include-exclude list” on page 488
“How files and directories are associated with a management class” on page 489

Contents of a management class

A management class contains policy for backup, archive, and space management
operations by clients. You can specify if and how a Tivoli Storage Manager for
Space Management client can migrate files to server storage with parameters in the
management class.

For clients using the server for backup and archive, you can choose what a
management class contains from the following options:
A backup copy group and an archive copy group
Typical end users need to back up and archive documents, spreadsheets,
and graphics.

486 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 A backup copy group only
Some users only want to back up files (such as working documents,
database, log, or history files that change daily). Some application clients
need only a backup copy group because they never archive files.
An archive copy group only
A management class that contains only an archive copy group is useful for
users who create:
v Point-in-time files. For example, an engineer can archive the design of an
electronic component and the software that created the design. Later, the
engineer can use the design as a base for a new electronic component.
v Files that are rarely used but need to be retained for a long time. A client
can erase the original file without affecting how long the archive copy is
retained in server storage. Examples include legal records, patient
records, and tax forms.

Attention: A management class that contains neither a backup nor an archive

copy group prevents a file from ever being backed up or archived. This type of
management class is not recommended for most users. Use such a management
class carefully to prevent users from mistakenly selecting it. If users bind their files
to a management class without copy groups, IBM Tivoli Storage Manager issues
warning messages.

Default management classes

Each policy set must include a default management class.

The default management class is used for the following purposes:

v To manage files that are not bound to a specific management class, as defined by
the INCLUDE option in the include-exclude list.
v To manage existing backup versions when an administrator deletes a
management class or a backup copy group from the server.
v To manage existing archive copies when an administrator deletes a management
class or an archive copy group from the server. The server does not rebind
archive copies, but does use the archive copy group (if one exists) in the default
management class.
v To manage files when a client node is assigned to a new policy domain and the
active policy set does not have management classes with the same names as that
to which the node's files are bound.

A typical default management class should perform the following things:

v Meet the needs of most users
v Contain both a backup copy group and an archive copy group
v Set serialization static or shared static to ensure the integrity of backed up and
archived files
v Retain backup versions and archive copies for a sufficient amount of time
v Retain directories for at least as long as any files are associated with the
directory

Other management classes can contain copy groups tailored either for the needs of
special sets of users or for the needs of most users under special circumstances.
Related reference:
“How files and directories are associated with a management class” on page 489

Chapter 13. Implementing policies for client data 487

 The include-exclude list
A user can define an include-exclude list to specify which files are eligible for the
different processes that the client can run. Include and exclude options in the list
determine which files are eligible for backup and archive services and which files
can be migrated from the client (space-managed).

The options also include how the server controls symbolic links and processing
such as image, compression and encryption.

If a user does not create an include-exclude list, the following default conditions
apply:
v All files belonging to the user are eligible for backup and archive services.
v The default management class governs backup, archive, and space-management
policies.

Figure 69 shows an example of an include-exclude list. The statements in this

example list perform the following actions:
v Excludes certain files or directories from backup, archive, and client migration
operations
Line 1 in Figure 69 means that the SSTEINER node ID excludes all core files
from being eligible for backup and client migration.
v Includes some previously excluded files
Line 2 in Figure 69 means that the files in the following directory are excluded:
– *:\home\ssteiner
The include statement that follows on line 3, however, means that the
[Link] file in that directory is eligible for backup and client migration.
v Binds a file to a specific management class
Line 4 in Figure 69 means that all files and subdirectories belonging to the
following directory are managed by the policy defined in the MCENGBK2
management class:
– *:\home\ssteiner\driver5

exclude *:\...\core
exclude *:\home\ssteiner\*
include *:\home\ssteiner\[Link]
include *:\home\ssteiner\driver5\...\* mcengbk2

Figure 69. Example of an include-exclude list

IBM Tivoli Storage Manager processes the include-exclude list from the bottom up,
and stops when it finds an include or exclude statement that matches the file it is
processing. Therefore, the order in which the include and exclude options are listed
affects which files are included and excluded. For example, suppose you switch the
order of two lines in the example, as follows:
include *:\home\ssteiner\[Link]
exclude *:\home\ssteiner\*

The exclude statement comes last, and excludes all files in the following directory:
v *:\home\ssteiner

488 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 When IBM Tivoli Storage Manager is processing the include-exclude list for the
[Link] file, it finds the exclude statement first. This time, the [Link] file
is excluded.

Some options are evaluated after the more basic include and exclude options. For
example, options that exclude or include files for compression are evaluated after
the program determines which files are eligible for the process being run.

You can create include-exclude lists as part of client options sets that you define
for clients.

For detailed information on the include and exclude options, see the user’s guide
for the appropriate client.
Related tasks:
“Creating client option sets on the server” on page 465

How files and directories are associated with a management

class
Binding is the process of associating a file with a management class. The policies
defined in the management class then apply to the bound files. The server binds a
file to a management class when a client backs up, archives, or migrates the file.

A client chooses a management class as follows:

v For backing up a file, a client can specify a management class in the client's
include-exclude list (include-exclude options file for UNIX and Linux clients), or
can accept the default management class.
v For backing up directories, the client can specify a management class by using
the DIRMC option in the client options file.

Important: It is recommended that you define a default management class. If no

management class is specified for a directory, the server chooses the
management class with the longest retention period in the backup copy group
(retention period for the only backup version). When two or more management
classes have the same, "longest" retention period, the Tivoli Storage Manager
client selects the management class whose name is last in alphabetical order.
v For backing up a file system or logical volume, a client can specify a
management class in the client's include-exclude list (include-exclude options file
for UNIX and Linux clients), or can accept the default management class.
v For archiving a file, the client can do one of the following tasks:
– Specify a management class in the client's include-exclude list (with either an
include option or an [Link] option)
– Specify a management class with the ARCHMC option on the archive command
– Accept the default management class
v For archiving directories, the client can specify a management class with the
archiving options, or the ARCHMC option.

Important: It is recommended that you define a default management class. If

the client does not specify any archiving options, the server assigns the default
management class to the archived directory. If the default management class has
no archive copy group, the server assigns the management class that currently
has the archive copy group with the shortest retention time. When two or more

Chapter 13. Implementing policies for client data 489

 management classes have the same, "shortest" retention period, the Tivoli
Storage Manager client selects the management class whose name is last in
alphabetical order.
v For migrating a file, a client can specify a management class in the client's
include-exclude options file, or can accept the default management class.

The default management class is the management class identified as the default in
the active policy set.

A management class specified with a simple include option can apply to one or
more processes on the client. More specific include options (such as
[Link]) allow the user to specify different management classes. Some
examples of how this works:
v If a client backs up, archives, and migrates a file to the same server, and uses
only a single include option, the management class specified for the file applies
to all three operations (backup, archive, and migrate).
v If a client backs up and archives a file to one server, and migrates the file to a
different server, the client can specify one management class for the file for
backup and archive operations, and a different management class for migrating.
v Clients can specify a management class for archiving that is different from the
management class for backup.
See the user's guide for the appropriate client for more details.

Effects of changing a management class

A file remains bound to a management class even if the attributes of the
management class or its copy groups change.

Procedure

The following scenario illustrates this process:

1. A file named [Link] is bound to the default management class that
contains a backup copy group specifying that up to three backup versions can
be retained in server storage.
2. During the next week, three backup versions of [Link] are stored in server
storage. The active and two inactive backup versions are bound to the default
management class.
3. The administrator assigns a new default management class that contains a
backup copy group specifying only up to two backup versions.
4. The administrator then activates the policy set, and the new default
management class takes effect.
5. [Link] is backed up again, bringing the number of versions to four. The
server determines that according to the new backup copy group only two
versions are to be retained. Therefore, the server marks the two oldest versions
for deletion (expired).
6. Expiration processing occurs. [Link] is still bound to the default
management class, which now includes new retention criteria. Therefore, the
two versions marked for deletion are purged, and one active and one inactive
backup version remain in storage.
Related reference:
“Running expiration processing to delete expired files” on page 512

490 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Rebinding files to management classes
Rebinding is the process of associating all versions of a file or a logical volume
image with a new management class.

Backup versions

The server rebinds backup versions of files and logical volume images in some
cases.

The following list highlights the cases when a server rebinds backup versions of
files:
v The user changes the management class specified in the include-exclude list and
does a backup.
v An administrator activates a policy set in the same policy domain as the client
node, and the policy set does not contain a management class with the same
name as the management class to which a file is currently bound.
v An administrator assigns a client node to a different policy domain, and the
active policy set in that policy domain does not have a management class with
the same name.

Backup versions of a directory can be rebound when the user specifies a different
management class using the DIRMC option in the client option file, and when the
directory gets backed up.

The most recently backed up files are active backup versions. Older copies of your
backed up files are inactive backup versions. You can configure management classes
to save a predetermined number of copies of a file. If a management class is saving
five backup copies, there would be one active copy saved and four inactive copies
saved. If a file from one management class is bound to a different management
class that retains a lesser number of files, inactive files are deleted.

If a file is bound to a management class that no longer exists, the server uses the
default management class to manage the backup versions. When the user does
another backup, the server rebinds the file and any backup versions to the default
management class. If the default management class does not have a backup copy
group, the server uses the backup retention grace period specified for the policy
domain.

Archive copies

Archive copies are never rebound because each archive operation creates a
different archive copy. Archive copies remain bound to the management class
name specified when the user archived them.

If the management class to which an archive copy is bound no longer exists or no

longer contains an archive copy group, the server uses the default management
class. If you later change or replace the default management class, the server uses
the updated default management class to manage the archive copy.

If the default management class does not contain an archive copy group, the server
uses the archive retention grace period specified for the policy domain.

Chapter 13. Implementing policies for client data 491

How Tivoli Storage Manager selects files for policy operations
The IBM Tivoli Storage Manager selects files for full and partial incremental
backups, selective backups, logical volume backups, archives, and automatic
migration from an HSM client (Tivoli Storage Manager for Space Management).

Incremental backup
Backup-archive clients can choose to back up their files using full or partial
incremental backup. A full incremental backup ensures that clients' backed-up files
are always managed according to policies. Clients are urged to use full incremental
backup whenever possible.

If the amount of time for backup is limited, clients may sometimes need to use
partial incremental backup. A partial incremental backup should complete more
quickly and require less memory. When a client uses partial incremental backup,
only files that have changed since the last incremental backup are backed up.
Attributes in the management class that would cause a file to be backed up when
doing a full incremental backup are ignored. For example, unchanged files are not
backed up even when they are assigned to a management class that specifies
absolute mode and the minimum days between backups (frequency) has passed.

The server also does less processing for a partial incremental backup. For example,
the server does not expire files or rebind management classes to files during a
partial incremental backup.

If clients must use partial incremental backups, they should periodically perform
full incremental backups to ensure that complete backups are done and backup
files are stored according to policies. For example, clients can do partial
incremental backups every night during the week, and a full incremental backup
on the weekend.

Performing full incremental backups is important if clients want the ability to

restore files to a specific time. Only a full incremental backup can detect whether
files have been deleted since the last backup. If full incremental backup is not done
often enough, clients who restore to a specific time may find that many files that
had actually been deleted from the workstation get restored. As a result, a client's
file system may run out of space during a restore process.
Related reference:
“Setting policy to enable point-in-time restore for clients” on page 527

Full incremental backup

When a user requests a full incremental backup, the IBM Tivoli Storage Manager
determines its eligibility.

Procedure

The IBM Tivoli Storage Manager ensures the following items are identified:
1. Checks each file against the user's include-exclude list:
v Files that are excluded are not eligible for backup.
v If files are not excluded and a management class is specified with the
INCLUDE option, IBM Tivoli Storage Manager uses that management class.
v If files are not excluded but a management class is not specified with the
INCLUDE option, IBM Tivoli Storage Manager uses the default management
class.

492 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 v If no include-exclude list exists, all files in the client domain are eligible for
backup, and IBM Tivoli Storage Manager uses the default management class.
2. Checks the management class of each included file:
v If there is a backup copy group, the process continues with step 3.
v If there is no backup copy group, the file is not eligible for backup.
3. Checks the mode, frequency, and serialization defined in the backup copy
group.
Mode Specifies whether the file is backed up only if it has changed since the
last backup (modified) or whenever a backup is requested (absolute).
Frequency
Specifies the minimum number of days that must elapse between
backups.

Tip: For Windows this attribute is ignored during a journal-based

backup.
Serialization
Specifies how files are handled if they are modified while being backed
up and what happens if modification occurs.
v If the mode is modified and the minimum number of days have elapsed
since the file was last backed up, IBM Tivoli Storage Manager determines if
the file has been changed since it was last backed up:
– If the file has been changed and the serialization requirement is met, the
file is backed up.
– If the file has not been changed, it is not backed up.
v If the mode is modified and the minimum number of days have not elapsed
since the file was last backed up, the file is not eligible for backup.
v If the mode is absolute, the minimum number of days have elapsed since the
file was last backed up, and the serialization requirement is met, the file is
backed up.
v If the mode is absolute and the minimum number of days have not elapsed
since the file was last backed up, the file is not eligible for backup.

Partial incremental backup

When a user requests a partial incremental backup, the IBM Tivoli Storage
Manager determines its eligibility.

Procedure

Tivoli Storage Manager ensures the following items are identified:

1. Checks each file against the user’s include-exclude list:
v Files that are excluded are not eligible for backup.
v If files are not excluded and a management class is specified with the
INCLUDE option, the server uses that management class.
v If files are not excluded but a management class is not specified with the
INCLUDE option, the server uses the default management class.
v If no include-exclude list exists, all files in the client domain are eligible for
backup, and the server uses the default management class.
2. Checks the management class of each included file:

Chapter 13. Implementing policies for client data 493

 v If there is a backup copy group, the process continues with step 3.
v If there is no backup copy group, the file is not eligible for backup.
3. Checks the date and time of the last incremental backup by the client, and the
serialization requirement defined in the backup copy group. (Serialization
specifies how files are handled if they are modified while being backed up and
what happens if modification occurs.)
v If the file has not changed since the last incremental backup, the file is not
backed up.
v If the file has changed since the last incremental backup and the serialization
requirement is met, the file is backed up.

Selective backup
When a user requests a selective backup, the IBM Tivoli Storage Manager ensures
its eligibility.

Procedure

IBM Tivoli Storage Manager ensures the following items are identified:
1. Checks the file against any include or exclude statements contained in the user
include-exclude list:
v Files that are not excluded are eligible for backup. If a management class is
specified with the INCLUDE option, IBM Tivoli Storage Manager uses that
management class.
v If no include-exclude list exists, the files selected are eligible for backup, and
IBM Tivoli Storage Manager uses the default management class.
2. Checks the management class of each included file:
v If the management class contains a backup copy group and the serialization
requirement is met, the file is backed up. Serialization specifies how files are
handled if they are modified while being backed up and what happens if
modification occurs.
v If the management class does not contain a backup copy group, the file is
not eligible for backup.

Results

An important characteristic of selective backup is that a file is backed up without

regard for whether the file has changed. This result may not always be what you
want. For example, suppose a management class specifies to keep three backup
versions of a file. If the client uses incremental backup, the file is backed up only
when it changes, and the three versions in storage will be at different levels. If the
client uses selective backup, the file is backed up regardless of whether it has
changed. If the client uses selective backup on the file three times without
changing the file, the three versions of the file in server storage are identical.
Earlier, different versions are lost.

494 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Logical volume backup
When a user requests a logical volume backup, the IBM Tivoli Storage Manager
determines its eligibility.

Procedure

IBM Tivoli Storage Manager ensures the following items are identified:
1. Checks the specification of the logical volume against any include or exclude
statements contained in the user include-exclude list:
v If no include-exclude list exists, the logical volumes selected are eligible for
backup, and IBM Tivoli Storage Manager uses the default management class.
v Logical volumes that are not excluded are eligible for backup. If the
include-exclude list has an INCLUDE option for the volume with a
management class specified, IBM Tivoli Storage Manager uses that
management class. Otherwise, the default management class is used.
2. Checks the management class of each included logical volume:
v If the management class contains a backup copy group and the logical
volume meets the serialization requirement, the logical volume is backed up.
Serialization specifies how logical volumes are handled if they are modified
while being backed up and what happens if modification occurs.
v If the management class does not contain a backup copy group, the logical
volume is not eligible for backup.

Archive
When a user requests the archiving of a file or a group of files, the IBM Tivoli
Storage Manager determine its eligibility.

Procedure

IBM Tivoli Storage Manager ensures the following items are identified:
1. Checks the files against the user’s include-exclude list to see if any
management classes are specified:
v IBM Tivoli Storage Manager uses the default management class for files that
are not bound to a management class.
v If no include-exclude list exists, IBM Tivoli Storage Manager uses the default
management class unless the user specifies another management class. See
the user’s guide for the appropriate client for details.
2. Checks the management class for each file to be archived.
v If the management class contains an archive copy group and the serialization
requirement is met, the file is archived. Serialization specifies how files are
handled if they are modified while being archived and what happens if
modification occurs.
v If the management class does not contain an archive copy group, the file is
not archived.

Results

If you need to frequently create archives for the same data, consider using instant
archive (backup sets) instead. Frequent archive operations can create a large
amount of metadata in the server database resulting in increased database growth
and decreased performance for server operations such as expiration. Frequently,
you can achieve the same objectives with incremental backup or backup sets.

Chapter 13. Implementing policies for client data 495

 Although the archive function is a powerful way to store inactive data with fixed
retention, it should not be used on a frequent and large scale basis as the primary
backup method.
Related concepts:
“Creating and using client backup sets” on page 543

Automatic migration from a client node

A file is eligible for automatic migration from an HSM client (Tivoli Storage
Manager for Space Management) if it meets certain criteria.

The criteria for a file to be eligible for automatic migration from an HSM client are
displayed in the following list:
v It resides on a node on which the root user has added and activated hierarchical
storage management. It must also reside in a local file system to which the root
user has added space management, and not in the root (/) or /tmp file system.
v It is not excluded from migration in the include-exclude list.
v It meets management class requirements for migration:
– The file is not a character special file, a block special file, a FIFO special file
(that is, a named pipe file) or a directory.
– The file is assigned to a management class that calls for space management.
– The management class calls for automatic migration after a specified number
of days, and that time has elapsed.
– A backup version of the file exists if the management class requires it.
– The file is larger than the stub file that would replace it (plus one byte) or the
file system block size, whichever is larger.

Creating your own policies

You can create your own policies by defining the parts of a policy and specifying
each attribute, or by copying existing policy parts and updating only those
attributes that you want to change.

Task Required Privilege Class

Define or copy a policy domain System
Update a policy domain over which you Restricted policy
have authority
Define, update, or copy policy sets and System or unrestricted policy
management classes in any policy domain
Define, update, or copy policy sets and Restricted policy
management classes in policy domains over
which you have authority
Define or update copy groups in any policy System or unrestricted policy
domain
Define or update copy groups that belong to Restricted policy
policy domains over which you have
authority
Assign a default management class to a System or unrestricted policy
nonactive policy set in any policy domain
Assign a default management class to a Restricted policy
nonactive policy set in policy domains over
which you have authority

496 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 Task Required Privilege Class
Validate and activate policy sets in any System or unrestricted policy
policy domain
Validate and activate policy sets in policy Restricted policy
domains over which you have authority
Start inventory expiration processing System

Table 42 shows that an advantage of copying existing policy parts is that some
associated parts are copied in a single operation.
Table 42. Cause and effect of copying existing policy parts
If you copy this... Then you create this...
Policy Domain A new policy domain with:
v A copy of each policy set from the original domain
v A copy of each management class in each original policy set
v A copy of each copy group in each original management class
Policy Set A new policy set in the same policy domain with:
v A copy of each management class in the original policy set
v A copy of each copy group in the original management class
Management Class A new management class in the same policy set and a copy of each
copy group in the management class

Example: sample policy objects

The sample policy objects example is used in several scenarios.

Figure 70 on page 498 shows the policies for an engineering department.

The domain contains two policy sets that are named STANDARD and TEST. The
administrator activated the policy set that is named STANDARD. When you
activate a policy set, the server makes a copy of the policy set and names it
ACTIVE. Only one policy set can be active at a time.

The ACTIVE policy set contains two management classes: MCENG and
STANDARD. The default management class is STANDARD.

Chapter 13. Implementing policies for client data 497

 Figure 70. An example of policy objects defined for an engineering department

Related tasks:
“Defining and updating an archive copy group” on page 508
Related reference:
“Defining and updating a policy domain”
“Defining and updating a policy set” on page 500
“Defining and updating a management class” on page 501
“Defining and updating a backup copy group” on page 502
“Assigning a default management class” on page 510
“Activating a policy set” on page 511
“Running expiration processing to delete expired files” on page 512

Defining and updating a policy domain

When you update or define a policy domain, you specify the backup retention
grace period, the archive retention grace period, and the destination for active
backup data.

See the following definitions:

Backup Retention Grace Period
Specifies the number of days to retain an inactive backup version when the
server cannot rebind the file to an appropriate management class. The
backup retention grace period protects backup versions from being
immediately expired when the management class to which a file is bound
no longer exists or no longer contains a backup copy group, and the
default management class does not contain a backup copy group.

498 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 Backup versions of the file managed by the grace period are retained in
server storage only for the backup retention grace period. This period
starts from the day of the backup. For example, if the backup retention
grace period for the STANDARD policy domain is used and set to 30 days,
backup versions using the grace period expire in 30 days from the day of
the backup.
Backup versions of the file continue to be managed by the grace period
unless one of the following events occur:
v The client binds the file to a management class containing a backup
copy group and then backs up the file
v A backup copy group is added to the file's management class
v A backup copy group is added to the default management class
Archive Retention Grace Period
Specifies the number of days to retain an archive copy when the
management class for the file no longer contains an archive copy group
and the default management class does not contain an archive copy group.
The retention grace period protects archive copies from being immediately
expired.
The archive copy of the file managed by the grace period is retained in
server storage for the number of days specified by the archive retention
grace period. This period starts from the day on which the file is first
archived. For example, if the archive retention grace period for the policy
domain STANDARD is used, an archive copy expires 365 days from the
day the file is first archived.
The archive copy of the file continues to be managed by the grace period
unless an archive copy group is added to the file's management class or to
the default management class.
Destination for Active Backup Data
Specifies the names active-data pools that store active versions of backup
data for nodes assigned to the domain. Before the IBM Tivoli Storage
Manager server writes data to an active-data pool, it verifies that the node
owning the data is assigned to a domain that has the active-data pool
defined in the list of active-data pools. If the server verifies that the node
meets this criteria, the data is stored in the active-data pool. If the node
does not meet the criteria, then the data is not stored in the active-data
pool.
If client backup is performed using simultaneous-write operations to an
active-data pool, the server performs the verification during backup
operations by IBM Tivoli Storage Manager backup-archive clients or by
application clients using the IBM Tivoli Storage Manager API. The
verification is also performed when active data is being copied using the
COPY ACTIVEDATA command.

Chapter 13. Implementing policies for client data 499

 Example: defining a policy domain
To create a new policy domain you can either copy an existing policy domain and
update the new domain, or define a new policy domain from scratch.

About this task

When you copy an existing domain, you also copy any associated policy sets,
management classes, and copy groups.

Procedure

For example, perform the following steps to copy and update an existing domain:
1. Copy the STANDARD policy domain to the ENGPOLDOM policy domain by
entering the following command:
copy domain standard engpoldom
ENGPOLDOM now contains the standard policy set, management class,
backup copy group, and archive copy group.
2. Update the policy domain ENGPOLDOM so that the backup retention grace
period is extended to 90 days and the archive retention grace period is
extended to two years. Specify an active-data pool as the destination for active
versions of backup data belonging to nodes assigned to the domain. Use
engactivedata as the name of the active-data pool, as in the following example:
update domain engpoldom description=’Engineering Policy Domain’
backretention=90 archretention=730 activedestination=engactivedata

Defining and updating a policy set

When you define or update a policy set, you must specify the policy domain
name.

See the following definition:

Policy domain name
Names the policy domain to which the policy set belongs

The policies in the new policy set do not take effect unless you make the new set
the ACTIVE policy set.
Related reference:
“Activating a policy set” on page 511

Example: defining a policy set

An administrator must develop new policies based on the existing STANDARD
policy set.

Procedure

To create the TEST policy set in the ENGPOLDOM policy domain, the
administrator performs the following steps:
1. Copy the STANDARD policy set and name the new policy set TEST:
copy policyset engpoldom standard test

Note: When you copy an existing policy set, you also copy any associated
management classes and copy groups.
2. Update the description of the policy set named TEST:

500 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 update policyset engpoldom test
description=’Policy set for testing’

Defining and updating a management class

When you define or update a management class, you must specify the policy
domain name, the policy set name, and the description.

See the following definitions:

Policy domain name
Names the policy domain to which the management class belongs.
Policy set name
Names the policy set to which the management class is assigned.
Description
Describes the management class. A clear description can help users to
choose an appropriate management class for their use.

The following four parameters apply only to HSM clients:

Whether space management is allowed
Specifies that the files are eligible for both automatic and selective
migration, only selective migration, or no migration.
How frequently files can be migrated
Specifies the minimum number of days that must elapse since a file was
last accessed before it is eligible for automatic migration.
Whether backup is required
Specifies whether a backup version of a file must exist before the file can
be migrated.
Where migrated files are to be stored
Specifies the name of the storage pool in which migrated files are stored.
Your choice could depend on factors such as:
v The number of client nodes migrating to the storage pool. When many
user files are stored in the same storage pool, volume contention can
occur as users try to migrate files to or recall files from the storage pool.
v How quickly the files must be recalled. If users need immediate access
to migrated versions, you can specify a disk storage pool as the
destination.

Attention: You cannot specify a copy storage pool or an active-data pool

as a destination.

Example: define a new management class

There are just two steps to creating a new management class.

Procedure

Perform the following steps to create a new management class:

1. Define a new management class MCENG by entering:
define mgmtclass engpoldom standard mceng
2. Update the description of the MCENG management class by entering:
update mgmtclass engpoldom standard mceng
description=’Engineering Management Class for Backup and Archive’

Chapter 13. Implementing policies for client data 501

 Defining and updating a backup copy group
When you are defining and updating a backup copy group, you will have to know
where to store it, how to manage files that are modified during backup, how to
designate the frequency of your backups, and how to retain the backup versions.
Related reference:
“Where to store backed-up files”
“How to manage files that are modified during backup”
“Defining the frequency of backing up files” on page 503
“Retaining backup versions” on page 504

Where to store backed-up files

Specify a storage pool where the server initially stores the files associated with this
backup copy group. This is called the destination.

Your choice can depend on factors such as the following items:

v Whether the server and the client nodes have access to shared devices on a
storage area network (SAN).
v The number of client nodes backing up to the storage pool. When many user
files are stored in the same storage pool, volume contention can occur as users
try to back up to or restore files from the storage pool.
v How quickly the files must be restored. If users need immediate access to
backup versions, you may want to specify a disk storage pool as the destination.

Attention: You cannot specify a copy storage pool or an active-data pool as the
destination.

How to manage files that are modified during backup

You can use the SERIALIZATION attribute on the DEFINE COPYGROUP command to
specify how files are managed if they are modified during a backup.

This attribute can be one of four values: STATIC, SHRSTATIC (shared static),
DYNAMIC, or SHRDYNAMIC (shared dynamic).

The value you choose depends on how you want IBM Tivoli Storage Manager to
manage files that are modified while they are being backed up.
Do not back up files that are modified during the backup
You will want to prevent the server from backing up a file while it is being
modified. Use one of the following values:
STATIC
Specifies that IBM Tivoli Storage Manager will attempt to back up
the file only once. If the file or directory is modified during a
backup, the server does not back it up.
SHRSTATIC (Shared static)
Specifies that if the file or directory is modified during a backup,
the server retries the backup as many times as specified by the
CHANGINGRETRIES option in the client options file. If the file is
modified during the last attempt, the file or directory is not backed
up.
Back up files that are modified during the backup
Some files are in constant use, such as an error log. Consequently, these

502 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 files may never be backed up when serialization is set to STATIC or
SHRSTATIC. To back up files that are modified during the backup, use one
of the following values:
DYNAMIC
Specifies that a file or directory is backed up on the first attempt,
even if the file or directory is modified during the backup.
SHRDYNAMIC (Shared dynamic)
Specifies that if a file or directory is modified during a backup, the
server retries the backup as many times as specified by the
CHANGINGRETRIES option in the client options file. The server backs
up the file on the last attempt, even if the file or directory is being
modified.

Attention:
v If a file is modified during backup and DYNAMIC or SHRDYNAMIC is
specified, then the backup may not contain all the changes and may not
be usable. For example, the backup version may contain a truncated
record. Under some circumstances, it may be acceptable to capture a
dynamic or “fuzzy” backup of a file (the file was changed during the
backup). For example, a dynamic backup of an error log file that is
continuously appended may be acceptable. However, a dynamic backup
of a database file may not be acceptable, since restoring such a backup
could result in an unusable database. Carefully consider dynamic
backups of files as well as possible problems that may result from
restoring potentially “fuzzy” backups.
v When certain users or processes open files, they may deny any other
access, including “read” access, to the files by any other user or process.
When this happens, even with serialization set to DYNAMIC or
SHRDYNAMIC, IBM Tivoli Storage Manager will not be able to open
the file at all, so the server cannot back up the file.

Defining the frequency of backing up files

You can specify how frequently files can be backed up with two parameters,
FREQUENCY and MODE.

See the following definitions:

Frequency
The frequency is the minimum number of days that must elapse between
full incremental backups.

Note: This attribute is ignored during a journal-based backup.

Mode The mode parameter specifies whether a file or directory must have been
modified to be considered for backup during a full incremental backup
process. IBM Tivoli Storage Manager does not check this attribute when a
user requests a partial incremental backup, a selective backup for a file, or
a backup of a logical volume. You can select from two modes:
Modified
A file is considered for full incremental backup only if it has
changed since the last backup. A file is considered changed if any
of the following items is different:
v Date on which the file was last modified
v File size

Chapter 13. Implementing policies for client data 503

 v File owner
v File permissions
Absolute
A file is considered for full incremental backup regardless of
whether it has changed since the last backup.

The server considers both parameters to determine how frequently files can be
backed up. For example, if frequency is 3 and mode is Modified, a file or directory
is backed up only if it has been changed and if three days have passed since the
last backup. If frequency is 3 and mode is Absolute, a file or directory is backed up
after three days have passed whether or not the file has changed.

Use the Modified mode when you want to ensure that the server retains multiple,
different backup versions. If you set the mode to Absolute, users may find that
they have three identical backup versions, rather than three different backup
versions.

Absolute mode can be useful for forcing a full backup. It can also be useful for
ensuring that extended attribute files are backed up, because Tivoli Storage
Manager does not detect changes if the size of the extended attribute file remains
the same.

When you set the mode to Absolute, set the frequency to 0 if you want to ensure
that a file is backed up each time full incremental backups are scheduled for or
initiated by a client.

Retaining backup versions

Multiple versions of files are useful when users continually update files and
sometimes need to restore the original file from which they started. The most
current backup version of a file is called the active version. All other versions are
called inactive versions.

You can specify the number of versions to keep by:

v Directly specifying the number of versions
You specify the number of backup versions with two parameters:
– Versions Data Exists (number of versions to keep when the data still exists
on the client node)
– Versions Data Deleted (number of versions to keep when the data no longer
exists on the client node)
v Specifying the number of days to keep each backup version
You specify the number of days to keep backup versions with two parameters:
– Retain Extra Versions (how many days to keep inactive backup versions; the
days are counted from the day that the version became inactive)
– Retain Only Versions (how many days to keep the last backup version of a
file that has been deleted)
v Specifying a combination of the number of versions and the days to keep them
Use a combination of the four parameters: Versions Data Exists, Versions Data
Deleted, Retain Extra Versions, and Retain Only Versions.

These parameters interact to determine the backup versions that the server retains.
When the number of inactive backup versions exceeds the number of versions
allowed (Versions Data Exists and Versions Data Deleted), the oldest version
expires and the server deletes the file from the database the next time expiration

504 IBM Tivoli Storage Manager for Windows: Administrator's Guide

processing runs. How many inactive versions the server keeps is also related to the
parameter for how long inactive versions are kept (Retain Extra Versions).
Inactive versions expire when the number of days that they have been inactive
exceeds the value specified for retaining extra versions, even when the number of
versions is not exceeded.

Important: A base file is not eligible for expiration until all its dependent subfiles
have been expired.

For example, see Table 43 and Figure 71. A client node has backed up the file
[Link] four times in one month, from March 23 to April 23. The settings in the
backup copy group of the management class to which [Link] is bound
determine how the server treats these backup versions. Table 44 on page 506 shows
some examples of how different copy group settings would affect the versions. The
examples show the effects as of April 24 (one day after the file was last backed
up).
Table 43. Status of [Link] as of april 24
Days the Version Has Been
Version Date Created Inactive
Active April 23 (not applicable)
Inactive 1 April 13 1 (since April 23)
Inactive 2 March 31 11 (since April 13)
Inactive 3 March 23 24 (since March 31)

Policy domain Server storage

[Link]
Active policy set Tuesday March 23

Wednesday March 31
Default management
Tuesday April 13
class
Friday April 23

Backup
copy group

Active version
Inactive
versions

Figure 71. Active and inactive versions of [Link]

Chapter 13. Implementing policies for client data 505

Table 44. Effects of backup copy group policy on backup versions for [Link] as of april 24. One day after the file
was last backed up.
Versions Versions Retain Retain Results
Data Data Extra Only
Exists Deleted Versions Version
4 versions 2 versions 60 days 180 days Versions Data Exists and Retain Extra Versions control the
expiration of the versions. The version created on March 23 is
retained until the client node backs up the file again (creating a
fourth inactive version), or until that version has been inactive for
60 days.

If the user deletes the [Link] file from the client node, the
server notes the deletion at the next full incremental backup of the
client node. From that point, the Versions Data Deleted and
Retain Only Version parameters also have an effect. All versions
are now inactive. Two of the four versions expire immediately (the
March 23 and March 31 versions expire). The April 13 version
expires when it has been inactive for 60 days (on June 23). The
server keeps the last remaining inactive version, the April 23
version, for 180 days after it becomes inactive.
NOLIMIT 2 versions 60 days 180 days Retain Extra Versions controls expiration of the versions. The
inactive versions (other than the last remaining version) are
expired when they have been inactive for 60 days.

If the user deletes the [Link] file from the client node, the
server notes the deletion at the next full incremental backup of the
client node. From that point, the Versions Data Deleted and
Retain Only Version parameters also have an effect. All versions
are now inactive. Two of the four versions expire immediately (the
March 23 and March 31 versions expire) because only two versions
are allowed. The April 13 version expires when it has been
inactive for 60 days (on June 22). The server keeps the last
remaining inactive version, the April 23 version, for 180 days after
it becomes inactive.
NOLIMIT NOLIMIT 60 days 180 days Retain Extra Versions controls expiration of the versions. The
server does not expire inactive versions based on the maximum
number of backup copies. The inactive versions (other than the
last remaining version) are expired when they have been inactive
for 60 days.

If the user deletes the [Link] file from the client node, the
server notes the deletion at the next full incremental backup of the
client node. From that point, the Retain Only Version parameter
also has an effect. All versions are now inactive. The three of four
versions will expire after each of them has been inactive for 60
days. The server keeps the last remaining inactive version, the
April 23 version, for 180 days after it becomes inactive.
4 versions 2 versions NOLIMIT NOLIMIT Versions Data Exists controls the expiration of the versions until
a user deletes the file from the client node. The server does not
expire inactive versions based on age.

If the user deletes the [Link] file from the client node, the
server notes the deletion at the next full incremental backup of the
client node. From that point, the Versions Data Deleted parameter
controls expiration. All versions are now inactive. Two of the four
versions expire immediately (the March 23 and March 31 versions
expire) because only two versions are allowed. The server keeps
the two remaining inactive versions indefinitely.

506 IBM Tivoli Storage Manager for Windows: Administrator's Guide

See Administrator's Reference for details about the parameters. The following list
gives some tips on using the NOLIMIT value:
Versions Data Exists
Setting the value to NOLIMIT may require increased storage, but that
value may be needed for some situations. For example, to enable client
nodes to restore files to a specific point in time, set the value for Versions
Data Exists to NOLIMIT. Setting the value this high ensures that the
server retains versions according to the Retain Extra Versions parameter
for the copy group.
Versions Data Deleted
Setting the value to NOLIMIT may require increased storage, but that
value may be needed for some situations. For example, set the value for
Versions Data Deleted to NOLIMIT to enable client nodes to restore files
to a specific point in time. Setting the value this high ensures that the
server retains versions according to the Retain Extra Versions parameter
for the copy group.
Retain Extra Versions
If NOLIMIT is specified, inactive backup versions are deleted based on the
Versions Data Exists or Versions Data Deleted parameters.
To enable client nodes to restore files to a specific point in time, set the
parameters Versions Data Exists or Versions Data Deleted to NOLIMIT.
Set the value for Retain Extra Versions to the number of days that you
expect clients may need versions of files available for possible
point-in-time restoration. For example, to enable clients to restore files from
a point in time 60 days in the past, set Retain Extra Versions to 60.
Retain Only Version
If NOLIMIT is specified, the last version is retained forever unless a user
or administrator deletes the file from server storage.
Related concepts:
“Enabling clients to use subfile backup” on page 552

Example: define a backup copy group

Define a backup copy group belonging to the MCENG management class in the
STANDARD policy set belonging to the ENGPOLDOM policy domain.

About this task

This new copy group must be able to complete the following tasks:
v Let users back up changed files, regardless of how much time has elapsed since
the last backup, using the default value 0 for the Frequency parameter
(frequency parameter not specified)
v Retain up to four inactive backup versions when the original file resides on the
user workstation, using the Versions Data Exists parameter (verexists=5)
v Retain up to four inactive backup versions when the original file is deleted from
the user workstation, using the Versions Data Deleted parameter
(verdeleted=4)
v Retain inactive backup versions for no more than 90 days, using the Retain
Extra Versions parameter (retextra=90)

Chapter 13. Implementing policies for client data 507

 v If there is only one backup version, retain it for 600 days after the original is
deleted from the workstation, using the Retain Only Version parameter
(retonly=600)
v Prevent files from being backed up if they are in use, using the Serialization
parameter (serialization=static)
v Store files in the ENGBACK1 storage pool, using the Destination parameter
(destination=engback1)

Procedure

Issue the following command to define the backup copy group:

define copygroup engpoldom standard mceng standard
destination=engback1 serialization=static
verexists=5 verdeleted=4 retextra=90 retonly=600

Defining and updating an archive copy group

To define or update an archive copy group, you must specify where the archived
files are to be stored, if files can be modified during archive, how long to retain an
archived copy, and the minimum amount of days to retain an object.

Procedure
1. Where archived files are to be stored, specify a defined storage pool as the
initial destination. Your choice can depend on factors such as:
v Whether the server and the client nodes have access to shared devices on a
SAN
v The number of client nodes archiving files to the storage pool. When many
user files are stored in the same storage pool, volume contention can occur as
users archive files to and retrieve files from the storage pool.
v How quickly the files must be restored. If users need immediate access to
archive copies, you could specify a disk storage pool as the destination.
v Whether the archive copy group is for a management class that is the default
for a policy domain. The default management class is used by clients
registered in the policy domain, when they do not specify a management
class for a file. This includes servers that are registered as clients to this
server.

Note: You cannot specify a copy storage pool or an active-data pool as a

destination.
2. If files can be modified during archive, specify how files are handled if they are
modified while being archived. This attribute, called serialization, can be one of
four values:
Static Specifies that if the file is modified during an archiving process, the
server does not archive it. IBM Tivoli Storage Manager does not retry
the archive.
Shared Static
Specifies that if the file is modified during an archive process, the
server does not archive it. However, IBM Tivoli Storage Manager retries
the archive process as many times as specified by the CHANGINGRETRIES
option in the client options file.
Dynamic
Specifies that a file is archived on the first attempt, even if the file is
being modified during the archive process.

508 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 Shared Dynamic
Specifies that if the file is modified during the archive attempt, the
server archives it on its last try even if the file is being modified. IBM
Tivoli Storage Manager retries the archive process as many times as
specified by the CHANGINGRETRIES option in the client options file.
For most files, set serialization to either static or shared static to prevent the
server from archiving a file while it is being modified.
However, you may want to define a copy group with a serialization of shared
dynamic or dynamic for files where log records are continuously added, such
as an error log. If you only have copy groups that use static or shared static,
these files may never be archived because they are constantly in use. With
shared dynamic or dynamic, the log files are archived. However, the archive
copy may contain a truncated message.
Attention: If a file is archived while it is in use (shared dynamic or dynamic
serialization), the copy may not contain all the changes and may not be usable.

Note: When certain users or processes open files, they deny read access to the
files for any other user or process. When this happens, even with serialization
set to dynamic or shared dynamic, the server does not back up the file.
3. How long to retain an archived copy specifies the number of days to retain an
archived copy in storage. When the time elapses, the archived copy expires and
the server deletes the file the next time expiration processing runs.
When a user archives directories, the server uses the default management class
unless the user specifies otherwise. If the default management class does not
have an archive copy group, the server binds the directory to the management
class that currently has the shortest retention time for archive. When you
change the retention time for an archive copy group, you may also be changing
the retention time for any directories that were archived using that copy group.
The user can change the archive characteristics by using Archive Options in the
interface or by using the ARCHMC option on the command.
4. The RETMIN parameter in archive copy groups specifies the minimum number of
days an object will be retained after the object is archived. For objects that are
managed by event-based retention policy, this parameter ensures that objects
are retained for a minimum time period regardless of when an event triggers
retention
After you have defined an archive copy group, using the RETMIN=n parameter,
ensure that the appropriate archive data will be bound to the management class
with this archive copy group. You can do this either by using the default
management class or by modifying the client options file to specify the
management class for the appropriate archive data.
Placing a deletion hold on an object does not extend its retention period. For
example, if an object is thirty days away from the end of its retention period
and it is placed on hold for ninety days, it will be eligible for expiration
immediately upon the hold being released.
Related concepts:
“Deletion hold” on page 515
Related tasks:
“Using virtual volumes to store data on another server” on page 732

Chapter 13. Implementing policies for client data 509

 Example: define an archive copy group
Define an archive copy group belonging to the MCENG class.

This copy group must:

v Allow users to archive a file if it is not in use (serialization=static)
v Retain the archive copy for 730 days (retver=730)
v Store files in the ENGARCH1 storage pool (destination=engarch1)
To define a STANDARD archive copy group to the MCENG management class in
the STANDARD policy set belonging to the ENGPOLDOM policy domain, enter:
define copygroup engpoldom standard mceng standard
type=archive destination=engarch1 serialization=static
retver=730

Assigning a default management class

After you have defined a policy set and the management classes that it contains,
you must assign a default management class for the policy set.
Related reference:
“Default management classes” on page 487

Example: assign a default management class

The example is to assign the STANDARD management class as the default
management class for the TEST policy set in the ENGPOLDOM policy domain.

Procedure

Issue the following command to complete the example:

assign defmgmtclass engpoldom standard standard

The STANDARD management class was copied from the STANDARD policy set to
the TEST policy set. Before the new default management class takes effect, you
must activate the policy set.
Related tasks:
“Example: defining a policy set” on page 500

Validating and activating a policy set

After you have defined a policy set and defined management classes to it, you can
validate the policy set and activate the policy set for the policy domain. Only one
policy set is active in a policy domain.

Validating a policy set

When you validate a policy set, the server examines the management class and
copy group definitions in the policy set and reports on conditions that need to be
considered if the policy set is activated.

Validation fails if the policy set does not contain a default management class.
Validation results in result in warning messages if any of the following conditions
exist.

Condition Reason for warning

The storage destinations specified for A backup, archive, or migration operation
backup, archive, or migration do not refer to will fail when the operation involves storing
defined storage pools. a file in a storage pool that does not exist.

510 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Condition Reason for warning
A storage destination specified for backup, The storage destination must be a primary
archive, or migration is a copy storage pool storage pool.
or an active-data pool.
The default management class does not When the default management class does not
contain a backup or archive copy group. contain a backup or archive copy group, any
user files bound to the default management
class are not backed up or archived.
The current ACTIVE policy set names a When users back up files that were bound to
management class that is not defined in the a management class that no longer exists in
policy set being validated. the active policy set, backup versions are
rebound to the default management class.

When the management class to which an

archive copy is bound no longer exists and
the default management class does not
contain an archive copy group, the archive
retention grace period is used to retain the
archive copy.
The current ACTIVE policy set contains copy When users perform a backup and the
groups that are not defined in the policy set backup copy group no longer exists in the
being validated. management class to which a file is bound,
backup versions are managed by the default
management class. If the default
management class does not contain a backup
copy group, backup versions are managed by
the backup retention grace period, and the
workstation file is not backed up.
A management class specifies that a backup The contradictions within the management
version must exist before a file can be classes can cause problems for Tivoli Storage
migrated from a client node, but the Manager for Space Management (HSM)
management class does not contain a backup users.
copy group.

Related reference:
“How files and directories are associated with a management class” on page 489
“Defining and updating a policy domain” on page 498

Activating a policy set

To activate a policy set, specify a policy domain and policy set name.

When you activate a policy set, the server performs a final validation of the
contents of the policy set and copies the original policy set to the ACTIVE policy
set.

You cannot update the ACTIVE policy set; the original and the ACTIVE policy sets
are two separate objects. For example, updating the original policy set has no effect
on the ACTIVE policy set. To change the contents of the ACTIVE policy set, you
must create or change another policy set and then activate that policy set.

If data retention protection is active, the following rules apply during policy set
validation and activation. The server can be a managed server and receive policy
definitions via enterprise configuration, but it will not be possible to activate
propagated policy sets if these rules are not satisfied.
v All management classes in the policy set to be validated and activated must
contain an archive copy group.

Chapter 13. Implementing policies for client data 511

 v If a management class exists in the active policy set, a management class with
the same name must exist in the policy set to be validated and activated.
v If an archive copy group exists in the active policy set, the corresponding copy
group in the policy set to be validated and activated must have RETVER and
RETMIN values at least as large as the corresponding values in the active copy
group.
Related concepts:
“Data retention protection” on page 514
Related tasks:
“Changing policy” on page 478

Example: validating and activating a policy set

Validating and activating the STANDARD policy set in the ENGPOLDOM policy
domain requires a two-step process.

Perform the following steps to complete the example:

1. To validate the STANDARD policy set, enter:
validate policyset engpoldom standard

Examine any messages that result and correct the problems.

2. To activate the STANDARD policy set, enter:
activate policyset engpoldom standard

Assigning client nodes to a policy domain

You can assign existing client nodes to a new policy domain, or create new client
nodes to be associated with an existing policy domain.

About this task

You can use the server command line to assign client nodes to a policy domain.

Procedure
v To assign the client node APPCLIENT1 to the ENGPOLDOM policy domain,
issue the following command:
update node appclient1 domain=engpoldom
v To create a new client node, NEWUSER, and assign it to the ENGPOLDOM
policy domain, issue the following command:
register node newuser newuser domain=engpoldom

Running expiration processing to delete expired files

Expiration processing deletes expired client files from the server storage. Expiration
processing also removes from the database any restartable restore sessions that
exceed the time limit for saving such sessions.

You can run expiration processing either automatically or by command. You

should ensure that expiration processing runs periodically to allow the server to
reuse storage pool space that is occupied by expired client files.

Note:
1. A base file is not eligible for expiration until all of its dependent subfiles have
been expired.

512 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 2. An archive file is not eligible for expiration if there is a deletion hold on it. If a
file is not held, it will be handled according to existing expiration processing.
Related concepts:
“Expiration processing of base files and subfiles” on page 554
“Deletion hold” on page 515

Running expiration processing automatically

You control automatic expiration processing by using the expiration interval option
(EXPINTERVAL) in the server options file ([Link]). You can also control
when restartable restore sessions expire with another server option,
RESTOREINTERVAL.

You can set the options by editing the [Link] file (see the Administrator's
Reference).

If you use the server options file to control automatic expiration, the server runs
expiration processing each time that you start the server. After that, the server runs
expiration processing at the interval you specified with the option, which is
measured from the start time of the server.

Using commands to control expiration processing

You can manually run expiration by issuing the EXPIRE INVENTORY command or
you can set a specific schedule for expiration processing by using the DEFINE
SCHEDULE command.

After issuing EXPIRE INVENTORY, expired files are deleted from the database
according to how you specify parameters on the command.

You can control how long the expiration process runs by using the DURATION
parameter with the EXPIRE INVENTORY command. You can run several (up to 40)
expiration processes in parallel by specifying RESOURCE=x, where x equals the
number of nodes that you want to process. Inventory expiration can also be
distributed across more than one resource on a file space level to help distribute
the workload for nodes with many file spaces.

You can use the DEFINE SCHEDULE command to set a specific schedule for this
command. This automatically starts inventory expiration processing. If you
schedule the EXPIRE INVENTORY command, set the expiration interval to 0 (zero) in
the server options so that the server does not run expiration processing when you
start the server.

When expiration processing runs, the server normally sends detailed messages
about policy changes made since the last time expiration processing ran. You can
reduce those messages by using the QUIET=YES parameter with the EXPIRE
INVENTORY command, or the following options:
v The Use Quiet Expiration option in the server options
When you use the quiet option or parameter, the server issues messages about
policy changes during expiration processing only when files are deleted, and either
the default management class or retention grace period for the domain has been
used to expire the files.

Chapter 13. Implementing policies for client data 513

 Additional expiration processing with disaster recovery
manager
If you have disaster recovery manager (DRM), one or more database backup
volumes may also be deleted during expiration processing.

These volumes may be deleted if the following conditions are true:

v The volume has a device type of SERVER
v The volume is not part of the most recent database backup series
v The last volume of the database backup series has exceeded the expiration value
specified with the SET DRMDBBACKUPEXPIREDAYS command
Related tasks:
“Moving copy storage pool and active-data pool volumes on-site” on page 1028

Protection and expiration of archive data

There are two separate ways to protect Tivoli Storage Manager archive objects so
that they will not be inadvertently deleted. One method is to activate data
retention protection on a server level.

For example, securities brokers and other regulated institutions enforce retention
requirements for certain records, including electronic mail, customer statements,
trade settlements, check images and new account forms. Data retention protection
prevents deliberate or accidental deletion of data until its specified retention
criterion is met.

Another method of additional protection is to place a deletion hold on an object

using the client API. For example, federal regulatory requirements allow a
broker-dealer to delete records when the regulatory retention period has lapsed,
but other legal requirements might mandate that the records continue to be
maintained. By using deletion hold, you ensure that data is not deleted until the
hold is released.

Data retention protection

Data retention protection ensures that archive objects are not deleted from the
Tivoli Storage Manager server until policy-based retention requirements for that
object are satisfied. It is enforced at the server level by the SET
ARCHIVERETENTIONPROTECTION command.

See the Administrator's Reference for more information.

Retention protection can only be activated on a new server that does not already
have stored objects (backup, archive, or space-managed). Activating retention
protection applies to all archive objects subsequently stored on that server. After
retention protection is set, the server cannot store backup objects, space-managed
objects, or backupsets. Retention protection cannot be added for an object that was
previously stored on a Tivoli Storage Manager server. After an object is stored with
retention protection, retention protection cannot be removed.

514 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 Retention protection is based on the retention criterion for each object, which is
determined by the RETVER parameter of the archive copy group of the management
class to which the object is bound. If an object uses event-based retention, the
object does not expire until whatever comes later: either the date the object was
archived plus the number of days in the RETMIN parameter or the date the event
was signaled plus the number of days specified in the RETVER parameter. On
servers that have retention protection enabled, the following operations will not
delete objects whose retention criterion is not satisfied:
v Requests from the client to delete an archive object
v DELETE FILESPACE (from either a client or administrative command)
v DELETE VOLUME DISCARDDATA=YES
v AUDIT VOLUME FIX=YES

Important: A cached copy of data can be deleted, but data in primary storage
pools, copy storage pools, and active-data pools can only be marked damaged
and is never deleted.

If your server has data retention protection activated, the following items are
restrictions:
v A registered node cannot be reassigned to a different policy domain.
v You cannot define a device class with a device type of SERVER.
v You can export data but it will not be retention protected after it is imported.
v You cannot import data to a retention protected server.

The server does not send a retention value to an EMC Centera storage device if
retention protection is not enabled. If this is the case, you can use a Centera
storage device as a standard device from which archive and backup files can be
deleted.
Related tasks:
Chapter 29, “Protecting and recovering the server infrastructure and client data,”
on page 869

Deletion hold
If a hold is placed on an object through the client API, the object is not deleted
until the hold is released.

See the Backup-Archive Clients Installation and User's Guide for more information.
There is no limit to how often you alternate holding and releasing an object. An
object can have only one hold on it at a time, so if you attempt to hold an object
that is already held, you will get an error message.

If an object with event-based policy is on hold, an event can still be signalled. The
hold will not extend the retention period for an object. If the retention period
specified in the RETVER and RETMIN parameters expires while the object is on hold,
the object will be eligible for deletion whenever the hold is released.

If an object is held, it will not be deleted whether or not data retention protection
is active. If an object is not held, it is handled according to existing processing such
as normal expiration, data retention protection, or event-based retention. Data that
is in deletion hold status can be exported. The hold status will be preserved when
the data is imported to another system.

The following deletion operations are prevented if a hold is on an object:

Chapter 13. Implementing policies for client data 515

 v Requests from the client to delete an archive object
v DELETE FILESPACE (from either a client or administrative command)
v DELETE VOLUME DISCARDDATA=YES
v AUDIT VOLUME FIX=YES

Note: A cached copy of data can be deleted, but data in primary storage pools,
copy storage pools, and active-data pools can only be marked damaged and is
never deleted.

Protecting data using the NetApp SnapLock licensed feature

The NetApp SnapLock licensed feature helps meet federal regulatory requirements
for archived data. The SnapLock feature allows Tivoli Storage Manager to set a
retention date for files and to commit a file to a WORM (write once, read many)
state.

Data stored with a retention date cannot be deleted from the file system before the
retention period expires. The SnapLock feature can only be used by Tivoli Storage
Manager servers that have data retention protection enabled.

Data archived by data retention protection servers and stored to NetApp NAS file
servers is stored as Tivoli Storage Manager FILE volumes. At the end of a write
transaction, a retention date is set for the FILE volume, through the SnapLock
interface. This date is calculated by using the RETVER and RETMIN parameters of the
archive copy group used when archiving the data. Having a retention date
associated with the FILE volume gives it a characteristic of WORM media by not
allowing the data to be destroyed or overwritten until the retention date has
passed. These FILE volumes are referred to as WORM FILE volumes. After a
retention date has been set, the WORM FILE volume cannot be deleted until the
retention date has passed. System Storage Archive Manager combined with
WORM FILE volume reclamation ensures protection for the life of the data.

Storage pools can be managed either by threshold or by data retention period. The
RECLAMATIONTYPE storage pool parameter indicates that a storage pool is managed
based on a data retention period. When a traditional storage pool is queried with
the FORMAT=DETAILED parameter, this output is displayed:
Reclamation Type: THRESHOLD

Tivoli Storage Manager servers that have data retention protection enabled through
System Storage Archive Manager and have access to a NetApp filer with the
SnapLock licensed feature can define a storage pool with RECLAMATIONTYPE set
to SNAPLOCK. This means that data created on volumes in this storage pool are
managed by retention date. When a SnapLock storage pool is queried with the
FORMAT=DETAILED parameter, the output displayed indicates that the storage
pools are managed by data retention period.
Reclamation Type: SNAPLOCK

See the NetApp document Data ONTAP Storage Management Guide for details on
the SnapLock filer. Note this is NetApp documentation.

Attention: It is not recommended that you use this feature to protect data with a
retention period of less than three months.
Related concepts:
“Data retention protection” on page 514

516 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Reclamation and the SnapLock feature
It is recommended that you set the NetApp default retention period to 30 days to
match the WORM FILE default reclamation period. Tivoli Storage Manager
reclaims any remaining data on a WORM FILE volume just before the retention
date expiration.

The reclamation of a WORM FILE volume to another WORM FILE volume before
the retention date expiration ensures that data is always protected by the SnapLock
feature.

Because this protection is at a Tivoli Storage Manager volume level, the data on the
volumes can be managed by Tivoli Storage Manager policy without consideration
of where the data is stored. Data stored on WORM FILE volumes is protected both
by data retention protection and by the retention period stored with the physical
file on the SnapLock volume. If a Tivoli Storage Manager administrator issues a
command to delete the data, the command fails. If someone attempt to delete the
file through a series of network file system calls, the SnapLock feature prevents the
data from being deleted.

During reclamation processing, if the Tivoli Storage Manager server cannot move
data from an expiring SnapLock volume to a new SnapLock volume, a warning
message is issued.

Retention periods
Tivoli Storage Manager policies manage the retention time for the WORM FILE
volume. The retention of some files might exceed the retention time for the WORM
FILE volume they were stored on. This could require moving them to another
volume to ensure that the files are stored on WORM media.

Some objects on the volume might need to be retained longer than other objects on
the volume for the following reasons:
v They are bound to management classes with different retention times.
v They cannot be removed because of a deletion hold.
v They are waiting for an event to occur before expiring.
v The retention period for a copy group is increased, requiring a longer retention
time than that specified in the SnapLock feature when the WORM FILE volume
was committed.

Use the DEFINE STGPOOL command to set up a storage pool for use with the
SnapLock feature. Selecting RECLAMATIONTYPE=SNAPLOCK enables Tivoli
Storage Manager to manage FILE volumes by a retention date. After a storage pool
has been set up as a SnapLock storage pool, the RECLAMATIONTYPE parameter
cannot be updated to THRESHOLD. When a SnapLock storage pool is defined, a
check is made to ensure that the directories specified in the device class are
SnapLock WORM volumes. When a file class is defined and storage pools are
created with the reclamation type of SNAPLOCK, all volumes must be WORM
volumes or the operation fails. If a device class is updated to contain additional
directories and there are SnapLock storage pools assigned to it, the same check is
made to ensure all directories are SnapLock WORM volumes.

There are three retention periods available in the NetApp SnapLock feature. These
must be configured correctly so that the Tivoli Storage Manager server can
properly manage WORM data stored in SnapLock volumes. The Tivoli Storage
Manager server sets the retention period for data being stored on NetApp
SnapLock volumes based on the values in the copy group for the data being
Chapter 13. Implementing policies for client data 517
 archived. The NetApp filer should not conflict with the ability of the Tivoli Storage
Manager server to set the retention period. The following settings are the Tivoli
Storage Manager recommendations for retention periods in the NetApp filer:
1. Minimum Retention Period Set the higher value: either 30 days or the
minimum number of days specified by any copy group (using a NetApp
SnapLock filer for WORM FILE storage) for the data retention period. The copy
group is the one in use storing data on NetApp SnapLock volumes.
2. Maximum Retention Period Leave default of 30 years. This allows the Tivoli
Storage Manager server to set the actual volume retention period based on the
settings in the archive copy group.
3. Default Retention Period Set to 30 days. If you do not set this value and you do
not set the maximum retention period, each volume's retention period will be
set to 30 years. If this occurs, the Tivoli Storage Manager server's ability to
manage expiration and reuse of NetApp SnapLock volumes will be largely
defeated in that no volume will be able to be reused for thirty years.

With the NetApp SnapLock retention periods appropriately set, Tivoli Storage
Manager can manage the data in SnapLock storage pools with maximum efficiency.
For each volume that is in a SNAPLOCK storage pool, a Tivoli Storage Manager
reclamation period is created. The Tivoli Storage Manager reclamation period has a
start date, BEGIN RECLAIM PERIOD, and an end date, END RECLAIM PERIOD.
View these dates by issuing the QUERY VOLUME command with the
FORMAT=DETAILED parameter on a SnapLock volume. For example:
Begin Reclaim Period: 09/05/2010
End Reclaim Period: 10/06/2010

When Tivoli Storage Manager archives files to a SnapLock volume, it keeps track
of the latest expiration date of those files, and the BEGIN RECLAIM PERIOD is set
to that latest expiration date. When more files are added to the SnapLock volume,
the starting date is set to that later date if there is a file with a later expiration date
than the one currently on the volume. The start date is set to the latest expiration
date for any file on that volume. The expectation is that all files on that volume
should have already either expired, or should be expiring on that day and the
following day there should be no valid data left on that volume.

The END RECLAIM PERIOD is set to a month later than the BEGIN RECLAIM
PERIOD. The retention date set in the NetApp filer for that volume is set to the
END RECLAIM PERIOD date. This means the NetApp filer will prevent any
deletion of that volume until the END RECLAIM PERIOD date has passed. This is
approximately a month after the data has actually expired in the Tivoli Storage
Manager server. If an END RECLAIM PERIOD date is calculated by the Tivoli
Storage Manager server for a volume, and the date is later than the current END
RECLAIM PERIOD, the new date will be reset in the NetApp filer for that volume
to the later date. This guarantees that the Tivoli Storage Manager WORM FILE
volume will not be deleted until all data on the volume has expired, or the data
has been moved to another SnapLock volume.

The Tivoli Storage Manager reclamation period is the amount of time between the
begin date and the end date. It is also the time period which the Tivoli Storage
Manager server has to delete volumes on which all the data has expired, or to
move files which have not expired on expiring SnapLock volumes to new
SnapLock volumes with new dates. This month is critical to how the server safely
and efficiently manages the data on WORM FILE volumes. Data on a SnapLock
volume typically expires by the time the beginning date arrives, and the volume

518 IBM Tivoli Storage Manager for Windows: Administrator's Guide

should be empty. When the end date arrives, the volume can be safely deleted
from the Tivoli Storage Manager inventory and the SnapLock filer.

However, some events may occur which mean that there is still valid data on a
SnapLock volume:
1. Expiration processing in the Tivoli Storage Manager server for that volume may
have been delayed or has not completed yet.
2. The retention parameters on the copy group or associated management classes
may have been altered for a file after it was archived, and that file is not going
to expire for some period of time.
3. A deletion hold may have been placed on one or more of the files on the
volume.
4. Reclamation processing has either been disabled or is encountering errors
moving data to new SnapLock volumes on a SnapLock storage pool.
5. A file is waiting for an event to occur before the Tivoli Storage Manager server
can begin the expiration of the file.

If there are files which have not expired on a SnapLock volume when the
beginning date arrives, they must be moved to a new SnapLock volume with a
new begin and end date. This will properly protect that data. However, if
expiration processing on the Tivoli Storage Manager server has been delayed, and
those files will expire as soon as expiration processing on the Tivoli Storage
Manager server runs, it is inefficient to move those files to a new SnapLock
volume. To ensure that unnecessary data movement does not occur for files which
are due to expire, movement of files on expiring SnapLock volumes will be
delayed some small number of days after the BEGIN RECLAIM PERIOD date.
Since the data is protected in the SnapLock filer until the END RECLAIM PERIOD
date, there is no risk to the data in delaying this movement. This allows Tivoli
Storage Manager expiration processing to complete. After that number of days, if
there is still valid data on an expiring SnapLock volume, it will be moved to a new
SnapLock volume, thus continuing the protection of the data.

Since the data was initially archived, there may have been changes in the retention
parameters for that data (for example, changes in the management class or copy
pool parameters) or there may be a deletion hold on that data. However, the data
on that volume will only be protected by SnapLock until the END RECLAIM
PERIOD date. Data that has not expired is moved to new SnapLock volumes
during the Tivoli Storage Manager reclamation period. If errors occur moving data
to a new SnapLock volume, a distinct warning message is issued indicating that
the data will soon be unprotected. If the error persists, it is recommended that you
issue a MOVE DATA command for the problem volume.

Attention: Disabling reclamation processing on a SnapLock storage pool is not

recommended because after the processing is disabled, the Tivoli Storage Manager
server has no way to issue warning messages that data will become unprotected.
This situation can also occur if reclamation and migration is disabled for the entire
server (for example, NOMIGRRECL set in the server options file). Be very careful
when managing SnapLock storage pools so that data doesn't inadvertently become
unprotected.

Chapter 13. Implementing policies for client data 519

 Configuring SnapLock for event-based retention
Data stored in SnapLock volumes that are managed by System Storage Archive
Manager and event-based retention can result in excessive reclamation, which
causes performance degradation of the server.

If data is managed by event-based retention, Tivoli Storage Manager initially sets

the retention period to the greater of the RETVER and RETMIN values for the
archive copy group. When the volume enters the reclamation period and data that
remains on the volume is moved, the retention period for the target volume is set
to the remaining retention period of the data, which is typically 0. The new
volume then enters the reclamation period shortly after receiving the data,
resulting in the reclamation of volumes that were just created.

You can avoid this situation by using the RETENTIONEXTENSION server option. This
option allows the server to set or extend the retention date of a SnapLock volume.
You can specify from 30 to 9999 days. The default is 365 days.

When selecting volumes in a SnapLock storage pool for reclamation, the server
checks if the volume is within the reclamation period.
v If the volume is not within the reclamation period, no action is taken. The
volume is not reclaimed, and the retention date is unchanged
v If the volume is within the reclamation period, the server checks if the percent of
reclaimable space on the volume is greater than the reclamation threshold of the
storage pool or of the threshold percentage passed in on the THRESHOLD
parameter of a RECLAIM STGPOOL command.
– If the reclaimable space is greater than the threshold, the server reclaims the
volume and sets the retention date of the target volume is set to the greater of
these values:
- The remaining retention time of the data plus 30 days for the reclamation
period.
- The RETENTIONEXTENSION value plus 30 days for the reclamation period.
– If the reclaimable space is not greater than the threshold, the server resets the
retention date of the volume by the amount specified in the
RETENTIONEXTENSION option. The new retention period is calculated by adding
the number of days specified to the current date.

In the examples described below, the SnapLock volume, VolumeA, is in a storage

pool whose reclamation threshold is set to 60%. The RETENTIONEXTENSION server
option is set to 365 days. The retention period VolumeA is in the reclamation
period. The following situations show how retention is affected:
v The reclaimable space on VolumeA is less than 60%. The retention date of
VolumeA is extended by 365 days.
v The reclaimable space on VolumeA is greater than 60%, and the remaining
retention time of the data is more than 365 days. VolumeA is reclaimed, and the
retention date of the target volume is set based on the remaining retention of the
data plus 30 days for the reclamation period.
v The reclaimable space on VolumeA is greater than 60%, and the retention time of
the data is less than 365 days. VolumeA is reclaimed, and its retention date is set
to 365 days, the RETENTIONEXTENTION value, plus 30 days for the reclamation
period.

520 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 Ensuring continuous data protection
Data that is stored on a volume with the SnapLock feature enabled and moved or
copied to a non-SnapLock volume loses the unique hardware protection that is
available through the NetApp WORM volumes.

The Tivoli Storage Manager server allows this type of movement, but if data is
moved from a WORM FILE volume to another type of media, the data may no
longer be protected from inadvertent or malicious deletion. If this data is on
WORM volumes to meet data retention and protection requirements for certain
legal purposes and is moved to other media, the data may no longer meet those
requirements. You should configure your storage pools so this type of data is kept
in storage pools which consist of SnapLock WORM volumes during the entire data
retention period.

Set up SnapLock volumes as Tivoli Storage Manager WORM

FILE volumes
When defining or updating configurations that involve SnapLock storage pools,
you should ensure that the storage pools selected for the NEXTSTGPOOL,
RECLAIMSTGPOOL, and COPYSTGPOOLS parameters have the RECLAMATIONTYPE=SNAPLOCK
option specified.

About this task

When you configure the storage pools this way, you ensure that your data is
properly protected. If you define a next, reclaim, copy storage pool, or active-data
pool without selecting the RECLAMATIONTYPE=SNAPLOCK option, you will not have a
protected storage pool. The command succeeds, but a warning message is issued.

Procedure

Complete the following steps to set up a SnapLock volume for use as a Tivoli
Storage Manager WORM FILE volume:
1. Install and set up SnapLock on the NetApp filer. See NetApp documentation
for more information.
2. Properly configure the minimum, maximum, and default retention periods. If
these retention periods are not configured properly, Tivoli Storage Manager will
not be able to properly manage the data and volumes.
3. Install and configure a Tivoli Storage Manager server with data retention
protection. Ensure the SET ARCHIVERETENTIONPROTECTION command is activated.
4. Set up policy by using the DEFINE COPYGROUP command. Select RETVER and
RETMIN values in the archive copy group which will meet your requirements
for protecting this data in WORM storage. If the RETVER or RETMIN values
are not set, the default management classes values will be used.
5. Set up storage by using the DEFINE DEVCLASS command.
v Use the FILE device class.
v Specify the DIRECTORY parameter to point to the directory or directories on
the SnapLock volumes.
6. Define a storage pool using the device class you defined above.
v Specify RECLAMATIONTYPE=SNAPLOCK.
7. Update the copy group to point to the storage pool you just defined.

Chapter 13. Implementing policies for client data 521

 8. Use the Tivoli Storage Manager API to archive your objects into the SnapLock
storage pool. This feature is not available on standard Tivoli Storage Manager
backup-archive clients.
Related reference:
“Retention periods” on page 517

Policy configuration scenarios

The scenarios that are published are designed to show you some cases for which
policy changes may be needed.
Related tasks:
“Configuring policy for direct-to-tape backups”
Related reference:
“Configuring policy for Tivoli Storage Manager application clients” on page 523
“Policy for logical volume backups” on page 523
“Configuring policy for NDMP operations” on page 524
“Configuring policy for LAN-free data movement” on page 525
“Policy for Tivoli Storage Manager servers as clients” on page 527
“Setting policy to enable point-in-time restore for clients” on page 527

Configuring policy for direct-to-tape backups

The server default policy enables client nodes to back up data to disk storage pools
on the server. As an alternative, you can configure a policy to store client data
directly in tape storage pools to reduce contention for disk resources.

About this task

If you back up directly to tape, the number of clients that can back up data at the
same time is equal to the number of drives available to the storage pool (through
the mount limit of the device class). For example, if you have one drive, only one
client at a time can back up data.

The direct-to-tape backup eliminates the need to migrate data from disk to tape.
However, performance of tape drives is often lower when backing up directly to
tape than when backing up to disk and then migrating to tape. Backing up data
directly to tape usually means more starting and stopping of the tape drive.
Backing up to disk then migrating to tape usually means the tape drive moves
more continuously, meaning better performance.

At the server command line, you can define a new policy domain that enables
client nodes to back up or archive data directly to tape storage pools. For example,
you can define a policy domain named DIR2TAPE with the following steps:

Procedure
1. Copy the default policy domain STANDARD as a template:
copy domain standard dir2tape
This command creates the DIR2TAPE policy domain that contains a default
policy set, management class, backup and archive copy group, each named
STANDARD.
2. Update the backup or archive copy group in the DIR2TAPE policy domain to
specify the destination to be a tape storage pool. For example, to use a tape
storage pool named TAPEPOOL for backup, issue the following command:

522 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 update copygroup dir2tape standard standard destination=tapepool
To use a tape storage pool named TAPEPOOL for archive, issue the following
command:
update copygroup dir2tape standard standard type=archive
destination=tapepool
3. Activate the changed policy set.
activate policyset dir2tape standard
4. Assign client nodes to the DIR2TAPE policy domain. For example, to assign a
client node named TAPEUSER1 to the DIR2TAPE policy domain, issue the
following command:
update node tapeuser1 domain=dir2tape

Configuring policy for Tivoli Storage Manager application

clients
The Tivoli Storage Manager application clients that use the server to store data
might require that you configure policies to make the most efficient use of server
storage. See the documentation for each application client for the policy
requirements.

Some of the application clients include a timestamp in each database backup.

Because the default policy for the server keeps one backup version of each unique
file, database backups that are managed by default policy are never deleted. The
backups are not deleted because each backup is uniquely named with its
timestamp. To ensure that the server deletes backups as required, configure policy
as recommended by the documentation for the application client.

Policy for logical volume backups

Consider defining a management class specifically for logical volume backups. To
enable clients to restore a logical volume and then reconcile the results of any file
backup operations since the logical volume backup was made, you must set up
management classes with the backup copy group set up differently from the
STANDARD.

The Versions Data Exists, Versions Data Deleted, and Retain Extra Versions
parameters work together to determine over what time period a client can restore a
logical volume image and reconcile later file backups. Also, you may have server
storage constraints that require you to control the number of backup versions
allowed for logical volumes. The server handles logical volume backups the same
as regular incremental or selective backups. Logical volume backups differ from
selective, incremental, or archive operations in that each file space that is backed
up is treated as a single large file.

Backups of logical volumes are intended to help speed the restoration of a

computer. One way to use the capability is to have users periodically (for example,
once a month) perform a logical volume backup, and schedule daily full
incremental backups. If a user restores a logical volume, the program first restores
the logical volume backup and then any files that were changed since the backup
(incremental or other file backup processes). The user can also specify that the
restore process reconcile any discrepancies that can result when files are deleted.

For example, a user backs up a logical volume, and the following week deletes one
or more files from the volume. At the next incremental backup, the server records
in its database that the files were deleted from the client. When the user restores
the logical volume, the program can recognize that files have been deleted since

Chapter 13. Implementing policies for client data 523

 the backup was created. The program can delete the files as part of the restore
process. To ensure that users can use the capability to reconcile later incremental
backups with a restored logical volume, you need to ensure that you coordinate
policy for incremental backups with policy for backups for logical volumes.

For example, you decide to ensure that clients can choose to restore files and
logical volumes from any time in the previous 60 days. You can create two
management classes, one for files and one for logical volumes. Table 45 shows the
relevant parameters. In the backup copy group of both management classes, set the
Retain Extra Versions parameter to 60 days.

In the management class for files, set the parameters so that the server keeps
versions based on age rather than how many versions exist. More than one backup
version of a file may be stored per day if clients perform selective backups or if
clients perform incremental backups more than once a day. The Versions Data
Exists parameter and the Versions Data Deleted parameter control how many of
these versions are kept by the server. To ensure that any number of backup
versions are kept for the required 60 days, set both the Versions Data Exists
parameter and the Versions Data Deleted parameter to NOLIMIT for the
management class for files. This means that the server retains backup versions
based on how old the versions are, instead of how many backup versions of the
same file exist.

For logical volume backups, the server ignores the frequency attribute in the
backup copy group.
Table 45. Example of backup policy for files and logical volumes
Parameter (backup copy Management Class for Files Management Class for
group in the management Logical Volumes
class)
Versions Data Exists NOLIMIT 3 versions
Versions Data Deleted NOLIMIT 1
Retain Extra Versions 60 days 60 days
Retain Only Version 120 days 120 days

Configuring policy for NDMP operations

You can register a network-attached storage (NAS) file server as a node, using
network data management protocol (NDMP) operations. Under the direction of the
Tivoli Storage Manager server, the NAS file server performs backup and restore of
file system and directory images to a tape library.

The Tivoli Storage Manager server initiates the backup, allocates a drive, and
selects and mounts the media. The NAS file server then transfers the data to tape.

Because the NAS file server performs the backup, the data is stored in its own
format. For most NAS file servers, the data is stored in the NDMPDUMP data
format. For NetApp file servers, the data is stored in the NETAPPDUMP data
format. For EMC file servers, the data is stored in the CELERRADUMP data
format. To manage NAS file server image backups, copy groups for NAS nodes
must point to a storage pool that has a data format of NDMPDUMP,
NETAPPDUMP, or CELERRADUMP.

The following backup copy group attributes are ignored for NAS images:

524 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 v Frequency
v Mode
v Retain Only Versions
v Serialization
v Versions Data Deleted

To set up the required policy for NAS nodes, you can define a new, separate policy
domain.

Backups for NAS nodes can be initiated from the server, or from a client that has
at least client owner authority over the NAS node. For client-initiated backups, you
can use client option sets that contain include and exclude statements to bind NAS
file system or directory images to a specific management class. The valid options
that can be used for a NAS node are: [Link], [Link], and
[Link]. NAS backups initiated from the Tivoli Storage Manager server with
the BACKUP NODE command ignore client options specified in option files or client
option sets. For details on the options see the Backup-Archive Clients Installation and
User's Guide for your particular client platform.

When the Tivoli Storage Manager server creates a table of contents (TOC), you can
view a collection of individual files and directories backed up via NDMP and
select which to restore. To establish where to send data and store the table of
contents, policy should be set so that:
v Image backup data is sent to a storage pool with a NDMPDUMP,
NETAPPDUMP or CELERRADUMP format.
v The table of contents is sent to a storage pool with either NATIVE or
NONBLOCK format.
Related tasks:
“Creating client option sets on the server” on page 465
Related reference:
Chapter 9, “Using NDMP for operations with NAS file servers,” on page 203

Configuring policy for LAN-free data movement

For LAN-free data movement, you can set up a SAN configuration in which a
client directly accesses a storage device to read or write data. LAN-free data
movement requires setup on the server and on the client, and the installation of a
storage agent on the client computer.

The storage agent transfers data between the client and the storage device. See
Storage Agent User's Guide for details. See the Web site for details on clients that
support the feature: [Link]
Software/Tivoli/Tivoli_Storage_Manager.

One task in configuring your systems to use this feature is to set up policy for the
clients. Copy groups for these clients must point to the storage pool that is
associated with the SAN devices. If you have defined a path from the client to a
drive on the SAN, drives in this storage pool can then use the SAN to send data
directly to the device for backup, archive, restore, and retrieve.

To set up the required policy, either define a new, separate policy domain, or
define a new management class in an existing policy domain.
Related tasks:

Chapter 13. Implementing policies for client data 525

 “Define a new policy domain”
“Configuring Tivoli Storage Manager for LAN-free data movement” on page 122
Related reference:
“Define a new management class in an existing policy domain” on page 527

Define a new policy domain

One way to configure policy for clients is to define a separate policy domain in
which the active policy set has a default management class with the required
settings. Then register all clients using SAN data transfer to that domain.

Procedure

Complete the following steps to define a new policy domain:

1. Create the policy domain for the clients. For example, to define a policy
domain that is named SANCLIENTS, issue the following command:
define domain sanclients
description=’Policy domain for clients using SAN devices’
2. Create a policy set in that domain. For example, to define the policy set that is
named BASE in the SANCLIENTS policy domain, issue the following
command:
define policyset sanclients base
3. Create the default management class for the policy set. First define the
management class, then assign the management class as the default for the
policy set.
For example, to define the management class that is named SANCLIENTMC,
issue the following command:
define mgmtclass sanclients base sanclientmc

Then assign the new management class as the default:

assign defmgmtclass sanclients base sanclientmc
4. Define the backup copy group in the default management class, as follows:
v Specify the DESTINATION, the name of the storage pool that is associated
with the SAN devices on the server.
The storage pool must already be set up. The storage pool must point to a
device class that is associated with the library for the SAN devices.
v Accept the default settings for all remaining parameters.
For example, to define the backup copy group for the SANCLIENTMC
management class, issue the following command:
define copygroup sanclients base sanclientmc standard destination=sanpool
5. Define the archive copy group in the default management class, as follows:
v Specify the DESTINATION, the name of the storage pool that is associated
with the SAN devices on the server.
The storage pool must already be set up. The storage pool must point to a
device class that is associated with the library for the SAN devices.
v Accept the default settings for all remaining parameters.
For example, to define the archive copy group for the SANCLIENTMC
management class, issue the following command:
define copygroup sanclients base sanclientmc standard
type=archive destination=sanpool
6. Activate the policy set.

526 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 For example, to activate the BASE policy set in the SANCLIENTS policy
domain, issue the following command:
activate policyset sanclients base
7. Register or update the application clients to associate them with the new policy
domain.
For example, to update the node SANCLIENT1, issue the following command:
update node sanclient1 domain=sanclients
Related tasks:
“Configuring Tivoli Storage Manager for LAN-free data movement” on page 122

Define a new management class in an existing policy domain

If you choose not to define a separate policy domain with the appropriate
management class as the default, you must define a new management class within
an existing policy domain and activate the policy set.

Because the new management class is not the default for the policy domain, you
must add an include statement to each client options file to bind objects to that
management class.

For example, suppose sanclientmc is the name of the management class that you
defined for clients that are using devices on a SAN. You want the client to be able
to use the SAN for backing up any file on the c drive. Put the following line at the
end of the client's include-exclude list:
include c:* sanclientmc

For details on the include-exclude list, see Backup-Archive Clients Installation and
User's Guide.

Policy for Tivoli Storage Manager servers as clients

One server (a source server) can be registered as a client to another server (the
target server). Data stored by the source server appears as archived files on the
target server. The source server is registered to a policy domain on the target
server, and uses the default management class for that policy domain.

In the default management class, the destination for the archive copy group
determines where the target server stores data for the source server. Other policy
specifications, such as how long to retain the data, do not apply to data stored for
a source server.
Related tasks:
“Using virtual volumes to store data on another server” on page 732

Setting policy to enable point-in-time restore for clients

To enable clients to restore backed-up files to a specific point in time, you must set
up the backup copy group differently from the STANDARD. The Versions Data
Exists, Versions Data Deleted, and Retain Extra Versions parameters work
together to determine over what time period a client can perform a point-in-time
restore operation.

For example, you decide to ensure that clients can choose to restore files from
anytime in the previous 60 days. In the backup copy group, set the Retain Extra
Versions parameter to 60 days. More than one backup version of a file may be
stored per day if clients perform selective backups or if clients perform incremental
backups more than once a day. The Versions Data Exists parameter and the

Chapter 13. Implementing policies for client data 527

 Versions Data Deleted parameter control how many of these versions are kept by
the server. To ensure that any number of backup versions are kept for the required
60 days, set both the Versions Data Exists parameter and the Versions Data
Deleted parameter to NOLIMIT. This means that the server essentially determines
the backup versions to keep based on how old the versions are, instead of how
many backup versions of the same file exist.

Keeping backed-up versions of files long enough to allow clients to restore their
data to a point in time can mean increased resource costs. Requirements for server
storage increase because more file versions are kept, and the size of the server
database increases to track all of the file versions. Because of these increased costs,
you may want to choose carefully which clients can use the policy that allows for
point-in-time restore operations.

Clients need to run full incremental backup operations frequently enough so that
IBM Tivoli Storage Manager can detect files that have been deleted on the client
file system. Only a full incremental backup can detect whether files have been
deleted since the last backup. If full incremental backup is not done often enough,
clients who restore to a specific time may find that many files that had actually
been deleted from the workstation get restored. As a result, a client’s file system
may run out of space during a restore process.

Important: The server will not attempt to retrieve client files from an active-data
pool during a point-in-time restore. Point-in-time restores require both active and
inactive file versions. Active-data pools contain only active file versions. For
optimal efficiency during point-in-time restores and to avoid switching between
active-data pools and primary or copy storage pools, the server retrieves both
active and inactive versions from the same storage pool and volumes.

Distributing policy using enterprise configuration

If you set up one Tivoli Storage Manager server as a configuration manager, you
can distribute policy to other Tivoli Storage Manager servers.

To distribute policy, you associate a policy domain with a profile. Managed servers
that subscribe to the profile then receive the following definitions:
v The policy domain itself
v Policy sets in that domain, except for the ACTIVE policy set
v Management classes in the policy sets
v Backup and archive copy groups in the management classes
v Client schedules associated with the policy domain
The names of client nodes and client-schedule associations are not distributed. The
ACTIVE policy set is also not distributed.

The distributed policy becomes managed objects (policy domain, policy sets,
management classes, and so on) defined in the database of each managed server.
To use the managed policy, you must activate a policy set on each managed server.
If storage pools specified as destinations in the policy do not exist on the managed
server, you receive messages pointing out the problem when you activate the
policy set. You can create new storage pools to match the names in the policy set,
or you can rename existing storage pools.

On the managed server you also must associate client nodes with the managed
policy domain and associate client nodes with schedules.

528 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 Related tasks:
“Setting up enterprise configurations” on page 701

Querying policy
You can request information about the contents of policy objects. You might want
to do this before creating new objects or when helping users to choose policies that
fit their needs.

Task Required Privilege Class

Query any policy domain, policy set, Any administrator
management class, or copy group

You can specify the output of a query in either standard or detailed format. The
examples in this section are in standard format.

On a managed server, you can see whether the definitions are managed objects.
Request the detailed format in the query and check the contents of the “Last
update by (administrator)” field. For managed objects, this field contains the string
$$CONFIG_MANAGER$$.

Querying copy groups

You can request information about backup copy groups through a command.

Issue the following command to request information about the backup copy group
(the default) in the ENGPOLDOM engineering policy domain:
query copygroup engpoldom * *

The following data shows the output from the query. It shows that the ACTIVE
policy set contains two backup copy groups that belong to the MCENG and
STANDARD management classes.

Policy Policy Mgmt Copy Versions Versions Retain Retain

Domain Set Name Class Group Data Data Extra Only
Name Name Name Exists Deleted Versions Version
--------- --------- --------- -------- -------- -------- -------- -------
ENGPOLDOM ACTIVE MCENG STANDARD 5 4 90 600
ENGPOLDOM ACTIVE STANDARD STANDARD 2 1 30 60
ENGPOLDOM STANDARD MCENG STANDARD 5 4 90 600
ENGPOLDOM STANDARD STANDARD STANDARD 2 1 30 60
ENGPOLDOM TEST STANDARD STANDARD 2 1 30 60

To request information about archive copy groups in the ENGPOLDOM

engineering policy domain, enter:
query copygroup engpoldom * type=archive

The following data shows the output from the query.

Chapter 13. Implementing policies for client data 529

 Policy Policy Mgmt Copy Retain
Domain Set Name Class Group Version
Name Name Name
--------- --------- --------- --------- --------
ENGPOLDOM ACTIVE MCENG STANDARD 730
ENGPOLDOM ACTIVE STANDARD STANDARD 365
ENGPOLDOM STANDARD MCENG STANDARD 730
ENGPOLDOM STANDARD STANDARD STANDARD 365
ENGPOLDOM TEST STANDARD STANDARD 365

Querying management classes

You can request information about management classes through a command.

Issue the following command to request information about management classes in

the ENGPOLDOM engineering policy domain:
query mgmtclass engpoldom * *

The following figure is the output from the query. It shows that the ACTIVE policy
set contains the MCENG and STANDARD management classes.

Policy Policy Mgmt Default Description

Domain Set Name Class Mgmt
Name Name Class ?
--------- --------- --------- --------- ------------------------
ENGPOLDOM ACTIVE MCENG No Engineering Management
Class with Backup and
Archive Copy Groups
ENGPOLDOM ACTIVE STANDARD Yes Installed default
management class
ENGPOLDOM STANDARD MCENG No Engineering Management
Class with Backup and
Archive Copy Groups
ENGPOLDOM STANDARD STANDARD Yes Installed default
management class
ENGPOLDOM TEST STANDARD Yes Installed default
management class

Querying policy sets

You can request information about policy sets through a command.

Issue the following command to request information about policy sets in the
ENGPOLDOM engineering policy domain:
query policyset engpoldom *

The following figure is the output from the query. It shows an ACTIVE policy set
and two inactive policy sets, STANDARD and TEST.

Policy Policy Default Description

Domain Set Name Mgmt
Name Class
Name
--------- --------- --------- ------------------------
ENGPOLDOM ACTIVE STANDARD Installed default policy
set
ENGPOLDOM STANDARD STANDARD Installed default policy
set
ENGPOLDOM TEST STANDARD Policy set for testing

530 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 Querying policy domains
You can request information about policy domains through a command.

Issue the following command to request information about a policy domain (for
example, to determine if any client nodes are registered to that policy domain):
query domain *

The following figure is the output from the query. It shows that both the
ENGPOLDOM and STANDARD policy domains have client nodes assigned to
them.

Policy Activated Activated Number of Description

Domain Policy Default Registered
Name Set Mgmt Nodes
Class
--------- --------- --------- ---------- ------------------------
APPCLIEN- BASE APPCLIEN- 1 Policy domain for
TS TMC application clients
ENGPOLDOM STANDARD STANDARD 21 Engineering Policy
Domain
STANDARD STANDARD STANDARD 18 Installed default policy
domain.

Deleting policy
When you delete a policy object, you also delete any objects belonging to it. For
example, when you delete a management class, you also delete the copy groups in
it.

You cannot delete the ACTIVE policy set or objects that are part of that policy set.

Task Required Privilege Class

Delete policy domains System
Delete any policy sets, management classes, System or unrestricted policy
or copy groups
Delete policy sets, management classes, or Restricted policy
copy groups that belong to policy domains
over which you have authority

You can delete the policy objects named STANDARD that come with the server.
However, all STANDARD policy objects are restored whenever you reinstall the
server.
Related concepts:
“Protection and expiration of archive data” on page 514

Chapter 13. Implementing policies for client data 531

 Deleting copy groups
You can delete a backup or archive copy group if it does not belong to a
management class in the ACTIVE policy set.

For example, to delete the backup and archive copy groups belonging to the
MCENG and STANDARD management classes in the STANDARD policy set,
enter:
delete copygroup engpoldom standard mceng type=backup
delete copygroup engpoldom standard standard type=backup
delete copygroup engpoldom standard mceng type=archive
delete copygroup engpoldom standard standard type=archive

Deleting management classes

You can delete a management class if it does not belong to the ACTIVE policy set.

For example, to delete the MCENG and STANDARD management classes from the
STANDARD policy set, enter:
delete mgmtclass engpoldom standard mceng
delete mgmtclass engpoldom standard standard

When you delete a management class from a policy set, the server deletes the
management class and all copy groups that belong to the management class in the
specified policy domain.

Deleting policy sets

Authorized administrators can delete any policy set other than the ACTIVE policy
set.

For example, to delete the TEST policy set from the ENGPOLDOM policy domain,
enter:
delete policyset engpoldom test

When you delete a policy set, the server deletes all management classes and copy
groups that belong to the policy set within the specified policy domain.

The ACTIVE policy set in a policy domain cannot be deleted. You can replace the
contents of the ACTIVE policy set by activating a different policy set. Otherwise,
the only way to remove the ACTIVE policy set is to delete the policy domain that
contains the policy set.

Deleting policy domains

You can delete a policy domain only if the domain has no client nodes registered
to it. To determine if any client nodes are registered to a policy domain, issue the
QUERY DOMAIN or the QUERY NODE command.

Before you begin

Move any client nodes to another policy domain, or delete the nodes.

Procedure

As an example, to delete the STANDARD policy domain, complete the following

steps:

532 IBM Tivoli Storage Manager for Windows: Administrator's Guide

1. Request a list of all client nodes assigned to the STANDARD policy domain by
issuing the following command:
query node * domain=standard
2. If client nodes are assigned to the policy domain, remove them in one of the
following ways:
a. Assign each client node to a new policy domain. For example, enter the
following commands:
update node htang domain=engpoldom
update node tomc domain=engpoldom
update node pease domain=engpoldom
If the ACTIVE policy set in ENGPOLDOM does not have the same
management class names as in the ACTIVE policy set of the STANDARD
policy domain, then backup versions of files may be bound to a different
management class name.
b. Delete each node from the STANDARD policy domain by first deleting all
file spaces belonging to the nodes, then deleting the nodes.
3. Delete the policy domain by issuing the following command:
delete domain standard

When you delete a policy domain, the server deletes the policy domain and all
policy sets (including the ACTIVE policy set), management classes, and copy
groups that belong to the policy domain.
Related reference:
“How files and directories are associated with a management class” on page 489

Chapter 13. Implementing policies for client data 533

534 IBM Tivoli Storage Manager for Windows: Administrator's Guide
Chapter 14. Managing data for client nodes
You might need help to generate backup sets and enable subfile backups for client
nodes. Subfile backups are only applicable to Windows clients. You can also use
data validation for client nodes so that any data corruption is identified when data
is sent over the network between the client and server.

Tasks:
“Validating a node's data during a client session” on page 536
“Securing communications” on page 831
“Encrypting data on tape” on page 536
“Setting up shredding” on page 540
“Generating client backup sets on the server” on page 544
“Restoring backup sets from a backup-archive client” on page 548
“Moving backup sets to other servers” on page 548
“Managing client backup sets” on page 549
“Enabling clients to use subfile backup” on page 552
“Managing storage usage for archives” on page 555
Concepts:
“Performance considerations for data validation” on page 536
“Securing sensitive client data” on page 539
“Creating and using client backup sets” on page 543

Validating a node's data

Data validation can identify data corruption during a client session when data is
sent between a client and the server. IBM Tivoli Storage Manager provides the
option of specifying whether a cyclic redundancy check (CRC) is performed during
a client session to validate data sent over the network between a client or a storage
agent and the server.

Cyclic redundancy checking is performed at the client when the client requests
services from the server. For example, the client issues a query, backup, or archive
request. The server also performs a CRC operation on the data sent by the client
and compares its value with the value calculated by the client. If the CRC values
do not match, the server will issue an error message once per session. Depending
on the operation, the client may attempt to automatically retry the operation.

After Tivoli Storage Manager completes the data validation, the client and server
discard the CRC values generated in the current session.

Data validation can be enabled for one or all of the following items:
v Tivoli Storage Manager client nodes.
v Tivoli Storage Manager storage agents. For details, refer to the Storage Agent
User's Guide for your particular operating system.

© Copyright IBM Corp. 1993, 2013 535

 See “Choosing when to enable data validation” on page 891 to help you determine
where to enable data validation.

Performance considerations for data validation

Consider the impact on performance when you decide whether data validation is
necessary for all nodes or some nodes. Data validation impacts performance
because additional CPU overhead is required on both the client and server to
calculate and compare CRC values.

This type of validation is independent from validating data written to a storage

pool volume. See “Data validation during audit volume processing” on page 890.

Validating a node's data during a client session

You can enable data validation for a node by using either the REGISTER NODE or
UPDATE NODE command. By default, data validation is set to NO.

Methods for enabling data validation for a node include choosing data validation
for individual nodes, specifying a set of nodes by using a wildcard search string,
or specifying a group of nodes in a policy domain.

For example, to enable data validation for existing node, ED, you can issue an
UPDATE NODE command. This user backs up the company payroll records weekly
and you have decided it is necessary to have all the user data validated: the data
itself and metadata.
update node ed validateprotocol=all

Later, the network has shown to be stable and no data corruption has been
identified when user ED has processed backups. You can then disable data
validation to minimize the performance impact of validating all of ED's data
during a client session. For example:
update node ed validateprotocol=no

Encrypting data on tape

It is often critical to secure client data, especially when that data might be of a
sensitive nature. To ensure that data for off-site volumes is protected, IBM tape
encryption technology is available.

This technology uses a stronger level of encryption by requiring 256-bit Advanced

Encryption Standard (AES) encryption keys. Keys are passed to the drive by a key
manager to encrypt and decrypt data.

IBM tape technology supports different methods of drive encryption for the
following devices:
v IBM 3592 generation 2 and generation 3
v IBM linear tape open (LTO) generation 4 and generation 5
Application encryption
Encryption keys are managed by the application, in this case, Tivoli
Storage Manager. Tivoli Storage Manager generates and stores the keys in
the server database. Data is encrypted during WRITE operations, when the
encryption key is passed from the server to the drive. Data is decrypted for
READ operations.

536 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 Attention: When using application encryption, you must take extra care
to secure database backups because the encryption keys that are used to
encrypt and decrypt data are stored in the server database. To restore your
data, you must have the correct database backup and corresponding
encryption keys to access your information. Ensure that you back up the
database frequently and safeguard the backups to prevent data loss or
theft. Anyone who has access to both the database backup and the
encryption keys has access to your data.
Use application-managed encryption for only storage pool volumes. Other
volumes such as backup-set tapes, export volumes, and database backups
are not encrypted using the application method.
Library encryption
Encryption keys are managed by the library. Keys are stored in an
encryption key manager and provided to the drive. If you set up the
hardware to use the library encryption, you can use this method by setting
the DRIVEENCRYPTION parameter in the device class definition to ALLOW.

Restriction: Only certain IBM libraries support IBM LTO-4 library

encryption.
System encryption
System encryption is available on AIX. Encryption keys that are provided
to the drive are managed by the device driver or operating system and
stored in an encryption key manager. If the hardware is set up to use
system encryption, you can use this method by setting the
DRIVEENCRYPTION parameter in the device class definition to ALLOW.

The methods of drive encryption that you can use with Tivoli Storage Manager are
set up at the hardware level. Tivoli Storage Manager cannot control or change
which encryption method is used in the hardware configuration. If the hardware is
set up for the application encryption method, Tivoli Storage Manager can turn
encryption on or off depending on the DRIVEENCRYPTION value on the device
class. For more information about specifying this parameter, see the following
topics:
v “Encrypting data with drives that are 3592 generation 2 and later” on page 185
v “Encrypting data using LTO generation 4 tape drives” on page 193
v “Enabling ECARTRIDGE drive encryption” on page 196 and “Disabling
ECARTRIDGE drive encryption” on page 197

Choosing an encryption method

Deciding on which encryption method you want to use depends on how you want
to manage your data. If you only want to encrypt storage pool volumes and
eliminate some encryption processing on your system, the Application method
should be enabled.

This method allows Tivoli Storage Manager to manage the encryption keys. When
using Application encryption, you must take extra care to secure database backups
since the encryption keys are stored in the server database. Without access to
database backups and matching encryption keys, you will not be able to restore
your data.

If you want to encrypt all of your data in a particular logical library or encrypt
data on more than just storage pool volumes, the System or Library method can be

Chapter 14. Managing data for client nodes 537

 used. These methods are virtually transparent to the server. Tivoli Storage Manager
is aware of them being used and displays informational messages when writing to
an empty volume.

Library managed encryption allows you to control which volumes are encrypted
through the use of their serial numbers. You can specify a range or set of volumes
to encrypt. With Application managed encryption, you can create dedicated storage
pools that only contain encrypted volumes. This way, you can use storage pool
hierarchies and policies to manage the way data is encrypted.

The Library and System methods of encryption can share the same encryption key
manager, which allows the two modes to be interchanged. However, this can only
occur if the encryption key manager is set up to share keys. Tivoli Storage
Manager cannot currently verify if encryption key managers for both methods are
the same. Neither can Tivoli Storage Manager share or use encryption keys
between the application method and either library or system methods of
encryption.

To determine whether or not a volume is encrypted and which method was used,
you can issue the QUERY VOLUME command with FORMAT=DETAILED. For more
information on data encryption using the backup-archive client, see the
Backup-Archive Clients Installation and User's Guide.

Changing your encryption method and hardware configuration

If you want to change the encryption method for a given set of volumes, the
volumes need to be returned to scratch status. Updating the parameter value will
only affect empty volumes.

For example, if you currently have Application managed encryption enabled, and
you decide that you don't want encryption enabled at all, only empty volumes will
be impacted by the change. Filling volumes will continue to be encrypted while
new volumes will not. If you do not want currently filling volumes to continue
being encrypted, the volume status should be changed to READONLY. This will
ensure that Tivoli Storage Manager does not append any more encrypted data to
the volumes. You can use the MOVE DATA command to transfer the data to a new
volume after the update of the DRIVEENCRYPTION parameter. The data will then
be available in an un-encrypted format.

When migrating from one hardware configuration to another, you will need to
move your data from the old volumes to new volumes with new encryption keys
and key managers. You can do this by setting up two logical libraries and storage
pools (each with a different encryption method) and migrating the data from the
old volumes to the new volumes. This will eliminate volumes that were encrypted
using the original [Link] that you have volumes that were encrypted
using the Library method and you want to migrate to the Application method.
Tivoli Storage Manager will be unable to determine which encryption keys are
needed for data on these volumes because the library's encryption key manager
stores these keys and Tivoli Storage Manager does not have access to them.
Table 46 on page 539 illustrates considerations for changing your hardware
encryption method.

538 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 Table 46. Hardware and encryption method compatibility
Volumes with Volumes with Volumes with
Application Library System
Volumes with Managed Managed Managed
No Encryption Encryption Encryption Encryption
Desired No Special Incompatible. Incompatible. Incompatible.
Hardware Consideration Scratch tape Scratch tape Scratch tape
Method = None labels will be labels will be labels will be
unreadable and unreadable and unreadable and
need to be need to be need to be
relabeled relabeled relabeled
Desired No Special No Special Incompatible Incompatible
Hardware Consideration Consideration
Method =
Application
Desired No Special Incompatible No Special Ensure the same
Hardware Consideration Consideration key bank/server
Method = is still used
Library
Desired No Special Incompatible Ensure the same No Special
Hardware Consideration key bank/server Consideration
Method = is still used
System

Restriction: If encryption is enabled for a device class, and the device class is
associated with a storage pool, the storage pool should not share a scratch pool
with other device classes that cannot be encrypted. If a tape is encrypted, and you
plan to use it on a drive that cannot be encrypted, you must manually relabel the
tape before it can be used on that drive.

Securing sensitive client data

After client data has been deleted, it might still be possible to recover it. For
sensitive data, this condition is a potential security exposure. The destruction of
deleted data, also known as shredding, lets you store sensitive data so that it is
overwritten one or more times after it is deleted.

This process increases the difficulty of discovering and reconstructing the data
later. Tivoli Storage Manager performs shredding only on data in random-access
disk storage pools. You can configure the server to ensure that sensitive data is
stored only in storage pools in which shredding is enforced (shred pools).

Shredding occurs only after a data deletion commits, but it is not necessarily
completed immediately after the deletion. The space occupied by the data to be
shredded remains occupied while the shredding takes place, and is not available as
free space for new data until the shredding is complete. When sensitive data is
written to server storage and the write operation fails, the data that was already
written is shredded.

Shredding performance is affected by the amount of data to be shredded, the

number of times that data is to be overwritten, and the speed of the disk and
server hardware. You can specify that the data is to be overwritten up to ten times.
The greater the number of times, the greater the security but also the greater the
impact on server performance. It is strongly recommended that write caching be

Chapter 14. Managing data for client nodes 539

 disabled for any disk devices used to store sensitive data. If write caching is
enabled, the overwrite operations are adversely affected.

Shredding can be done either automatically after the data is deleted or manually
by command. The advantage of automatic shredding is that it is performed
without administrator intervention whenever deletion of data occurs. This limits
the time that sensitive data might be compromised. Automatic shredding also
limits the time that the space used by deleted data is occupied. The advantage of
manual shredding is that it can be performed when it will not interfere with other
server operations.

Setting up shredding
You must configure Tivoli Storage Manager so that data identified as sensitive is
stored only in storage pools that will enforce shredding after that data is deleted.

Procedure
Complete the following steps to set up your shredding configuration:
1. Specify that you want data to be shredded either automatically after it is
deleted or manually by an administrator. You can specify how shredding is to
be done by setting the SHREDDING server option.
shredding automatic

You can also set the shredding option dynamically by using the SETOPT
command.
2. Set up one or more random access disk storage pool hierarchies that will
enforce shredding and specify how many times the data is to be overwritten
after deletion. For example,
define stgpool shred2 disk shred=5
define stgpool shred1 disk nextstgpool=shred2 shred=5
3. Define volumes to those pools, and specify disks for which write caching can
be disabled.
define volume shred1 j:\storage\[Link] formatsize=100
define volume shred2 m:\storage\[Link] formatsize=100
4. Define and activate a policy for the sensitive data. The policy will bind the data
to a management class whose copy groups specify shred storage pools.
define domain shreddom
define policyset shreddom shredpol
define mgmtclass shreddom shredpol shredclass
define copygroup shreddom shredpol shredclass type=backup
destination=shred1
define copygroup shreddom shredpol shredclass type=archive
destination=shred1
activate policyset shreddom shredpol
5. Identify those client nodes whose data should be shredded after deletion, and
assign them to the new domain.
update node engineering12 domain=shreddom

Results

If you have specified manual shredding with the SHREDDING server option, you can
start the shredding process by issuing the SHRED DATA command. This command
lets you specify how long the process will run before it is canceled and how the
process responds to an I/O error during shredding. For objects that cannot be
shredded, the server reports each object.

540 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 Note: If you specify manual shredding, run the SHRED DATA command regularly, at
least as often as you perform other routine server-maintenance tasks (for example,
expiration, reclamation, and so on). Doing so can prevent performance degradation
of certain server processes (in particular, migration). For best results, run SHRED
DATA after any operation (for example, expiration and migration) that deletes files
from a shred pool.

To see the status and amount of data waiting to be shredded, you can issue the
QUERY SHREDSTATUS command. The server reports a summary of the number and
size of objects waiting to be shredded. To display detailed information about data
shredding on the server, issuing the following command:
query shredstatus format=detailed

Figure 72 displays a detailed report for the storage pool.

Shredding Objects Occupied Data left

Active Awaiting Space to shred
Shred (MB) (MB)
--------- -------- -------- ----------
NO 4 182 364

Figure 72. Querying shredding status

When data shredding completes, a message is issued that reports the amount of
data that was successfully shredded and the amount of data that was skipped, if
any.

Ensuring that shredding is enforced

It is important to ensure that sensitive data is stored only in shred storage pools.
Only data that is in a shred pool is shredded after being deleted.

Some changes to objects and some server operations involving the moving or
copying of data could result in sensitive data that cannot be shredded. This would
compromise the intent and value of shredding.

Table 47 describes these types of operations.

Table 47. Commands affecting the shredding process
Command Operation
BACKUP STGPOOL To back up a shred pool to a copy storage pool you must set
the SHREDTONOSHRED parameter to YES. If this value is not
specified, the server issues an error message and does not
allow the backup. If this value is specified, the server does not
issue a warning message when the shred pool is backed up,
and that data cannot be shredded.
COPY ACTIVEDATA To copy data from a shred pool to an active-data pool you
must set the SHREDTONOSHRED parameter to YES. If this value is
not specified, the server issues an error message and does not
allow the data to be copied. If this value is specified, the
server does not issue a warning when data from the shred
pool is copied, and that data cannot be shredded.

Chapter 14. Managing data for client nodes 541

 Table 47. Commands affecting the shredding process (continued)
Command Operation
DEFINE STGPOOL Tivoli Storage Manager does not require that the next storage
pool for a shred pool also be a shred pool. When you define a
storage pool and you specify a non-shred pool as the next
storage pool, a warning message is issued, but you can choose
to continue with the definition. After the storage pool is
defined and migration occurs, no message is issued. However,
the data in the next storage pool cannot be shredded.
EXPIRE INVENTORY Data in a shred pool is deleted after inventory expiration and
DELETE FILESPACE after deletion of a file space or volume. After the data is
DELETE VOLUME deleted, it is shredded. However, if data is deleted that has
copies in both shred and non-shred pools, the server shreds
only those copies in the shred pools and does not issue a
warning message before deletion. The data non-shred pools
cannot be shredded.
EXPORT NODE To export data from a shred pool you must set the
EXPORT SERVER ALLOWSHREDDABLE parameter to YES. If this value is specified,
and the exported data includes data from shred pools, that
data cannot be shredded. If the export operation includes data
from shred pools, the server does not issue a warning
message.
GENERATE BACKUPSET To include data from a shred pool when you generate a
backup set, you must set the ALLOWSHREDDABLE parameter to
YES. If this value is specified and the backup set includes data
from shred pools, that data cannot be shredded, and no
warning message is issued.
MOVE DATA You cannot move data from a shred pool to a destination that
is not another shred pool unless you set the SHREDTONOSHRED
parameter to YES. When the move is complete, the original
data is shredded but the data in the destination storage pool
cannot be shredded.
UPDATE STGPOOL The server issues a warning message for any of the storage
pool updates listed below. You can choose to continue with the
operation, but the result is that some or all sensitive data
cannot be shredded.
v For a shred pool, add or change a non-shred pool as the
next pool in the hierarchy. This includes copy storage pools
and active data storage pools.
v Change a pool from non-shred to shred when the
NEXTSTGPOOL attribute specifies a non-shred pool.
v Change a pool with COPYSTGPOOLS or
ACTIVEDATAPOOLS attributes from non-shred to shred.
v Change a pool from shred to non-shred by setting the
SHRED parameter to 0.

542 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Creating and using client backup sets
A backup set is a collection of backed-up data from one client, which is stored and
managed as a single object on specific media in server storage. The server creates
copies of active versions of a client's backed up objects that are within the one or
more file spaces specified with the GENERATE BACKUPSET command, and consolidates
them onto sequential media.

Currently, the backup object types that are supported for backup sets include
directories, files, and image data. If you are upgrading from Tivoli Storage
Manager Express®, backup sets can also contain data from Data Protection for
Microsoft SQL and Data Protection for Microsoft Exchange servers. The backup set
process is also called instant archive.

You can generate backup sets on the server for individual client nodes or for
groups of nodes. A node group is a group of client nodes that are acted upon as a
single entity. If you specify one or more node groups, the server generates a
backup set for each node and places all of the backup sets together on a single set
of output volumes. To create a node group, use the DEFINE NODEGROUP
command, and then use the DEFINE NODEGROUPMEMBER command to add nodes to
the group. For details, see the Administrator's Reference. The client node for which a
backup set is generated must be registered to the server.

Centera storage devices do not support backup sets.

The media might be directly readable by something such as the following device:
v A CD-ROM, JAZ, or ZIP drive attached to a client's computer.

Administrators can generate multiple copies of backup sets that correspond to

some point-in-time. The backup sets can be retained for various time periods. This
is an efficient way to create long-term storage of periodic backups, without
requiring the data to be sent over the network again.

While an administrator can generate a backup set from any client's backed up files,
backup sets can only be used by a backup-archive client.

You cannot generate a backup set with files that were backed up to Tivoli Storage
Manager by using NDMP. However, you can create a backup set with files that
were backed up using NetApp SnapShot Difference.

When generating backup sets, the server searches for active file versions in an
active-data storage pool associated with a FILE device class, if such a pool exists.
For details about the complete storage-pool search-and-selection order, see
“Selection for active file versions” on page 243.

Data from a shred storage pool will not be included in a backup set unless you
explicitly permit it by setting the ALLOWSHREDDABLE parameter to YES in the
GENERATE BACKUPSET command. If this value is specified, and the client node data
includes data from shred pools, that data cannot be shredded. The server does not
issue a warning if the backup set operation includes data from shred pools. For
more information about shredding, see “Securing sensitive client data” on page
539.

For details about creating and using backup sets, see the following sections:
v “Generating client backup sets on the server” on page 544

Chapter 14. Managing data for client nodes 543

 v “Restoring backup sets from a backup-archive client” on page 548
v “Moving backup sets to other servers” on page 548
v “Managing client backup sets” on page 549

Generating client backup sets on the server

You can generate backup sets on the server for client nodes. The client node for
which a backup set is generated must be registered to the server. An incremental
backup must be completed for a client node before the server can generate a
backup set for the client node.

About this task

The GENERATE BACKUPSET command runs as a background process on the server. If

you cancel the background process created by this command, the media may not
contain a complete backup set. You must have system or restricted policy over the
domain to which the node is assigned to generate a backup set.

Generate backup set processing attempts to process all available objects onto the
backup set media. However, objects may be skipped due to being unavailable on
the server or other errors (I/O, media, hardware) that can occur at the time of
backup set generation. Some errors may lead to termination of processing before
all available data can be processed. For example, if the source data for a backup set
is on multiple sequential volumes and the second or subsequent segment of an
object spanning volumes is on a volume that is unavailable, processing is
terminated.

If objects are skipped or other problems occur to terminate processing, review all
of the messages associated with the process to determine whether or not it should
be run again. To obtain a complete backup set, correct any problems that are
indicated and reissue the GENERATE BACKUPSET command.

Procedure

To improve performance when generating backup sets, you can do one or both of
the following tasks:
v Collocate the primary storage pool in which the client node data is stored. If a
primary storage pool is collocated, client node data is likely to be on fewer tape
volumes than it would be if the storage pool were not collocated. With
collocation, less time is spent searching database entries, and fewer mount
operations are required.
v Store active backup data in an active-data pool associated with a FILE device
class. When generating a backup set, the server will search this type of
active-data pool for active file versions before searching other possible sources.

Results

See the following sections:

v “Choosing media for generating the backup set” on page 545
v “Selecting a name for the backup set” on page 545
v “Setting a retention period for the backup set” on page 546
v “Example: generating a client backup set” on page 546
v “Generating backup sets to a specific point-in-time” on page 546
v “Generating backup sets with multiple data types” on page 547

544 IBM Tivoli Storage Manager for Windows: Administrator's Guide

v “Creating a single set of backup volumes with data from multiple nodes” on
page 547

Choosing media for generating the backup set

To generate a backup set, you must specify a device class that is associated with
the media to which the backup set will be written.

About this task

You can write backup sets to sequential media: sequential tape and device class
FILE. The tape volumes containing the backup set are not associated with storage
pools and, therefore, are not migrated through the storage pool hierarchy.

For device class FILE, the server creates each backup set with a file extension of
OST. You can copy FILE device class volumes to removable media that is
associated with CD-ROM, JAZ, or ZIP devices, by using the REMOVABLEFILE
device type.

You can determine whether to use scratch volumes when you generate a backup
set. If you do not use specific volumes, the server uses scratch volumes for the
backup set.

You can use specific volumes for the backup set. If there is not enough space to
store the backup set on the volumes, the server uses scratch volumes to store the
remainder of the backup set.

Procedure

Consider the following items when you select a device class for writing the backup
set:
v Generate the backup set on any sequential access devices whose device types are
supported on both the client and server. If you do not have access to compatible
devices, you will need to define a device class for a device type that is
supported on both the client and server.
v Ensure that the media type and recording format used for generating the backup
set is supported by the device that will be reading the backup set.
v You must restore, with the IBM Tivoli Storage Manager server, backup sets that
are written to more than one volume and generated to a REMOVABLEFILE
device. Issue the RESTORE BACKUPSET command and specify -location=server to
indicate that the backup set is on the Tivoli Storage Manager server.

Results

For more information, see “Configuring removable media devices” on page 89.

Selecting a name for the backup set

The server adds a unique suffix to the name you specify for the backup set. For
example, if you name the backup set mybackupset, the server adds a unique
extension, such as 3099, to the name. This allows you to create backup sets with
the same name without overwriting previous backup sets.

To later display information about this backup set, you can include a wildcard
character with the name, such as mybackupset*, or you can specify the fully
qualified name, such as mybackupset.3099.

Chapter 14. Managing data for client nodes 545

 Setting a retention period for the backup set
You can set the retention period, specified as a number of days, to retain the
backup set on the server. You can specify a number between zero and 30,000 days.

Backup sets are retained on the server for 365 days if you do not specify a value.
The server uses the retention period to determine when to expire the volumes on
which the backup set resides.

Example: generating a client backup set

Generate a backup set on CD-ROM that the client can later use to restore the data.
Because CD-ROM is a read-only medium, you must create a backup set on a
device that the server can write to before you create the backup set on the
CD-ROM. Do not exceed the CD-ROM data capacity.

Procedure

Complete the following steps to generate a backup set on a CD-ROM:

1. Define a device class whose device type is FILE. Name the device class
CDFILE:
define devclass cdfile devtype=file maxcapacity=640M dir=/backupset
2. Create a backup set with a device type of FILE for client node JOHNSON.
Name the backup set PROJECT and retain it for 90 days.
generate backupset johnson project devclass=file
volumes=BK1,BK2,BK3 retention=90

Attention: Volumes=BK1,BK2,BK3 specifies that the backup set is put into

files named BK1, BK2 and BK3. This assumes that the backup set is large
enough to require three 650MB files.
3. Use any CD-ROM authoring software to put these files onto CD-ROMs. For
this example, the CD-ROM volume names are BK1, BK2, and BK3.
v The authoring software must label each CD-ROM.
v The label must exactly match the name of the file.
For an example of using the backup set on the CD-ROM, see “Moving backup
sets to other servers” on page 548.

Generating backup sets to a specific point-in-time

You can generate a backup set to a specific point-in-time by using the PITDATE and
PITTIME parameters on the GENERATE BACKUPSET command. When these dates are
specified, the new backup set will contain files that were active at the specified
date and time, even if the files are inactive at the time the GENERATE BACKUPSET
command is issued.

Backup sets are generated to a point-in-time by using one of two date and time
specifications: the date and time specified on the GENERATE BACKUPSET command, or
the date and time the that the GENERATE BACKUPSET command was issued.

Point-in-time backup set generation works best if a recent date and time are
specified. Files that have expired, or are marked as expire-immediately cannot be
included in the backup set.

546 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Generating backup sets with multiple data types
You can use the GENERATE BACKUPSET command to generate backup sets that
contain either file or image data. You can use the DATATYPE parameter to specify the
types of data to be included. A separate backup set is generated for each specified
data type, but all the backup sets are stored together on a single set of output
media.

You can use the DATATYPE parameter to limit the backup set to only one data type.
For example, you might do this if you don't want to store redundant data on the
backup set media. Alternatively, you can specify that both file and image backup
data be included from a machine in order to reduce the number of tapes that must
be included in your off-site tape rotation.

Image backup sets include the image and all files and directories changed or
deleted since the image was backed up so that all backup sets on the media
represent the same point in time. Tables of contents are automatically generated for
any backup sets that contain image or application data. If the GENERATE BACKUPSET
command cannot generate a table of contents for one of these backup sets, then it
will fail.

For file level backup sets, the table of contents generation is optional. By default,
the command attempts to create a table of contents for file level backup sets, but it
will not fail if a table of contents is not created. You can control the table of
contents option by specifying the TOC parameter.

Creating a single set of backup volumes with data from multiple

nodes
On the GENERATE BACKUPSET command, you can specify multiple nodes or node
groups, and you can use wildcards with node names.

A separate backup set is generated for each specified node, but all of the backup
sets will be stored together on the same set of output volumes. the backup set for
each node has its own entry in the database. The QUERY BACKUPSET command will
display information about all backup sets, whether they are on their own tape or
stacked together with other backup sets onto one tape.

On the DEFINE BACKUPSET command, you can also specify multiple nodes or node
groups, and you can use wildcards with node names. DEFINE BACKUPSET
determines what backup sets are on the set of tapes and defines any that match the
specified nodes. Specifying only a single wildcard character ('*') for the node name
has the effect of defining all the backup sets on the set of tapes. Conversely, you
can define only those backup sets belonging to a particular node by specifying just
the name of that node. Backup sets on tapes belonging to nodes that are not
specified on the command are not defined. They will still exist on the tape, but
cannot be accessed.

The QUERY, UPDATE, and DELETE BACKUPSET commands also allow the specification of
node group names in addition to node names. When you delete backup sets, the
volumes on which the backup sets are stored are not returned to scratch as long as
any backup set on the volumes remain active.

Chapter 14. Managing data for client nodes 547

 Restoring backup sets from a backup-archive client
Backup-archive client nodes can restore their backup sets either directly from the
server or by using a device attached to the client's computer that will read the
media in which the backup set is stored. This second method only works for
file-level backup sets.

Backup sets can only be used by a backup-archive client, and only if the files in the
backup set originated from a backup-archive client.

For more information about restoring backup sets, see the Backup-Archive Clients
Installation and User's Guide for your particular operating system.

Selecting individual files for restore

You can query a backup set table of contents to select individual files for restore.
Table of contents are generated when a new backup set is created. They contain
entries for each object stored in the backup set. Entries detail the position of the
object within the backup set.

In order to query the contents of a backup set and choose files to restore, tables of
contents need to be loaded into the server database. The backup-archive client can
specify more than one backup set table of contents to be loaded to the server at the
beginning of a restore session.

Restoring image data from backup sets

Backup sets containing image data can be used during a disaster recovery
situation, for example, when a hard drive crashes and needs to be replaced.
Individual file restore is not possible using an image backup, so backup sets
containing normal file system data should be maintained for most other restores. A
backup set may contain image data or file system data, including files and
directories, but not both.

Image backups and restores require a table of contents when generating a backup
set for image data. If the table of contents existed but was deleted for some reason
then the image backup set cannot be restored until the table of contents is
regenerated with the GENERATE BACKUPSETTOC command.

Moving backup sets to other servers

You can define (move) a backup set generated on one server to another Tivoli
Storage Manager server. Any client backup set that you generate on one server can
be defined to another server as long as the servers share ../common device type.

About this task

The level of the server defining the backup set must be equal to or greater than the
level of the server that generated the backup set.

Task Required Privilege Class

Define a backup set If the REQSYSAUTHOUTFILE server option is set
to YES, system privilege is required. If the
REQSYSAUTHOUTFILE server option is set to
NO, system or restricted policy over the
domain to which the node is assigned is
required.

548 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 Procedure

Using the example described in “Example: generating a client backup set” on page
546, you can make the backup set that was copied to the CD-ROM available to
another server by issuing the following command:
define backupset johnson project devclass=cdrom volumes=BK1,BK2,BK3
description="backup set copied to CD-ROM"

Results

If you have multiple servers connecting to different clients, the DEFINE BACKUPSET
command makes it possible for you to take a previously generated backup set and
make it available to other servers. The purpose is to allow the user flexibility in
moving backup sets to different servers, thus allowing the user the ability to
restore their data from a server other than the one on which the backup set was
created.

Important:
1. Devclass=cdrom specifies a device class of type REMOVABLEFILE that points
to your CD-ROM drive. CD-ROMs have a maximum capacity of 650MB.
2. Volumes=BK1,BK2,BK3 specifies the names of the volumes containing the
backup set. The volume label of these CD-ROMs must match the name of the
file on the volume exactly.

Managing client backup sets

You can update, query, and delete backup sets.

About this task

Task Required Privilege Class
Update the retention period assigned to a System or restricted policy over the domain
backup set to which the node is assigned
Display information about backup sets Any administrator
Display information about backup set System or restricted policy over the domain
contents to which the node is assigned
Delete backup set If the REQSYSAUTHOUTFILE server option is set
to YES, system privilege is required. If the
REQSYSAUTHOUTFILE server option is set to
NO, system or restricted policy over the
domain to which the node is assigned is
required.

Generating a table of contents for a backup set

Tables of contents for backup sets are created as part of backup set generation.
v By default, when generating a backup set containing files and directories, the
server will attempt to create a table of contents for a new backup set, but will
not fail the backup set generation process if it cannot do so. You can, however,
choose to require a table of contents by setting the TOC parameter to YES.
v When generating a backup set that contains API data or images, the server will
require that a table of contents is generated. You cannot override this default.
In either case, if a table of contents is required and the server cannot create it, the
backup set generation process will fail.

Chapter 14. Managing data for client nodes 549

 Tables of contents:
v Reside on the server even if the backup set's media has been moved off-site.
v Can be generated for existing backup sets that do not contain a table of contents.
v Can be re-generated when a backup set is defined on a new server, or if using a
user-generated copy on a different medium.

The GENERATE BACKUPSETTOC command allows a table of contents to be created for

backup sets that do not have one. It may be used after a backup set is added to the
server via the DEFINE BACKUPSET command or for backup sets that were generated
by an earlier release of the Tivoli Storage Manager server.

Backup set tables of contents are stored in the storage pool identified by the
TOCDESTINATION attribute of the backup copy group associated with the
management class to which the backup set is bound. The management class to
which the backup set is bound will either be the default management class in the
policy domain in which the backup set's node is registered, or the management
class specified by the TOCMGmtclass parameter of the GENERATE BACKUPSET,
GENERATE BACKUPSETTOC, or DEFINE BACKUPSET command. Tables of contents for
backup sets are retained until the backup set with which they are associated
expires or is deleted. They are not subject to the policy associated with their
management class. You can issue the QUERY BACKUPSET command to show whether
a given backup set has a table of contents or not. Output from the QUERY BACKUPSET
command can be filtered based on the existence of a table of contents. This allows
you to determine which backup sets may need to have a new table of contents
created, or conversely, which backup sets could be used with the client's file-level
restore.

Updating the retention period of a backup set

When you want to change the number of days the server retains a backup set,
update the retention period that is associated with the backup set.

Procedure

To update the retention period assigned to backup set named ENGDATA.3099,

belonging to client node JANE, to 120 days, issue the following example command:
update backupset jane engdata.3099 retention=120

Displaying backup set information

To view information about backup sets, you can use the QUERY BACKUPSET
command. The output that is displayed lists information such as the name of the
client node whose data is contained in the backup set as well as the description of
the backup set, assuming one has been used.

About this task

The following figure shows the report that is displayed after you enter:
query backupset f=d

550 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 Node Name: JANE
Backup Set Name: MYBACKUPSET.3099
Date/Time: 09/04/2002 [Link]
Retention Period: 60
Device Class Name: DCFILE
Description:
Filespace names: \\jane\c$ \\jane\d$
Volume names: /tsmpool/[Link] /tsmpool/[Link]

The FORMAT=DETAILED parameter on the QUERY BACKUPSET provides the client file
spaces contained in the backup set and the list of volumes of the backup set.

Displaying contents of backup sets

You can display information about the contents of backup sets by using the QUERY
BACKUPSETCONTENTS command. When you issue the query, the server displays only
one backup set at a time.

The server displays information about the files and directories that are contained in
a backup set. After you issue the query backupsetcontents jane engdata.3099
command, the following output is displayed:

Node Name Filespace Client’s Name for File

Name
------------------------ ---------- ---------------------------
JANE /srvr /deblock
JANE /srvr /deblock.c
JANE /srvr /[Link]
JANE /srvr /[Link]
JANE ... ......

Tip: To display the contents of an image backup set, specify DATATYPE=IMAGE on the
QUERY BACKUPSETCONTENTS command.

Displaying file space and file names

File space names and file names that can be in a different code page or locale than
the server do not display correctly in the Operations Center or the administrative
command-line interface. The data itself is backed up and can be restored properly,
but the file space or file name may display with a combination of invalid
characters or blank spaces.

If the file space name is Unicode enabled, the name is converted to the server's
code page for display. The results of the conversion for characters not supported
by the current code page depends on the operating system. For names that Tivoli
Storage Manager is able to partially convert, you may see question marks (??),
blanks, unprintable characters, or “...”. These characters indicate to the
administrator that files do exist. If the conversion is not successful, the name is
displayed as "...". Conversion can fail if the string includes characters that are not
available in the server code page, or if the server has a problem accessing system
conversion routines.

Chapter 14. Managing data for client nodes 551

 Deleting backup sets
When the server creates a backup set, the retention period assigned to the backup
set determines how long the backup set remains in the database.

Procedure

To delete a backup set named ENGDATA.3099, belonging to client node JANE,

created before 11:59 p.m. on March 18, 1999, issue the following command:
delete backupset jane engdata.3099 begindate=03/18/1999 begintime=23:59

To delete all backup sets belonging to client node JANE, created before 11:59 p.m.
on March 18, 1999, enter:
delete backupset jane * begindate=03/18/1999 begintime=23:59

Results

When that date passes, the server automatically deletes the backup set when
expiration processing runs. However, you can also manually delete the client's
backup set from the server before it is scheduled to expire by using the DELETE
BACKUPSET command.

After a backup set is deleted, the volumes return to scratch status if Tivoli Storage
Manager acquired them as scratch volumes. Scratch volumes associated with a
device type of FILE are deleted. However, if multiple backup sets are stored on the
volumes, all of the backup sets must expire or be deleted before the volumes are
returned to scratch.

Enabling clients to use subfile backup

A basic problem that remote and mobile users face today is connecting to storage
management services by using modems with limited bandwidth or poor line
quality. This creates a need for users to minimize the amount of data they send
over the network, as well as the time that they are connected to the network.

To help address this problem, you can use subfile backups. When a client's file has
been previously backed up, any subsequent backups are typically made of the
portion of the client's file that has changed (a subfile), rather than the entire file. A
base file is represented by a backup of the entire file and is the file on which
subfiles are dependent. If the changes to a file are extensive, a user can request a
backup on the entire file. A new base file is established on which subsequent
subfile backups are dependent.

This type of backup makes it possible for mobile users to reduce connection time,
network traffic, and the time it takes to do a backup.

To enable this type of backup, see “Setting up clients to use subfile backup” on
page 553.

Subfile backups

Assume that on a Monday, a user requests an incremental backup of a file called

[Link]. The user makes daily updates to the [Link] file and requests
subsequent backups.

552 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 The following table describes how Tivoli Storage Manager manages backups of this
file.

Day of
subsequent
Version backup What Tivoli Storage Manager backs up
One Monday The entire [Link] file (the base file)
Two Tuesday A subfile of [Link]. The server compares the file backed up
on Monday with the file that needs to be backed up on
Tuesday. A subfile containing the changes between the two
files is sent to the server for the backup.
Three Wednesday A subfile of [Link]. Tivoli Storage Manager compares the
file backed up on Monday with the file that needs to be
backed up on Wednesday. A subfile containing the changes
between the two files is sent to the server for the backup.

Related reference:
“Setting policy to enable point-in-time restore for clients” on page 527
“Policy for logical volume backups” on page 523

Setting up clients to use subfile backup

The subfile backup pertains to the sections of the files that have changed.

Procedure

To enable subfile backup, complete the following tasks:

1. On the server: You must set up the server to allow clients to back up subfiles.
Issue the SET SUBFILE command:
set subfile client
2. On the clients: The SUBFILEBACKUP, SUBFILECACHEPATH, and SUBFILECACHESIZE
options must be set in the client's options file ([Link]).
You can control these options from the server by including them in client
option sets. For example, you can disable subfile backup for individual client
nodes by setting SUBFILEBACKUP=NO in the client option set associated with
the client node. See “Creating client option sets on the server” on page 465 for
how to set up and use client option sets.
See Backup-Archive Clients Installation and User's Guide for more information
about the options.

Managing subfile backups

Tivoli Storage Manager manages subfiles that are restored, exported, imported, or
added to a backup set.

Chapter 14. Managing data for client nodes 553

 Restoring subfiles
When a client issues a request to restore subfiles, Tivoli Storage Manager restores
subfiles along with the corresponding base file back to the client. This process is
transparent to the client. That is, the client does not have to determine whether all
subfiles and corresponding base file were restored during the restore operation.

You can define (move) a backup set that contains subfiles to an earlier version of a
server that is not enabled for subfile backup. That server can restore the backup set
containing the subfiles to a client not able to restore subfiles. However, this process
is not recommended as it could result in a data integrity problem.

Exporting and importing subfiles

When subfiles are exported during an export operation, Tivoli Storage Manager
also exports the corresponding base file to volumes you specify.

When the base file and its dependent subfiles are imported from the volumes to a
target server and import processing is canceled while the base file and subfiles are
being imported, the server automatically deletes any incomplete base files and
subfiles that were stored on the target server.

Expiration processing of base files and subfiles

Because subfiles are useless without the corresponding base file, the server
processes base files eligible for expiration differently.

For example, when expiration processing runs, Tivoli Storage Manager recognizes a
base file as eligible for expiration but does not delete the file until all its dependent
subfiles have expired. For more information on how the server manages file
expiration, see “Running expiration processing to delete expired files” on page 512.

Adding subfiles to backup sets

When a subfile is added to a backup set, Tivoli Storage Manager includes its
corresponding base file with the backup set.

If the base file and dependent subfiles are stored on separate volumes when a
backup set is created, additional volume mounts may be required to create the
backup set.

Deleting base files

If a base file is deleted as a result of processing a DELETE VOLUME command, the
server recognizes its dependent subfiles and deletes them from the server as well.
Subfiles without the corresponding base file are incomplete and useless to the user.

Archiving data
Managing archive data on the server becomes important when you have client
nodes that archive large numbers (hundreds or thousands) of files every day.

If you archive files with automated tools that start the command-line client or API,
you might encounter large numbers. If performance degrades over time during an
archive operation, or you have a large amount of storage that is used by archives,
consider advanced techniques. See “Archive operations overview” on page 555 and
“Managing storage usage for archives” on page 555.

554 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Archive operations overview
IBM Tivoli Storage Manager allows clients to organize archived files into packages.
The description field of an archive request serves as the way to identify the
package.

All files that are archived with the same description become members of the same
archive package. If the user does not specify a description when archiving, the
client program provides a default description with each archive request. The
default description includes the date.

When files are archived, the client program archives the paths (directories) to those
files to preserve access permissions which are specific to the operating system.
Directories are also included in archive packages. If the same directory is archived
with different descriptions, the directory is stored once with each package. If a
command line user issues a QUERY ARCHIVE command, multiple entries for the same
directory may appear. Closer inspection shows that each entry has a different
description.

The GUI and Web client programs allow a user to navigate through a client node's
archives by first displaying all descriptions (the package identifiers), then the
directories, and finally the files. Users can retrieve or delete individual files or all
files in a directory. Command line client and API users can specify a description
when they archive files, or when they send requests to query, retrieve or delete
archived files.

When retrieving files, the server searches for the most current file versions. It will
search in an active-data storage pool associated with a FILE device class, if such a
pool exists.

Managing storage usage for archives

The amount of server database storage needed for archive operations is affected by
the use of descriptions for archiving and by the archiving of directories.

Consider the following two actions that you can take to minimize the storage
usage:
Minimize the number of unique descriptions
You can reduce storage usage by archiving more files into fewer packages
(by reducing the number of unique descriptions). The amount of storage
used for directories is also affected by the number of packages. If you
archive a file three different times using three different descriptions, the
server stores both the file and the directory three times, once in each
package. If you archives the same file three different times using just one
description, the server stores the file three times, but the directory is stored
just one time.
Archive directories only if needed
Archiving directories might be necessary if the directories are needed to
group files for query or retrieve, or if the directory-level access permission
information needs to be archived.

The users of the GUI and Web client programs need descriptions to aid in
navigation, to find archived files. You can minimize storage usage for archives by
reducing the number of packages. For client nodes that are always accessed via the
command-line interface you can also use some other techniques.

Chapter 14. Managing data for client nodes 555

 See the following topics for more information:
“Minimizing archive storage use”
“Reducing archive storage use for client nodes with many archives”
“Preventing archive of directory-level access permissions” on page 557

Minimizing archive storage use

You can minimize the storage used for new archives for a client node that is
always accessed through the command-line interface by ensuring that each file
archived by the client node has a unique name and that you use the same
description or a limited number of unique descriptions to group and identify
archived files for the client node.

If the user follows these guidelines, the client node will have one or a limited
number of archive packages. Because of the small number of packages, there are
only small numbers of copies of each directory entry. The savings in storage space
that result are noticeable when files with the same path specification are archived
multiple times over multiple days.

Having a unique description for an archive is not necessary if archived files

themselves have unique names. For example, files might already have a date or
time stamp encoded in their names. The same description can be used each time a
client node archives files. The user must specify the description to avoid getting
the default for the description, which is Archive Date: yyyy/mm/dd. The user can
specify a description as short as one character, or as simple as a blank space (but
not a null entry). For example, the user enters this command:
dsmc archive c:\[Link] -description=" "

If the user follows these guidelines, you do not need to convert the node because
converting the node affects archive descriptions. A client node that is used with
these guidelines has only one or a small number of archive descriptions, so there is
no improvement in performance to be gained by converting the node.

See the Backup-Archive Clients Installation and User's Guide for details about archive
operations and client options.

Reducing archive storage use for client nodes with many

archives
If a client node already has a large number of archive packages, you can reduce
the storage use by updating that node's archives.

Do not run the UPDATE ARCHIVE command while any other processing for the node
is running. If this command is issued for a node with any other object insertion or
deletion activity occurring at the same time, locking contention may occur. This
may result in processes and sessions hanging until the resource timeout is reached
and the processes and sessions terminate.

When you update archives for a node, you have two choices for the action to take:
Delete directory entries in all archive packages
This action preserves the archive packages, but removes directory entries
for all packages, reducing the amount of storage used for archives. Do this
only when directory entries that include access permissions are not needed
in the archive packages, and the paths are not needed to query or retrieve
a group of files. The amount of reduction depends on the number of
packages and the number of directory entries. For example, to remove
directory entries for the client node SNOOPY, enter this command:

556 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 update archive snoopy deletedirs

Attention: After you delete the directory entries, the directory entries
cannot be recreated in the archive packages. Do not use this option if users
of the client node need to archive access permissions for directories.
Reduce the number of archive packages to a single package for the node
This action removes all unique descriptions, thereby reducing the number
of archive packages to one for the client node. Do this only when the
descriptions are not needed and are causing large use of storage. This
action also removes directory entries in the archive packages. Because there
is now one package, there is one entry for each directory. For example, to
reduce the archive packages to one for the client node SNOOPY, enter this
command:
update archive snoopy resetdescriptions

After updating the archives for a node in this way, keep the archive
package count to a minimum.

Attention: You cannot recreate the packages after the descriptions have
been deleted. Do not use this option if users of the client node manage
archives by packages, or if the client node is accessed via the GUI or Web
client interface.

Preventing archive of directory-level access permissions

Command-line users who do not need to archive the directory-level access
permissions can reduce storage requirements by using the v2archive option with
the archive request. This option prevents the creation of directory entries in archive
packages.

See Backup-Archive Clients Installation and User's Guide for details about the option.

Tip: The GUI and Web client programs use the directories to allow users to
navigate to the archived files. This option is not recommended for GUI or Web
client interface users.

Chapter 14. Managing data for client nodes 557

558 IBM Tivoli Storage Manager for Windows: Administrator's Guide
Chapter 15. Scheduling operations for client nodes
Scheduling client operations can provide better protection for data, because
operations can occur consistently without user intervention. Scheduling can also
maximize utilization of resources. When client backups are scheduled at times of
lower usage, the impact on the network is minimized.

About this task

You can schedule operations such as the following:

v Archiving and retrieving client data.
v Running operating system commands.
v Running macro or command files that contain operating system commands,
commands, or both. You can schedule a command file to run on clients or
application clients.
v Backing up and restoring client data and application client data.

Administrators can perform the following activities to schedule client operations:

Tasks:
“Scheduling a client operation” on page 560
“Starting the scheduler on the clients” on page 562
“Displaying information about schedules” on page 571
“Starting the scheduler on the clients” on page 562
“Displaying information about schedules” on page 571
“Creating schedules for running command files” on page 564
“Updating the client options file to automatically generate a new password” on page 565

Prerequisites to scheduling operations

To interact with Tivoli Storage Manager for scheduling operations, client machines
must meet certain prerequisites.
v The client node must be registered with the server. For information, see
Chapter 11, “Adding client nodes,” on page 419.
v The client options file ([Link]) must contain the network address of the server
that the client will contact for services. See “Connecting nodes with the server”
on page 425 for more information.
v The scheduler must be configured and started on the client machine. See the
Backup-Archive Clients Installation and User's Guide for details.

© Copyright IBM Corp. 1993, 2013 559

Scheduling a client operation
To automate client operations, you can define new schedules.

About this task

When you define a schedule, you assign it to a specific policy domain. You can
define more than one schedule for each policy domain.

Procedure

To set up a client schedule on the server, complete the following steps:

1. Define a schedule by using the DEFINE SCHEDULE command. See “Defining client
schedules” on page 561.
2. Associate client nodes with the schedule by using the DEFINE ASSOCIATION
command. See “Associating client nodes with schedules” on page 562.
3. Ensure that the clients start the client scheduler. See “Starting the scheduler on
the clients” on page 562.
4. View schedule information and check that the schedule completed successfully
by using the QUERY SCHEDULE and QUERY EVENT commands. See “Displaying
schedule information” on page 563.

Example

The following example shows the commands that are used to schedule client
backups for three registered client nodes that are assigned to the STANDARD
policy domain.
1. Schedule an incremental backup and associate the schedule with the clients.
define schedule standard daily_incr action=incremental -
starttime=23:00

define association standard daily_incr bill,mark,mercedes

The schedule, named DAILY_INCR, is for the Tivoli Storage Manager default
policy domain, named STANDARD. The default specifies backup to the disk
storage pool BACKUPPOOL. This schedule calls for a schedule window with
the following characteristics:
v Begins on the date the schedule is defined (the default) at 11:00 p.m.
v Lasts for 1 hour (the default)
v Is repeated daily (the default)
v Stays in effect indefinitely (the default)
2. Start the client scheduler.
For the schedules to become active for a workstation, a user must start the
scheduler from the node.
dsmc schedule

To help ensure that the scheduler is running on the clients, start the client
acceptor daemon (CAD) or client acceptor service. The include-exclude list on
each client also affects which files are backed up or archived by the two
schedules that are defined in the preceding steps. For example, if a file is
excluded from backup with an EXCLUDE statement, the file is not backed up
when the DAILY_INCR schedule runs.

560 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 3. Because the DAILY_INCR schedule is to run daily, you can verify whether it is
working on the day after you define the schedule and associate it with clients.
If the schedule runs successfully, the status is Completed.
query event standard daily_incr begindate=today-1

You can limit the query of events to display only schedules that did not run as
expected. For example, you can use the following command daily to see which
clients did not run the DAILY_INCR schedule the previous day:
query event standard daily_incr begindate=today-1 -
exceptionsonly=yes
Schedules that did not complete successfully have a status of Failed, Missed, or
Severed.
4. Check the results of the schedule DAILY_INCR on one of the clients that was
associated with that schedule. For most clients, information about what
happens when a schedule runs is stored in the file [Link]. See the
Backup-Archive Clients Installation and User's Guide.

What to do next

You can modify, copy, and delete any schedule that you create. For more
information, see Chapter 16, “Managing schedules for client nodes,” on page 569.

Defining client schedules

When you schedule client operations, you need to know what operation you want
to run and the time, day, and frequency to run your operation.

About this task

Task Required Privilege Class
Define client schedules for any policy System or unrestricted policy
domain
Define client schedules for specific policy System, unrestricted policy, or restricted
domains policy for those domains

To define a schedule for daily incremental backups, use the DEFINE SCHEDULE
command. You must specify the policy domain to which the schedule belongs and
the name of the schedule (the policy domain must already be defined). For
example:
define schedule engpoldom daily_backup starttime=21:00
duration=2 durunits=hours

This command results in the following:

v Schedule DAILY_BACKUP is defined for policy domain ENGPOLDOM.
v The scheduled action is an incremental backup, which is the default.
v The priority for the operation is 5, which is the default. If schedules conflict, the
schedule with the highest priority (lowest number) runs first.
v The schedule window begins at 9:00 p.m., and the schedule itself has 2 hours to
start.
v The start window is scheduled every day, which is the default.
v The schedule never expires, which is the default.
v The schedule style is classic, which is the default.

Chapter 15. Scheduling operations for client nodes 561

 As a best practice, define schedules with durations longer than 10 minutes. This
action gives the Tivoli Storage Manager scheduler enough time to process the
schedule and prompt the client.

To change the defaults, see the DEFINE SCHEDULE command in the Administrator's
Reference.

Associating client nodes with schedules

Client nodes process operations according to the schedules associated with the
nodes. A client node can be associated with more than one schedule. However, a
node must be assigned to the policy domain to which a schedule belongs.

About this task

You must have system privilege, unrestricted policy, or restricted policy (for the
policy domain to which the schedule belongs) to associate client nodes with
schedules. Issue the DEFINE ASSOCIATION command to associate client nodes with a
schedule.

Procedure

Complete the following step to associate the ENGNODE client node with the
WEEKLY_BACKUP schedule, both of which belong to the ENGPOLDOM policy
domain:
define association engpoldom weekly_backup engnode

Results

After a client schedule is defined, you can associate client nodes with it by
identifying the following information:
v Policy domain to which the schedule belongs
v List of client nodes to associate with the schedule

Starting the scheduler on the clients

The client scheduler must be started before work scheduled by an administrator
can be initiated.

Administrators must ensure that users start the Tivoli Storage Manager scheduler
on the client or application client directory, and that the scheduler is running at the
schedule start time. After the client scheduler starts, it continues to run and
initiates scheduled events until it is stopped.

The way that users start the Tivoli Storage Manager scheduler varies, depending
on the operating system that the machine is running. The user can choose to start
the client scheduler automatically when the operating system is started, or can
start it manually at any time. The user can also have the client acceptor manage
the scheduler, starting the scheduler only when needed. For instructions on these
tasks, see the Backup-Archive Clients Installation and User's Guide.

The client and the Tivoli Storage Manager server can be set up to allow all sessions
to be initiated by the server. See “Server-initiated sessions” on page 433 for
instructions.

Note: Tivoli Storage Manager does not recognize changes that you made to the
client options file while the scheduler is running. For Tivoli Storage Manager to

562 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 use the new values immediately, you must stop the scheduler and restart it.

Displaying schedule information

When you request information about schedules, the server displays details. You
must have administrative privilege to display information about scheduled
operations.

About this task

v Schedule name
v Policy domain name
v Type of operation to perform
v Start date and time for the initial startup window
v Duration of the startup window
v Time period between startup windows (if using a classic schedule)
v Day (classic schedule) or days (enhanced schedule) of the week on which
scheduled operations can begin

The following output shows an example of a report for a classic schedule that is
displayed after you enter:
query schedule engpoldom
Domain * Schedule Name Action Start Date/Time Duration Period Day
------------ - ---------------- ------ -------------------- -------- ------ ---
ENGPOLDOM MONTHLY_BACKUP Inc Bk 09/04/2002 [Link] 2 H 2 Mo Sat
ENGPOLDOM WEEKLY_BACKUP Inc Bk 09/04/2002 [Link] 4 H 1 W Sat

For enhanced schedules, the standard schedule format displays a blank period
column and an asterisk in the day of week column. Issue FORMAT=DETAILED to
display complete information about an enhanced schedule. Refer to the
Administrator's Reference for command details. The following output shows an
example of a report for an enhanced schedule that is displayed after you enter:
query schedule engpoldom
Domain * Schedule Name Action Start Date/Time Duration Period Day
------------ - ---------------- ------ -------------------- -------- ------ ---
ENGPOLDOM MONTHLY_BACKUP Inc Bk 09/04/2002 [Link] 2 H 2 Mo Sat
ENGPOLDOM WEEKLY_BACKUP Inc Bk 09/04/2002 [Link] 4 H (*)

Checking the status of scheduled operations

A schedule completes successfully if the command associated with the schedule is
successfully issued. The success of the issued command is independent on the
success of the schedule.

About this task

You need to ask these two questions:

v Did the schedule run successfully?
To determine the success of a scheduled operation, query the server. Each
scheduled client operation is called an event, and is tracked by the server. You
can get information about projected and actual scheduled processes by using the
QUERY EVENT command. You can get information about scheduled processes
that did not complete successfully by using exception reporting with this
command.

Chapter 15. Scheduling operations for client nodes 563

 For example, you can issue the following command to find out which events
were missed (did not start) in the ENGPOLDOM policy domain for the
WEEKLY_BACKUP schedule in the previous week:
query event engpoldom weekly_backup begindate=-7 begintime=now
enddate=today endtime=now exceptionsonly=yes
For more information about managing event records, see “Managing event
records” on page 573.
v Did the operation or commands run as a result of the schedule run successfully?
To determine the success of the commands issued as the result of a successful
schedule, you can:
– Check the client's schedule log.
The schedule log is a file that contains information such as the statistics about
the backed-up objects, the name of the server backing up the objects, and the
time and date of the next scheduled operation. By default, Tivoli Storage
Manager stores the schedule log as a file called [Link] and places the
file in the directory where the Tivoli Storage Manager backup-archive client is
installed. Refer to Backup-Archive Clients Installation and User's Guide for more
information.
– Check the server's activity log.
Search or query the activity log for related messages. For example, search for
messages that mention the client node name, within the time period that the
schedule ran. For example:
query actlog begindate=02/23/2001 enddate=02/26/2001 originator=client
nodename=hermione
– Issue the QUERY EVENT command with FORMAT=DETAILED, and view the
Result field of the output screen. For example:
query event nodes=joe domain2 standard begindate=02/26/2002 enddate=02/27/2002
format=detailed
Refer to Backup-Archive Clients Installation and User's Guide for an explanation
of the Result field.

Creating schedules for running command files

For some clients, you may want to run a command for a different application
before running a Tivoli Storage Manager backup. For example, you may want to
stop a database application, back up files with Tivoli Storage Manager, and then
restart the application. To do this, you can schedule the running of a command file.
Application clients require schedules that run command files.

About this task

A command file (also known as a macro or batch file on different operating

systems) is stored on the client. This file contains a sequence of commands that are
intended to be run during a scheduled start date and time window. Commands
can include operating system commands, the Tivoli Storage Manager client's
DSMC command, and commands for other applications.

The default schedules do not support command files so you must create a new
schedule to schedule command files.

To use command files, administrators must create schedules with the

ACTION=MACRO parameter. For example, you can define a schedule called
DAILY_INCR that will process a command file called c:\[Link] on the client:

564 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 define schedule standard daily_incr description="daily incremental file"
action=macro objects="c:\[Link]" starttime=18:00 duration=5
durunits=minutes period=1 perunits=day dayofweek=any

Associate the client with the schedule and ensure that the scheduler is started on
the client or application client directory. The schedule runs the file called
c:\[Link] once a day between 6:00 p.m. and 6:05 p.m., every day of the week.

Updating the client options file to automatically generate a new

password
If the server uses password authentication, clients must use passwords. Passwords
are then also required for the server to process scheduled operations for client
nodes.

About this task

If a password expires and is not updated, scheduled operations fail. You can
prevent failed operations by allowing Tivoli Storage Manager to generate a new
password when the current password expires. If you set the PASSWORDACCESS
option to GENERATE in the Tivoli Storage Manager client options file, [Link],
Tivoli Storage Manager automatically generates a new password for your client
node each time it expires, encrypts and stores the password in a file, and retrieves
the password from that file during scheduled operations. You are not prompted for
the password.

The PASSWORDACCESS GENERATE option is also required in other situations,

such as when you want to use the Web backup-archive client to access a client
node. See the Backup-Archive Clients Installation and User's Guide for more
information.

Configuring the scheduler to run under the site-server account

Standard Tivoli Storage Manager security requirements apply to the Microsoft
Exchange application client. That is, the application client must be registered to
Tivoli Storage Manager and use the appropriate node name and password when
connecting with the Tivoli Storage Manager server.

About this task

To access the Exchange Server APIs, the application client must be running under
the Site Services Account. The Site Services Account is the account under which the
Exchange services are running. The scheduler service must logon using this
account. The scheduler service account information can be specified using the
services applet in the Windows Control Panel. For more information about the Site
Services Account, see the Microsoft Exchange Server documentation.

Chapter 15. Scheduling operations for client nodes 565

Overview of the Tivoli Storage Manager scheduler running as a
Windows service
Client nodes on Windows systems run scheduled operations through the Tivoli
Storage Manager scheduler, as do other client nodes. However, client nodes on
Windows benefit from running the Tivoli Storage Manager scheduler as a service.
This feature allows the Tivoli Storage Manager scheduler to run as a protected
account, but does not require you to be logged on to run the service.

The Client Acceptor daemon (CAD) cannot be used by a client node when
SESSIONINITIATION=SERVERONLY.

Figure 73 shows three Windows machines configured for scheduling.

Figure 73. Overview of the Tivoli Storage Manager Scheduler as a Windows Service

Machine 1 — Remote Client

Machine 1 contains only a backup-archive client. It has been configured to
back up its file system to the server on Machine 4. The Tivoli Storage
Manager scheduler has been configured. The client uses the information in

566 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 the client options file ([Link]) to connect to the server. The client options
file contains the server address of the server (Machine 4).
Machine 2 — Tivoli Storage Manager application client
Machine 2 contains an application client and a backup-archive client. Note
that the backup-archive client is not connected to the server. The
backup-archive client on Machine 2 has been installed only to provide
access to the DSMCUTIL utility, not to back up the resident file system.
After the backup-archive client has been used to install the scheduler, it
may continue to be required for updating the password or other
scheduling service parameters. The application client uses the information
in its client options file to connect to the server.
Machine 3 — Tivoli Storage Manager application client and backup-archive
client on the same machine
Machine 3 contains an application client and a backup-archive client. The
application client backs up databases and transaction logs to the server.
The backup-archive client backs up other files on Machine 3's file system,
including the database administrator's online documentation and
performance records. The scheduler wizard is used to install a scheduler
service for the backup-archive client and a second scheduler service for the
application client. The two clients use the information in their client
options files to connect to the server.
Machine 4 — Server and local backup-archive client on the same machine
Machine 4 contains a server and a local backup-archive client. For this
client (packaged with the Tivoli Storage Manager server), the scheduler is
configured during setup, and the administrator needs only to update the
password.

Chapter 15. Scheduling operations for client nodes 567

568 IBM Tivoli Storage Manager for Windows: Administrator's Guide
Chapter 16. Managing schedules for client nodes
You can manage and coordinate Tivoli Storage Manager schedules for registered
client nodes.

About this task

Administrators can perform the following tasks:

Tasks:
“Managing node associations with schedules” on page 571
“Specifying one-time actions for client nodes” on page 583
“Managing event records” on page 573
“Managing the throughput of scheduled operations” on page 576
“Managing IBM Tivoli Storage Manager schedules”

For a description of what Tivoli Storage Manager views as client nodes, see
Chapter 11, “Adding client nodes,” on page 419. For information about the
scheduler and creating schedules, see Chapter 15, “Scheduling operations for client
nodes,” on page 559

Managing IBM Tivoli Storage Manager schedules

You can perform the following activities to manage schedules.

About this task

Task Required Privilege Class
Verify that the schedule ran Any administrator
Add, copy, modify, or delete client schedules System or unrestricted policy
in any policy domain
Add, copy, modify, or delete client schedules System, unrestricted policy, or restricted
for specific policy domains policy for those domains
Display information about scheduled Any administrator
operations

Adding new schedules

You can add and associate schedules by using the Tivoli Storage Manager
command-line interface.

About this task

You can add new Tivoli Storage Manager schedules by using the DEFINE
SCHEDULE command.

After you add a new schedule, associate the node with the schedule. For more
information, see “Defining client schedules” on page 561.

© Copyright IBM Corp. 1993, 2013 569

 Copying existing schedules
You can create new schedules by copying existing schedules to the same policy
domain or a different policy domain. The schedule description and all schedule
parameter values are copied to the new schedule. You can then modify the new
schedule to meet site-specific requirements.

About this task

Client node associations are not copied to the new schedule. You must associate
client nodes with the new schedule before it can be used. The associations for the
old schedule are not changed.

To copy the WINTER schedule from policy domain DOMAIN1 to DOMAIN2 and
name the new schedule WINTERCOPY, enter:
copy schedule domain1 winter domain2 wintercopy

For information, see “Associating client nodes with schedules” on page 562.

Modifying schedules
You can modify existing schedules by issuing the UPDATE SCHEDULE command.

About this task

You can use the server command line to modify existing schedules in a policy
domain.

Procedure

To modify the ENGWEEKLY client schedule in the ENGPOLDOM policy domain, issue the
following command:
update schedule engpoldom engweekly period=5 perunits=days

Results

The ENGWEEKLY schedule is updated so that the incremental backup period is now
every five days.

Deleting schedules
When you delete a schedule, Tivoli Storage Manager deletes all client node
associations for that schedule.

About this task

To delete the schedule WINTER in the ENGPOLDOM policy domain, enter:

delete schedule engpoldom winter

Rather than delete a schedule, you may want to remove all nodes from the
schedule and save the schedule for future use. For information, see “Removing
nodes from schedules” on page 573.

See “Associating client nodes with schedules” on page 562 for more information.

570 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 Displaying information about schedules
You can display information about schedules.

About this task

The following information is displayed:

v Schedule name
v Policy domain name
v Type of operation to be performed
v Start date and time for the initial startup window
v Duration of the startup window
v Time period between startup windows (if using a classic schedule)
v Day (classic schedule) or days (enhanced schedule) of the week on which
scheduled operations can begin

The following output shows an example of a report for a classic schedule that is
displayed after you enter:
query schedule engpoldom
Domain * Schedule Name Action Start Date/Time Duration Period Day
------------ - ---------------- ------ -------------------- -------- ------ ---
ENGPOLDOM MONTHLY_BACKUP Inc Bk 09/04/2002 [Link] 2 H 2 Mo Sat
ENGPOLDOM WEEKLY_BACKUP Inc Bk 09/04/2002 [Link] 4 H 1 W Sat

For enhanced schedules, the standard schedule format displays a blank period
column and an asterisk in the day of week column. Issue FORMAT=DETAILED to
display complete information about an enhanced schedule. Refer to the
Administrator's Reference for command details. The following output shows an
example of a report for an enhanced schedule that is displayed after you enter:
query schedule engpoldom
Domain * Schedule Name Action Start Date/Time Duration Period Day
------------ - ---------------- ------ -------------------- -------- ------ ---
ENGPOLDOM MONTHLY_BACKUP Inc Bk 09/04/2002 [Link] 2 H 2 Mo Sat
ENGPOLDOM WEEKLY_BACKUP Inc Bk 09/04/2002 [Link] 4 H (*)

Managing node associations with schedules

You can add and delete node associations from schedules. Nodes can be associated
with more than one schedule.

About this task

You can perform the following activities to manage associations of client nodes
with schedules.

Task Required Privilege Class

Add new nodes to existing schedules System or restricted policy over the domain
to which the node is assigned
Move nodes to existing schedules System or restricted policy over the domain
to which the node is assigned
Delete nodes associated with a schedule System or restricted policy over the domain
to which the node is assigned

Chapter 16. Managing schedules for client nodes 571

 Task Required Privilege Class
Display nodes associated with a specific Any administrator
schedule

Adding new nodes to existing schedules

You can add new nodes to existing schedules by associating the node with the
schedule.

About this task

To associate client nodes with a schedule, you can use the following method:
Issue the DEFINE ASSOCIATION command from the command-line interface.
For more information, see “Associating client nodes with schedules” on page 562.

Moving nodes from one schedule to another

You can move a node from one schedule to another schedule.

Procedure
1. Associate the node to the new schedule. For information, see “Adding new
nodes to existing schedules.”
2. Delete the association of that node from the original schedule.
Related tasks:
“Associating client nodes with schedules” on page 562
“Removing nodes from schedules” on page 573

Displaying nodes associated with schedules

You can display information about the nodes that are associated with a specific
schedule.

About this task

For example, you should query an association before deleting a client schedule.

Figure 74 shows the report that is displayed after you enter:

query association engpoldom

Policy Domain Name: ENGPOLDOM

Schedule Name: MONTHLY_BACKUP
Associated Nodes: MAB SSTEINER

Policy Domain Name: ENGPOLDOM

Schedule Name: WEEKLY_BACKUP
Associated Nodes: MAB SSTEINER

Figure 74. Query association output

572 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 Removing nodes from schedules
When you remove the association of a node to a client schedule, the client no
longer runs operations specified by the schedule. However, the remaining client
nodes still use the schedule.

About this task

To delete the association of the ENGNOD client with the ENGWEEKLY schedule,
in the policy domain named ENGPOLDOM, enter:
delete association engpoldom engweekly engnod

Instead of deleting a schedule, you may want to delete all associations to it and
save the schedule for possible reuse in the future.

Managing event records

Each scheduled client operation is called an event. All scheduled events, including
their status, are tracked by the server. An event record is created in the server
database whenever a scheduled event is completed or missed.

About this task

You can perform the following activities to manage event records:

Task Required Privilege Class

Display information about scheduled events Any administrator
Set the retention period for event records System
Delete event records System or unrestricted policy

Displaying information about scheduled events

To help manage schedules for client operations, you can request information about
scheduled and completed events by using the QUERY EVENT command.

About this task

v To get information about past and projected scheduled processes, use a simple
query for events. If the time range you specify includes the future, the results
show which events should occur in the future based on current schedules.
v To get information about scheduled processes that did not complete successfully,
use the exceptions-only option with the query.

To minimize the processing time when querying events:

v Minimize the time range
v For client schedules, restrict the query to those policy domains, schedules, and
client node names for which information is required

You can also find information about scheduled events by checking the log file
described in “Checking the schedule log file” on page 575.

Chapter 16. Managing schedules for client nodes 573

 Displaying all client schedule events
You can display information about all client events by issuing the QUERY EVENT
command. The information includes events for both successful and failed
schedules. If the administrator specifies a time range that includes the future,
Tivoli Storage Manager displays future events with a status of future.

About this task

Figure 75 shows an example of a report for client node GOODELL that is

displayed after you enter:
query event standard weekly_backup node=goodell enddate=today+7

Scheduled Start Actual Start Schedule Name Node Name Status

-------------------- -------------------- ------------- ------------- ---------
09/04/2002 [Link] 09/04/2002 [Link] WEEKLY_BACKUP GOODELL Started
09/16/2002 [Link] WEEKLY_BACKUP GOODELL Future

Figure 75. Events for a node

Displaying events that ended unsuccessfully

You can display information about scheduled events that ended unsuccessfully by
using exception reporting.

About this task

For example, you can issue the following command to find out which events were
missed in the previous 24 hours, for the DAILY_BACKUP schedule in the
STANDARD policy domain:
query event standard daily_backup begindate=-1 begintime=now
enddate=today endtime=now exceptionsonly=yes

Figure 76 shows an example of the results of this query. To find out why a
schedule was missed or failed, you may need to check the schedule log on the
client node itself. For example, a schedule can be missed because the scheduler
was not started on the client node.

Scheduled Start Actual Start Schedule Name Node Name Status

-------------------- -------------------- ------------- ------------- ---------
09/04/2002 [Link] DAILY_BACKUP ANDREA Missed
09/04/2002 [Link] DAILY_BACKUP EMILY Missed

Figure 76. Exception report of events

Displaying past events

If you query the server for events, the server may display past events even if the
event records have been deleted.

About this task

Such events are displayed with a status of Uncertain, indicating that complete
information is not available because the event records have been deleted. To
determine if event records have been deleted, check the message that is issued
after the DELETE EVENT command is processed.

574 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 Checking the schedule log file
The Tivoli Storage Manager client stores detailed information about each scheduled
event in a file. This file contains information such as the statistics about the
backed-up objects, the name of the server to which the objects are backed up, and
the time and date of the next scheduled operation.

About this task

The default name for the schedule log file is [Link]. The file is located in
the directory where the Tivoli Storage Manager backup-archive client is installed.
You can override this file name and location by specifying the SCHEDLOGNAME option
in the client options file. See the Backup-Archive Clients Installation and User's
Guide for more information.

Managing event records in the server database

By default, the server retains event records for 10 days before automatically
removing them from the database. The server automatically deletes event records
from the database after the event retention period has passed and after the startup
window for the event has elapsed.

About this task

You can specify how long event records stay in the database before the server
automatically deletes them by using the SET EVENTRETENTION command. You
can also manually delete event records from the database, if database space is
required.

Setting the event retention period

You can modify the retention period for event records in the database.

About this task

To change the retention period to 15 days, enter:

set eventretention 15

Manually deleting event records

You may want to manually delete event records to increase available database
space.

About this task

For example, to delete all event records written prior to 11:59 p.m. on June 30,
2002, enter:
delete event 06/30/2002 23:59

Chapter 16. Managing schedules for client nodes 575

Managing the throughput of scheduled operations
In the Tivoli Storage Manager environment where many nodes attempt to initiate
scheduled operations simultaneously, you may have to manage scheduling
throughput. You can choose a scheduling mode, and you can control how often
client nodes contact the server to perform a scheduled operation.

About this task

Administrators can perform the following activities to manage the throughput of

scheduled operations.

Task Required Privilege Class

Modify the default scheduling mode System
Modify the scheduling period for System
incremental backup operations
Balance the scheduled workload for the System
server
Set the frequency at which client nodes System
contact the server

Modifying the default scheduling mode

Tivoli Storage Manager provides two scheduling modes: client-polling and
server-prompted. The mode indicates how client nodes interact with the server for
scheduling operations.

About this task

With client-polling mode, client nodes poll the server for the next scheduled event.
With server-prompted mode, the server contacts the nodes at the scheduled start
time. By default, the server permits both scheduling modes. The default (ANY)
allows nodes to specify either scheduling mode in their client options files. You can
modify this scheduling mode.

If you modify the default server setting to permit only one scheduling mode, all
client nodes must specify the same scheduling mode in their client options file.
Clients that do not have a matching scheduling mode will not process the
scheduled operations. The default mode for client nodes is client-polling.

The scheduler must be started on the client node's machine before a schedule can
run in either scheduling mode.

For more information about modes, see “Overview of scheduling modes” on page
577.

By default, clients contact the server (client-polling scheduling mode and

SESSIONINITIATION=CLIENTORSERVER). If SESSIONINITIATION is set to the default
CLIENTORSERVER value, you can use either client-polling or server-prompted
scheduling modes. The client might start sessions with the server by
communicating on the TCP/IP port that was defined with a server option.
Server-prompted scheduling also can be used to prompt the client to connect to the
server.

576 IBM Tivoli Storage Manager for Windows: Administrator's Guide

You can instead prevent clients from starting sessions, and allow only the server to
start sessions with clients.

Procedure

To limit the start of backup-archive client sessions to the server only, complete the
following steps for each node:
1. Use the REGISTER NODE command or the UPDATE NODE command to change the
value of the SESSIONINITIATION parameter to SERVERONLY, Specify the high-level
address and low-level address options. These options must match what the
client is using, otherwise the server will not know how to contact the client.
2. Set the scheduling mode to server-prompted. All sessions must be started by
server-prompted scheduling on the port that was defined for the client with the
REGISTER NODE or the UPDATE NODE commands.
3. Ensure that the scheduler on the client is started. You cannot use the client
acceptor (dsmcad) to start the scheduler when SESSIONINITIATION is set to
SERVERONLY.

Overview of scheduling modes

With client-polling mode, client nodes poll the server for the next scheduled event.
With server-prompted mode, the server contacts the nodes at the scheduled start
time.

See Table 49 on page 578 and Table 48 for the advantages and disadvantages of
client-polling and server-prompted modes.
Table 48. Client-Polling mode
How the mode works Advantages and disadvantages

1. A client node queries the server at v Useful when a high percentage of clients
prescribed time intervals to obtain a start the scheduler manually on a daily
schedule. This interval is set with a client basis, for example when their workstations
option, QUERYSCHEDPERIOD. For are powered off nightly.
information about client options, refer to v Supports randomization, which is the
the appropriate Backup-Archive Clients random distribution of scheduled start
Installation and User's Guide. times. The administrator can control
2. At the scheduled start time, the client randomization. By randomizing the start
node performs the scheduled operation. times, Tivoli Storage Manager prevents all
3. When the operation completes, the client clients from attempting to start the
sends the results to the server. schedule at the same time, which could
overwhelm server resources.
4. The client node queries the server for its
next scheduled operation. v Valid with all communication methods.

Chapter 16. Managing schedules for client nodes 577

 Table 49. Server-Prompted mode
How the mode works Advantages and disadvantages

1. The server contacts the client node when v Useful if you change the schedule start
scheduled operations need to be time frequently. The new start time is
performed and a server session is implemented without any action required
available. from the client node.
2. When contacted, the client node queries v Useful when a high percentage of clients
the server for the operation, performs the are running the scheduler and are waiting
operation, and sends the results to the for work.
server. v Useful if you want to restrict sessions to
server-initiated.
v Does not allow for randomization of
scheduled start times.
v Valid only with client nodes that use
TCP/IP to communicate with the server.

Modifying the scheduling mode on the server

If you modify the default so that the server permits only one scheduling mode for
the server, all clients must specify the same scheduling mode in their client options
file. Clients that do not have a matching scheduling mode do not process
scheduled operations.

About this task

Client-Polling Scheduling Mode: To have clients poll the server for scheduled
operations, enter:
set schedmodes polling

Ensure that client nodes specify the same mode in their client options files.

Server-Prompted Scheduling Mode: To have the server prompt clients for

scheduled operations, enter:
set schedmodes prompted

Ensure that client nodes specify the same mode in their client options files.

Any Scheduling Mode: To return to the default scheduling mode so that the
server supports both client-polling and server-prompted scheduling modes, enter:
set schedmodes any

Client nodes can then specify either polling or prompted mode.

Modifying the default scheduling mode on client nodes

Users can set the scheduling mode on client nodes.

About this task

They specify either the client-polling or the server-prompted scheduling mode on

the command line or in the client user options file. (On UNIX and Linux systems,
root users set the scheduling mode in the client system options file.)

For more information, refer to the appropriate Backup-Archive Clients Installation and
User's Guide.

578 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Specifying the schedule period for incremental backup
operations
When you define a backup copy group, you specify the copy frequency, which is
the minimum interval between successive backups of a file.

About this task

When you define a schedule, you specify the length of time between processing of
the schedule. Consider how these interact to ensure that the clients get the backup
coverage that you intend.

See “Defining and updating a backup copy group” on page 502.

Balancing the scheduled workload for the server

You can control the server’s workload and ensure that the server can perform all
scheduled operations within the specified window.

About this task

To enable the server to complete all schedules for clients, you may need to use trial
and error to control the workload. To estimate how long client operations take, test
schedules on several representative client nodes. Keep in mind, for example, that
the first incremental backup for a client node takes longer than subsequent
incremental backups.

You can balance the server’s scheduled workload by:

v Adjusting the number of sessions that the server allocates to scheduled
operations
v Randomizing scheduled start time for client operations (if clients use
client-polling scheduling mode)
v Increasing the length of the startup window

Setting the number of sessions the server allocates to scheduled

operations
The maximum number of concurrent client/server sessions is defined by the
MAXSESSIONS server option.

About this task

Of these sessions, you can set a maximum percentage to be available for processing
scheduled operations. Limiting the number of sessions available for scheduled
operations ensures that sessions are available when users initiate any unscheduled
operations, such as restoring file or retrieving files.

If the number of sessions for scheduled operations is insufficient, you can increase
either the total number of sessions or the maximum percentage of scheduled
sessions. However, increasing the total number of sessions can adversely affect
server performance. Increasing the maximum percentage of scheduled sessions can
reduce the server availability to process unscheduled operations.

For example, assume that the maximum number of sessions between client nodes
and the server is 80. If you want 25% of these sessions to be used by for scheduled
operations, enter:
set maxschedsessions 25

Chapter 16. Managing schedules for client nodes 579

 The server then allows a maximum of 20 sessions to be used for scheduled
operations.

The following table shows the trade-offs of using either the SET
MAXSCHEDSESSIONS command or the MAXSESSIONS server option.

An administrator can... Using... With the result

Increase the total number of MAXSESSIONS server option May adversely affect the
sessions server’s performance
Increase the total number of SET MAXSCHEDSESSIONS May reduce the server’s
sessions allocated to command ability to process
scheduled operations unscheduled operations

For information about the MAXSESSIONS option and the SET

MAXSCHEDSESSIONS command, refer to Administrator's Reference.

Randomizing schedule start times

To randomize start times for schedules, means to scatter each schedule’s start time
across its startup window.

About this task

A startup window is defined by the start time and duration during which a
schedule must be initiated. For example, if the start time is 1:00 a.m. and the
duration is 4 hours, the startup window is 1:00 a.m. to 5:00 a.m. For the
client-polling scheduling mode, specify the percentage of the startup window that
the server can use to randomize start times for different client nodes that are
associated with a schedule.

If you set randomization to 0, no randomization occurs. This process can result in

communication errors if many client nodes try to contact the server at the same
instant.

The settings for randomization and the maximum percentage of scheduled sessions
can affect whether schedules are successfully completed for client nodes. Users
receive a message if all sessions are in use when they attempt to process a
schedule. If this happens, you can increase randomization and the percentage of
scheduled sessions that are allowed to make sure that the server can handle the
workload. The maximum percentage of randomization that is allowed is 50%. This
limit ensures that half of the startup window is available for trying again, the
scheduled commands that failed.

To set randomization to 50%, enter:

set randomize 50

It is possible, especially after a client node or the server has been restarted, that a
client node may not poll the server until after the beginning of the startup window
in which the next scheduled event is to start. In this case, the starting time is
randomized over the specified percentage of the remaining duration of the startup
window.

Consider the following situation:

v The schedule start time is 8:00 a.m. and its duration is 1 hour. Therefore the
startup window for the event is from 8:00 to 9:00 a.m.
v Ten client nodes are associated with the schedule.

580 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 v Randomization is set to 50%.
v Nine client nodes poll the server before 8:00 a.m.
v One client node does not poll the server until 8:30 a.m.

The result is that the nine client nodes that polled the server before the beginning
of the startup window are assigned randomly selected starting times between 8:00
and 8:30. The client node that polled at 8:30 receives a randomly selected starting
time that is between 8:30 and 8:45.

Increasing the length of the schedule startup window

Increasing the size of the startup window (by increasing the schedule’s duration)
can also affect whether a schedule completes successfully.

About this task

A larger startup window gives the client node more time to attempt initiation of a
session with the server.

Controlling how often client nodes contact the server

To control how often client nodes contact the server to perform a scheduled
operation, an administrator can set the frequency for certain events.

About this task

v How often nodes query the server
v The number of command retry attempts
v The amount of time between retry attempts

Users can also set these values in their client user options files. (Root users on
UNIX and Linux systems set the values in client system options files.) However,
user values are overridden by the values that the administrator specifies on the
server.

The communication paths from client node to server can vary widely with regard
to response time or the number of gateways. In such cases, you can choose not to
set these values so that users can tailor them for their own needs.
Related tasks:
“Setting how often clients query the server”
“Setting the number of command retry attempts” on page 582
“Setting the amount of time between retry attempts” on page 582

Setting how often clients query the server

When scheduling client nodes with client-polling scheduling, you can specify how
often the nodes query the server for a schedule. If nodes poll frequently for
schedules, changes to scheduling information (through administrator commands)
are propagated more quickly to the nodes. However, increased polling by client
nodes also increases network traffic.

About this task

For the client-polling scheduling mode, you can specify the maximum number of
hours that the scheduler on a client node waits between attempts to contact the
server to obtain a schedule. You can set this period to correspond to the frequency
with which the schedule changes are being made. If client nodes poll more
frequently for schedules, changes to scheduling information (through administrator
commands) are propagated more quickly to client nodes.

Chapter 16. Managing schedules for client nodes 581

 If you want to have all clients using polling mode contact the server every 24
hours, enter:
set queryschedperiod 24

This setting has no effect on clients that use the server-prompted scheduling mode.

The clients also have a QUERYSCHEDPERIOD option that can be set on each
client. The server value overrides the client value once the client successfully
contacts the server.

Setting the number of command retry attempts

You can specify the maximum number of times the scheduler on a client node can
retry a scheduled command that fails.

About this task

The maximum number of command retry attempts does not limit the number of
times that the client node can contact the server to obtain a schedule. The client
node never gives up when trying to query the server for the next schedule.

Be sure not to specify so many retry attempts that the total retry time is longer
than the average startup window.

If you want to have all client schedulers retry a failed attempt to process a
scheduled command up to two times, enter:
set maxcmdretries 2

Maximum command retries can also be set on each client with a client option,
MAXCMDRETRIES. The server value overrides the client value once the client
successfully contacts the server.

Setting the amount of time between retry attempts

You can specify the length of time that the scheduler waits between command
retry attempts. Command retry attempts occur when a client node is unsuccessful
in establishing a session with the server or when a scheduled command fails to
process.

About this task

Typically, this setting is effective when set to half of the estimated time it takes to
process an average schedule. If you want to have the client scheduler retry every
15 minutes any failed attempts to either contact the server or process scheduled
commands, enter:
set retryperiod 15

You can use this setting in conjunction with the SET MAXCMDRETRIES command
(number of command retry attempts) to control when a client node contacts the
server to process a failed command. See “Setting the number of command retry
attempts.”

The retry period can also be set on each client with a client option, RETRYPERIOD.
The server value overrides the client value once the client successfully contacts the
server.

582 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Specifying one-time actions for client nodes
You can use the DEFINE CLIENTACTION command to specify that one or more
client nodes perform a one-time action if the client schedulers are active.

About this task

If the scheduling mode is set to prompted, the client performs the action within 3
to 10 minutes. If the scheduling mode is set to polling, the client processes the
command at its prescribed time interval. The time interval is set by the
QUERYSCHEDPERIOD client option. The DEFINE CLIENTACTION command
causes Tivoli Storage Manager to automatically define a schedule and associate
client nodes with that schedule. With the schedule name provided, you can later
query or delete the schedule and associated nodes. The names of one-time client
action schedules can be identified by a special character followed by numerals, for
example @1.

The schedule name and association information is returned to the server console or
the administrative client with messages ANR2500I and ANR2510I.

For example, you can issue a DEFINE CLIENTACTION command that specifies an
incremental backup command for client node HERMIONE in domain
ENGPOLDOM:
define clientaction hermione domain=engpoldom action=incremental

Tivoli Storage Manager defines a schedule and associates client node HERMIONE
with the schedule. The server assigns the schedule priority 1, sets the period units
(PERUNITS) to ONETIME, and determines the number of days to keep the
schedule active based on the value set with SET CLIENTACTDURATION
command.

For a list of valid actions, see the DEFINE CLIENTACTION command in the
Administrator's Reference. You can optionally include the OPTIONS and OBJECTS
parameters.

Determining how long the one-time schedule remains active

You can determine how long schedules that were defined via DEFINE
CLIENTACTION commands remain active by using the SET
CLIENTACTDURATION command.

About this task

This SET CLIENTACTDURATION command allows you to specify the number of

days that schedules that were created with the DEFINE CLIENTACTION
command are active. These schedules are automatically removed from the database
whether the associated nodes have processed the schedule or not, after the
specified number of days. The following example specifies that schedules for client
actions be active for 3 days:
set clientactduration 3

If the duration of client actions is set to zero, the server sets the DURUNITS
parameter (duration units) as indefinite for schedules defined with DEFINE
CLIENTACTION command. The indefinite setting for DURUNITS means that the
schedules are not deleted from the database.

Chapter 16. Managing schedules for client nodes 583

584 IBM Tivoli Storage Manager for Windows: Administrator's Guide
Part 4. Maintaining the server
To help you maintain server operations, Tivoli Storage Manager allows you to
automate tasks that should occur regularly, monitor processes, and ensure the
availability and integrity of the database. In addition, Tivoli Storage Manager
provides the tools needed to set up and manage a network of servers, and to move
data from one server to another.

© Copyright IBM Corp. 1993, 2013 585

586 IBM Tivoli Storage Manager for Windows: Administrator's Guide
Chapter 17. Managing the storage environment with the
Operations Center
The Operations Center provides visibility to the key aspects of your Tivoli Storage
Manager environment, including clients, servers, policy, storage pools, and storage
devices. Through the Operations Center, you can identify potential issues in your
storage environment, monitor alerts, and access the Tivoli Storage Manager
command line.

Operations Center overview

The Operations Center includes an Overview page where you can view high-level
status information. From the Overview page, you can navigate to other pages that
provide more details.

The Operations Center also includes controls for accessing the command line (for
example, to issue server commands) and for updating certain configuration
settings. Table 50 on page 588 describes the primary controls in the Operations
Center.

The Overview page of the Operations Center includes the following sections:
TSM Clients
A client is considered to be at risk if a problem occurs during its most
recent scheduled backup, or during a backup that starts within a specified
time interval. The term backup also includes archive operations for Tivoli
Storage Manager for Enterprise Resource Planning clients.
Use the information that is shown in the TSM Clients section to check for
failed or missed backups and to determine whether backup schedules are
running as expected.
Alerts An alert represents one or more error messages that are issued by a Tivoli
Storage Manager server.
Use the information that is shown in the Alerts section to determine
whether any client-related or server-related errors require attention.
Network Traffic
Use the Network Traffic section to verify that the amount of data that was
recently sent to and from the Tivoli Storage Manager servers is within the
expected range.
TSM Servers
The status indicator shows whether the Tivoli Storage Manager servers that
are managed by the Operations Center are available to provide data
protection services to clients.
You can also view information about the following server components:
Inventory
The inventory consists of the server database and its active and
archive logs. Use the Inventory section to complete the following
tasks:
v Verify that the database file system has enough free space.

© Copyright IBM Corp. 1993, 2013 587

 v Verify that there is enough free space in the file systems that
contain the active log and archive log.
v Verify that database backups are running as expected.

Tip: Database backup information is not reflected in the status

summary, but you can click the Inventory section to open the
TSM Servers page, which shows the time of the last database
backup.
Storage Pools and Storage Devices
Use the Storage Pools and Storage Devices section to complete the
following tasks:
v Check the status of your storage pools to ensure that there is
enough space available.
v For DISK device classes, determine whether any volumes are
offline or have a read-only access state.
v For tape or shared FILE device classes, determine whether there
are any libraries, drives, or paths that are offline.
v For FILE device classes that are not shared, determine whether
any directories are offline or if there is not enough free space to
allocate scratch volumes.
Table 50. Primary controls in the Operations Center
Control How to access the control
Command line To open the command-line interface, hover over the globe icon

in the Operations Center menu bar, and click Command

Line.

In the command-line interface, you can issue commands to

manage Tivoli Storage Manager servers that are configured as
hub or spoke servers.
Settings To view or change global configuration settings, hover over the

globe icon in the Operations Center menu bar, and click

Settings.

For example, you can change the following settings from the
Settings page:
v The frequency at which status data is refreshed
v The duration that alerts remain active, inactive, or closed
v The conditions that indicate that clients are at risk

588 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 Table 50. Primary controls in the Operations Center (continued)
Control How to access the control
Help To view general help for the Operations Center, including

message help, hover over the help icon

?
in the Operations
Center menu bar, and click Documentation.

To view help for the page that is open, hover over the same
icon, and click the respective page title.

The general help includes information about how to use the

Operations Center, such as information about the following
subjects:
v Alert monitoring and associated tasks
v Managing clients, policy, servers, storage pools, and storage
devices
v Issuing commands from the Operations Center
command-line interface
Log out To disconnect from the Operations Center, hover over the
administrator name in the Operations Center menu bar, and
click Log Out.

Opening the Operations Center

The Overview page is the default initial view in the Operations Center. However,
in your web browser, you can bookmark the page that you want to open when
you log in to the Operations Center.

Procedure
1. In a web browser, enter the following address, where hostname represents the
name of the computer where the Operations Center is installed, and secure_port
represents the port number that the Operations Center uses for HTTPS
communication:
[Link]
For more information about this port number, see the Operations Center
installation checklist in the Installation Guide.
2. Log in, using an administrator ID that is registered on the Operations Center
hub server.
In the Overview page, you can view summary information for the following
items:
v Clients, including clients that are at risk of being unprotected
v Alerts
v Network traffic
v Servers, including the server database and logs
v Storage pools and devices, including capacity usage
You can view more details by clicking most of these items or by using the
Operations Center menu bar.

Monitoring from a mobile device: To remotely monitor the storage

environment, you can view the Overview page of the Operations Center in the
web browser of a mobile device. The Operations Center supports the Apple
Safari web browser on the iPad. Other mobile devices can also be used.

Chapter 17. Managing the storage environment with the Operations Center 589
Alert monitoring overview
Alert monitoring in Tivoli Storage Manager can help identify and track relevant
problems on the server.

An alert is a notification of a relevant problem on the Tivoli Storage Manager

server and is triggered by a Tivoli Storage Manager server message. You can define
which server messages trigger alerts, and only those messages are reported as
alerts in the Operations Center or in an email.

For a newly installed server or for an upgraded server without defined alerts, a
default set of messages is defined to trigger alerts. You can add messages to, or
remove messages from, the default set.

When you configure a Tivoli Storage Manager server as a hub or spoke server for
the Operations Center, alert monitoring is activated by default, but email
notification about alerts is not activated by default.

Alert monitoring characteristics that you might want to configure

You can configure alert monitoring and its characteristics. For example, you might
want to configure the following characteristics:
v Configure email notification for administrators about alerts.
v Change which messages trigger alerts (by updating the default set).
v Change how often alerts are updated in the server database or how often alerts
remain in a certain state.

Server commands for configuring alert monitoring

To configure alert monitoring, use the following server commands, which are
grouped according to the general configuration task to which they apply.
Query the settings for alert monitoring
v QUERY MONITORSETTINGS
Activate alert monitoring
v SET ALERTMONITOR
Send email alerts to administrators
v SET ALERTEMAIL
v SET ALERTEMAILFROMADDR
v SET ALERTEMAILSMTPHOST
v SET ALERTEMAILSMTPPORT
v SET ALERTSUMMARYTOADMINS
v REGISTER ADMIN
v UPDATE ADMIN
v QUERY ADMIN
Define or query which messages trigger alerts
v DEFINE ALERTTRIGGER
v UPDATE ALERTTRIGGER
v DELETE ALERTTRIGGER
v QUERY ALERTTRIGGER

590 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 Define how often alerts are updated in the server database
v SET ALERTUPDATEINTERVAL
Define how often alerts remain in a certain state
v SET ALERTACTIVEDURATION
v SET ALERTINACTIVEDURATION
v SET ALERTCLOSEDDURATION
Query or update the state of an alert
v QUERY ALERTSTATUS
v UPDATE ALERTSTATUS

Changing the Operations Center configuration

You might want to change the Operations Center configuration. Some
configuration changes require stopping and restarting the Operations Center web
server, or restarting the initial configuration wizard.

Changing the hub server

You can remove the hub server of the Operations Center, and configure another
server as the hub server.

Procedure
1. Restart the initial configuration wizard of the Operations Center.
2. Use the wizard to configure the Operations Center to connect to the new hub
server.

Removing a spoke server

You can remove a spoke server from the Operations Center. For example, you
might want to move the spoke server from one hub server to another hub server,
or you might want to decommission the spoke server.

Procedure
1. Remove the spoke server from the group of servers that is managed by the hub
server by completing the following steps:
a. From a Tivoli Storage Manager command line, issue the following
command to the hub server:
QUERY MONITORSETTINGS
b. From the output of the command, copy the name that is in the Monitored
Group field.
c. Issue the following command to the hub server:
DELETE GRPMEMBER group_name member_name
where group_name represents the name of the monitored group, and
member_name represents the name of the spoke server.
2. Optional: Disable alerting and monitoring on the spoke server by issuing the
following commands to the spoke server. Do not complete this step if you want
to move a spoke server from one hub server to another hub server.
SET STATUSMONITOR OFF
SET ALERTMONITOR OFF
3. Optional: Delete the spoke-server definition on the hub server by issuing the
following command to the hub server. Do not complete this step if the

Chapter 17. Managing the storage environment with the Operations Center 591
 spoke-server definition is used for other purposes, for example, enterprise
configuration, command routing, storing virtual volumes, or library
management.
DELETE SERVER server_name
where server_name represents the name of the spoke server.

Restoring the configuration to the preconfigured state

If certain problems occur, you might want to restore the Operations Center
configuration to the preconfigured state where the Tivoli Storage Manager servers
are not defined as hub or spoke servers.

Procedure
To restore the configuration, complete the following steps:
1. Stop the Operations Center web server.
2. Unconfigure the hub server by completing the following steps:
a. Issue the following commands on the hub server, where
IBM-OC-hub_server_name is the name of the administrator ID that was
created automatically when you initially configured the hub server:
SET MONITORINGADMIN ""
SET MONITOREDSERVERGROUP ""
SET STATUSMONITOR OFF
SET ALERTMONITOR OFF
REMOVE ADMIN IBM-OC-hub_server_name
b. Reset the password for the hub server by issuing the following command
on the hub server:
SET SERVERPASSWORD ""

Attention: Do not complete this step if the hub server is configured with
other servers for other purposes, such as library sharing, exporting and
importing of data, or node replication.
3. Unconfigure spoke servers by completing the following steps:
a. On the hub server, issue the following command for each spoke server,
where IBM-OC-hub_server_name is the name of the server group that was
created automatically when you configured the first spoke server, and
spoke_server_name is the server name that is set for the spoke server:
DELETE GRPMEMBER IBM-OC-hub_server_name spoke_server_name
b. Check whether any spoke servers remain as members of the server group
by issuing the following command:
QUERY SERVERGROUP IBM-OC-hub_server_name
c. After all spoke servers are deleted from the server group, issue the
following commands on the hub server:
DELETE SERVERGROUP IBM-OC-hub_server_name
SET MONITOREDSERVERGROUP ""
d. On each spoke server, issue the following commands:
REMOVE ADMIN IBM-OC-hub_server_name
SETOPT PUSHSTATUS NO
SET ALERTMONITOR OFF
SET STATUSMONITOR OFF
e. Delete the definition of the hub server by issuing the following command
on each spoke server:
DELETE SERVER hub_server_name

592 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 Attention: Do not complete this step if the definition is used for other
purposes, such as library sharing, exporting and importing of data, or node
replication.
f. Delete the definition of each spoke server by issuing the following command
on the hub server:
DELETE SERVER spoke_server_name

Attention: Do not complete this step if the definition is used for other
purposes, such as library sharing, exporting and importing of data, or node
replication.
4. Restore the default settings on each server by issuing the following commands:
SET STATUSREFRESHINTERVAL 5
SET STATUSRETENTION 48
SET ALERTUPDATEINTERVAL 10
SET ALERTACTIVEDURATION 480
SET ALERTINACTIVEDURATION 480
SET ALERTCLOSEDDURATION 60
SET STATUSATRISKINTERVAL TYPE=AP INTERVAL=24
SET STATUSATRISKINTERVAL TYPE=VM INTERVAL=24
SET STATUSATRISKINTERVAL TYPE=SY INTERVAL=24
SET STATUSSKIPASFAILURE YES TYPE=ALL
5. Restart the initial configuration wizard of the Operations Center.

Starting and stopping the web server

The web server of the Operations Center runs as a service and starts automatically.
You might need to stop and start the web server, for example, to make
configuration changes.

Procedure

Stop and start the web server.

v From the Services window, stop or start the Tivoli Storage Manager Operations
Center service.

Restarting the initial configuration wizard

You might need to restart the initial configuration wizard of the Operations Center,
for example, to make configuration changes.

Before you begin

To change the following settings, use the Settings page in the Operations Center
rather than restarting the initial configuration wizard:
v The frequency at which status data is refreshed
v The duration that alerts remain active, inactive, or closed
v The conditions that indicate that clients are at risk
The Operations Center help includes more information about how to change these
settings.

About this task

To restart the initial configuration wizard, you must delete a properties file that
includes information about the hub server connection. However, any alerting,
monitoring, at-risk, or multiserver settings that were configured for the hub server

Chapter 17. Managing the storage environment with the Operations Center 593
 are not deleted. These settings are used as the default settings in the configuration
wizard when the wizard restarts.

Procedure
1. Stop the Operations Center web server.
2. On the computer where the Operations Center is installed, go to the following
directory, where installation_dir represents the directory in which the
Operations Center is installed:
v installation_dir\ui\Liberty\usr\servers\guiServer
For example:
v c:\Program Files\Tivoli\TSM\ui\Liberty\usr\servers\guiServer
3. In the guiServer directory, delete the [Link] file.
4. Start the Operations Center web server.
5. Open the Operations Center.
6. Use the configuration wizard to reconfigure the Operations Center. Specify a
new password for the monitoring administrator ID.
7. On any spoke servers that were previously connected to the hub server, update
the password for the monitoring administrator ID by issuing the following
command from the Tivoli Storage Manager command-line interface:
UPDATE ADMIN IBM-OC-hub_server_name new_password

Restriction: Do not change any other settings for this administrator ID. After
you specify the initial password, this password is managed automatically by
the Operations Center.
Related tasks:
“Opening the Operations Center” on page 589

594 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Chapter 18. Managing server operations
Administrators can perform such server operations as licensing purchased features,
starting and halting the server, and monitoring server information.

About this task

When possible, avoid running simultaneous data movement or delete operations,

or running simultaneous storage pool data access operations. One or more of the
operations might terminate prematurely if the data involved in a data access
operation is also involved in a data movement operation.

Examples of a data access operation include restoring, retrieving, or recalling data

from a client. Data access operations also include the GENERATE BACKUPSET and
EXPORT commands, as well as storage pool backups. Examples of data movement
and delete operations include migration, reclamation, MOVE DATA, EXPIRE INVENTORY,
DELETE FILESPACE, and DELETE VOLUME when DISCARRDATA=YES.

Also, if a delete operation is processing data involved in either a data access or a

data movement operation, either of the operations might terminate prematurely.

See the following topics:

Tasks:
“Licensing IBM Tivoli Storage Manager”
“Starting the Tivoli Storage Manager server” on page 604
“Moving the Tivoli Storage Manager server to another system” on page 608
“Date and time on the server” on page 609
“Managing server processes” on page 609
“Preempting operations” on page 611
“Setting the server name” on page 613
“Add or update server options” on page 614
“Getting help on commands and error messages” on page 616

Licensing IBM Tivoli Storage Manager

There are tasks involved when licensing an IBM Tivoli Storage Manager system,
including registering, saving and auditing.

Task Required Privilege Class

System
Register licenses
Audit licenses
Display license information Any administrator

For current information about supported clients and devices, visit the IBM Tivoli
Storage Manager home page at [Link]
Overview/Software/Tivoli/Tivoli_Storage_Manager.

© Copyright IBM Corp. 1993, 2013 595

 The base IBM Tivoli Storage Manager feature includes the following support:
v An unlimited number of administrative clients.
v Enterprise Administration, which includes: command routing, enterprise
configuration, and enterprise logging (server-to-server).
v Server-to-server Virtual Volume capabilities (does not include database and
storage pool backup).
v Network Enabler (network connections for clients).

Registering licensed features

You must register a new license if you want to add support for any of the features
that are not already in your existing license agreement. Tivoli Storage Manager
uses a license file and the REGISTER LICENSE command to complete this task.

Licenses are stored in enrollment certificate files, which contain licensing

information for the server product. The enrollment certificate files are on the
installation media, and are placed on the server during installation. When
registered, the licenses are stored in a NODELOCK file within the current directory.

To register a license, you must issue the REGISTER LICENSE command. The
command registers new licenses for server components, including Tivoli Storage
Manager (base), Tivoli Storage Manager Extended Edition, and System Storage
Archive Manager. You must specify the name of the enrollment certificate file
containing the license to be registered when you issue the REGISTER LICENSE
command. To unregister licenses, erase the NODELOCK file found in the server
instance directory and reregister the licenses.

The file specification can contain a wildcard character (*). The following are
possible certificate file names:
[Link]
Registers IBM Tivoli Storage Manager base edition.
[Link]
Registers IBM Tivoli Storage Manager Extended Edition. This includes the
disaster recovery manager, large libraries, and NDMP.
[Link]
Registers the System Storage Archive Manager. This is required to enable
Data Retention Protection and Expiration and Deletion Suspension
(Deletion Hold).
*.lic Registers all IBM Tivoli Storage Manager licenses for server components.

Notes:
v You cannot register licenses for components that are licensed on the basis of
processors. For example, Tivoli Storage Manager for Mail, Tivoli Storage
Manager for Databases, Tivoli Storage Manager for Enterprise Resource
Planning, Tivoli Storage Manager for Hardware, and Tivoli Storage Manager for
Space Management.

596 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 Attention:
v Save the installation media that contains your enrollment certificate files. You
might need to register your licenses again for any of the following reasons:
– The server is corrupted.
– The server is moved to a different computer.
– The NODELOCK file is destroyed or corrupted. IBM Tivoli Storage Manager
stores license information in the NODELOCK file, which is located in the
directory from which the server is started.

Monitoring licenses
When license terms change (for example, a new license is specified for the server),
the server conducts an audit to determine if the current server configuration
conforms to the license terms. The server also periodically audits compliance with
license terms. The results of an audit are used to check and enforce license terms.

If 30 days have elapsed since the previous license audit, the administrator cannot
cancel the audit. If an IBM Tivoli Storage Manager system exceeds the terms of its
license agreement, one of the following occurs:
v The server issues a warning message indicating that it is not in compliance with
the licensing terms.
v If you are running in Try Buy mode, operations fail because the server is not
licensed for specific features.

You must contact your IBM Tivoli Storage Manager account representative to
modify your agreement.

An administrator can monitor license compliance by:

Auditing licenses
Use the AUDIT LICENSES command to compare the current configuration
with the current licenses.

Note: During a license audit, the server calculates, by node, the amount of
backup, archive, and space management storage in use. This calculation
can take a great deal of CPU time and can stall other server activity. Use
the AUDITSTORAGE server option to specify that storage is not to be
calculated as part of a license audit.
Displaying license information
Use the QUERY LICENSE command to display details of your current
licenses and determine licensing compliance.
Scheduling automatic license audits
Use the SET LICENSEAUDITPERIOD command to specify the number of
days between automatic audits.

Chapter 18. Managing server operations 597

 Role of processor value units in assessing licensing
requirements
You can obtain information about the number of client and server devices that are
managed by Tivoli Storage Manager, and the utilization of processor value units
(PVUs) by server devices. Use this data to assess licensing requirements for the
Tivoli Storage Manager system.

Important: The PVU calculations that are provided by Tivoli Storage Manager are
considered estimates and are not legally binding. The PVU information reported by
Tivoli Storage Manager is not considered an acceptable substitute for the IBM
License Metric Tool.

Collecting PVU information

Review the following procedure to understand how Tivoli Storage Manager

estimates PVUs. The procedure also describes the steps that you can take to make
the estimates more accurate.
1. When you install the Tivoli Storage Manager V6.3 server, or upgrade the server
from a previous version, Common Inventory Technology is installed
automatically. Common Inventory Technology is a software component that
collects the metrics that are used to calculate PVUs.
2. When Tivoli Storage Manager clients connect to the Tivoli Storage Manager
server, the clients send Common Inventory Technology information to the
Tivoli Storage Manager server. This information includes the number of client
and server devices, the number of physical processors, the processor vendor
and type, and related information. The server will estimate PVUs only for
nodes that are classified as server nodes. If you change the default classification
for a device, the classification that you specify is used for PVU estimation. For
more information about devices and nodes, see Device classification.

Tip: Common Inventory Technology is not installed automatically on client

nodes that are running on Linux operating systems. If you plan to obtain PVU
information from a client node that is running on a Linux operating system,
install Common Inventory Technology on the client node. For details, see the
topic about installing UNIX and Linux backup-archive clients in the IBM Tivoli
Storage Manager Backup-Archive Clients Installation and User's Guide.
3. The Tivoli Storage Manager server stores the information in the Tivoli Storage
Manager database.
4. The Tivoli Storage Manager server generates PVU estimates. The estimates are
based on the values that are stored in the Tivoli Storage Manager database and
on PVU values that are defined for each processor type in the
IBM_ProcessorValueUnitTable.xml file. The IBM_ProcessorValueUnitTable.xml
file is installed automatically during the Tivoli Storage Manager server
installation. You also can download the latest version of the file from an IBM
FTP site. For more information, see Estimating processor value units.
5. You can browse PVU information by issuing commands that provide PVU
estimates. For more information about obtaining and adjusting PVU estimates,
see Estimating processor value units.

The following graphic illustrates the PVU estimation process.

598 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 Common Inventory Tivoli Storage
Technology component Manager database

Metrics used to
calculate processor
value units (PVUs)

Processor descriptions PVU values for

and client/server processor types
Tivoli Storage classifications IBM PVU table
Manager clients Tivoli Storage
Manager server

Commands for PVU estimation

query pvuestimate
select * from
pvuestimate_details

Figure 77. PVU estimation overview

Device classification
For purposes of PVU calculation, you can classify devices, such as workstations
and servers, as client nodes, server nodes, or other. By default, devices are
classified as client or server:
Client Backup-archive clients that run on Microsoft Windows 7, Microsoft
Windows XP Professional, and Apple systems are classified as client
devices.
Server Backup-archive clients that run on all platforms except for Microsoft
Windows 7, Microsoft Windows XP Professional, and Apple systems are
classified as server devices. All other node types are also classified as
server devices. The server on which Tivoli Storage Manager is running is
classified as a server device.

You can change the node classification to reflect how the device is used in the
system. For example, if a node is classified as a server, but functions as a client,
you can reclassify it as a client. If a node is not used in the system, you can
reclassify it as other.

When you assign a classification, consider the services that are associated with the
device. For example, a Microsoft Windows XP Professional notebook might be a

Chapter 18. Managing server operations 599

 client device if it is distributed to a university staff member as a single-use
workstation. The same computer might be a server device if it is distributed to a
department administrator, who provides additional services such as file sharing on
the computer to multiple workstations.

In a Tivoli Storage Manager system, you can assign multiple client node names to
the same physical workstation. For example, a clustering solution can have several
node names that are defined in the Tivoli Storage Manager server environment to
provide protection if a failover occurs. Redundant node names, or node names that
manage data for physical workstations that no longer exist, should not be counted
for licensing purposes. In this case, you might classify the node as other by using
the UPDATE NODE command.

Limitations

The PVU calculations are estimates because the software cannot determine all of
the factors that are required for a final number. The following factors affect the
accuracy of the calculations:
v PVU estimates are provided only for Tivoli Storage Manager V6.3 server devices
that have established a connection with the Tivoli Storage Manager server since
the installation of or upgrade to Tivoli Storage Manager V6.3.
v The default classification of nodes is based on assumptions, as described in
“Device classification” on page 599.
v The PVU estimate might not reflect the actual number of processors or processor
cores in use.
v The PVU estimate might not reflect cluster configurations.
v The PVU estimate might not reflect virtualization, including VMware and AIX
LPAR and WPAR.
v Common Inventory Technology might not be able to identify some processors,
and some processors might not have corresponding entries in the PVU table.

Formula for PVU estimation

You must have the following information to calculate PVUs:
v The node classification. PVUs are calculated only for server devices.
v The number of processors on the node.
v The processor vendor, brand, type (number of cores), and model. This
information is used to look up the PVU value in the IBM PVU table.

Calculate the PVUs for a node by using the following formula:

PVUs = number of processors on the node * processor type (core count) * pvu value

To calculate the total PVUs, sum the PVUs for all nodes.

Related information
Table 51. Information about PVUs and licensing
Information type Location
IBM PVU table [Link]
tivoli_support/misc/CandO/PVUTable/
PVU calculator [Link]
howtobuy/passportadvantage/
valueunitcalculator/[Link]

600 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 Table 51. Information about PVUs and licensing (continued)
Information type Location
PVU frequently asked questions [Link]
passportadvantage/[Link]
PVU licensing for distributed software [Link]
passportadvantage/
pvu_licensing_for_customers.html
PVU terminology [Link]
passportadvantage/
pvu_terminology_for_customers.html

Estimating processor value units

To estimate the number of processor value units (PVUs) for your system, use the
QUERY PVUESTIMATE, SELECT * FROM PVUESTIMATE_DETAILS, or the QUERY NODE
command.

Before you begin

Before you begin, review the information about how PVUs are estimated and what
the limitations are. For more information, see “Role of processor value units in
assessing licensing requirements” on page 598. Tivoli Storage Manager offers
several options for viewing PVU information. Select the option that best meets
your needs. To export the PVU estimates to a spreadsheet, use the SELECT * FROM
PVUESTIMATE_DETAILS command.

Important: The PVU calculations that are provided by Tivoli Storage Manager are
considered estimates and are not legally binding.

Procedure

To estimate PVUs, complete the following steps:

1. To ensure that you have the latest IBM PVU table, go to the FTP site at
[Link]
2. Download the file for the latest table and rename it
IBM_ProcessorValueUnitTable.xml.
3. Copy the renamed file to the Tivoli Storage Manager server instance directory.
Alternatively, you can store the file in the server installation directory, where it
replaces the installed file. Then, restart the server. A version of the IBM PVU
table is also provided with the Tivoli Storage Manager installation.
4. Obtain the type of PVU information that you require:

Chapter 18. Managing server operations 601

 Option Description
Product-based information To obtain a PVU report that is categorized
by product, issue the QUERY PVUESTIMATE
command. For more information about the
QUERY PVUESTIMATE command, see the Tivoli
Storage Manager Administrator's Reference.
Remember: The reports that are produced
by the QUERY PVUESTIMATE and SQL SELECT *
FROM PVUESTIMATE_DETAIL commands
provide information about all client and
server devices that are managed by the
Tivoli Storage Manager server. However,
PVU estimates are provided only for the
V6.3 server devices that established a
connection with the Tivoli Storage Manager
server since the installation of or upgrade to
Tivoli Storage Manager V6.3.
Node-based information To obtain a PVU report that provides
detailed information at the node level, issue
the SELECT * FROM PVUESTIMATE_DETAILS
command. For more information about the
SELECT command, see the Tivoli Storage
Manager Administrator's Reference. After you
obtain data from the SELECT * FROM
PVUESTIMATE_DETAILS command output, you
can move the data to a spreadsheet. Then,
modify the data to gain a more accurate
estimate.

To export data to a spreadsheet, use the

-comma and > output redirection options of
the dsmadmc administrative command-line
client with the SELECT administrative
command. For example, complete the
following steps:
1. dsmadmc -comma select * from
pvuestimate_details >
[Link]
2. Import the resulting CSV file into a
spreadsheet program.
Complete information To view complete information, including
PVU-related fields that are provided by
Common Inventory Technology, such as
processor vendor, brand, type, model, and
processor count for selected nodes, specify
the DETAILED value for the FORMAT parameter.
For more information about the QUERY NODE
command, see the Tivoli Storage Manager
Administrator's Reference.

5. To obtain a more accurate PVU estimate, you might want to change the
classifications of nodes. To change node classifications, issue the UPDATE NODE
command. For more information about the UPDATE NODE command, see the
Tivoli Storage Manager Administrator's Reference.
6. To calculate the PVUs for a node, use the following formula: PVUs = number of
processors on the node * processor type (core count) * pvu value. To
calculate the total PVUs, sum the PVUs for all nodes. For more information
about the PVU estimation formula, see Formula for PVU estimation.

602 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 7. After you generate a PVU report, more analysis might include removing
redundancies, deleting obsolete information from the report, and accounting for
known systems that are not logged in to and connected to the server.

Tip: If you cannot obtain PVU information from a client node that is running
on a Linux operating system, ensure that Common Inventory Technology is
installed on that client node. After you install Common Inventory Technology,
obtain a new PVU estimate.

Collecting processor value unit information in a VMware host

environment
If you are installing Tivoli Storage Manager on a system that uses VMware
virtualization software, ensure that PVU information is collected in the virtual
environment. To collect PVU information in a virtual environment, you must have
VMware ESX 4 software installed on your system, and you must set up and run
hardware scans.

Procedure

Complete the following steps:

1. Obtain the Tivoli Storage Manager backup-archive client installation package
for the Linux x86 operating system.
For information about obtaining installation packages, see the Passport
Advantage website at [Link]
passportadvantage/[Link].
2. Install the Tivoli Storage Manager backup-archive client on a computer in your
system.
3. From the ../CIT/VMware/esx-4 directory, download the dispatcher,
[Link], [Link], and cpuid files.
4. Copy the downloaded files to the ESX 4 host operating system.

Tip: You can copy the files to any location on the host operating system, but
ensure that all files are copied to the same directory.
5. Ensure that guest virtual machines are running. This step is necessary to ensure
that the guest virtual machines are detected during the hardware scan.
6. To collect PVU information, issue the following command:
retrieve -v

What to do next

If you restart the host machine or change the configuration, run the retrieve
command again to ensure that current information is retrieved.

Tip: When the IBM Tivoli Storage Manager for Virtual Environments license file is
installed on a VMware vStorage backup server, the platform string that is stored
on the Tivoli Storage Manager server is set to TDP VMware for any node name
that is used on the server. The reason is that the server is licensed for Tivoli
Storage Manager for Virtual Environments. The TDP VMware platform string can
be used for PVU calculations. If a node is used to back up the server with standard
backup-archive client functions, such as file-level and image backup, interpret the
TDP VMware platform string as a backup-archive client for PVU calculations.

Chapter 18. Managing server operations 603

Starting the Tivoli Storage Manager server
You can select different methods for starting the Tivoli Storage Manager server,
depending on the operating system on which the server is installed.

To review the methods that are available if the server is installed on Windows, see
“Starting the server on Windows.”

Starting the server on Windows

You can start the Tivoli Storage Manager server as a Windows service, which is the
preferred method, or in the foreground.

About this task

On the Windows operating system, you must close all applications before you log
off. As a production server, Tivoli Storage Manager must be available to clients 24
hours a day. At many sites, it is a security exposure to leave an administrator ID
logged on at an unattended computer. The solution is to run the server as a
Windows service. When you run the server as a service, it can be configured to
start automatically when the system is restarted.

Alternatively, if you plan to configure the server or use it in a test environment,

you might want to start the server in the foreground. When you start the server in
the foreground, Tivoli Storage Manager provides a special administrator user ID
that is named SERVER_CONSOLE. All server messages are displayed directly on
the screen. The messages can be useful for debugging startup problems.

If you install a Tivoli Storage Manager server and run it in the foreground, you
must stop the server in the foreground before you start the server as a service.
Similarly, if you start the server as a Windows service, you must stop the server
before you can successfully start it in the foreground.

Procedure

Select one of the following methods for starting the server:

v To start the server as a Windows service, follow the instructions in “Starting the
server as a Windows service.”
v To start the server in the foreground, follow the instructions in “Starting the
server in the foreground” on page 606.

Starting the server as a Windows service

If you are running Tivoli Storage Manager on a Windows operating system, you
can start the server as a service.

Before you begin

You must configure the options for starting the server as a service. Complete the
following steps:
1. From the Windows Start menu, click Run, type [Link], and click OK.
2. In the Services window, select the server instance that you want to start as a
service, and click Properties. For example, select TSM INST1, and click
Properties.
3. To ensure that the server service starts automatically, click the General tab.
From the Startup type list, select Automatic.

604 IBM Tivoli Storage Manager for Windows: Administrator's Guide

4. To set the user for starting the server service, complete the following actions:
a. Click the Log On tab.
b. Select This account, and browse for the user ID that owns the server DB2
instance and has permissions for starting the server.
c. In the Select User window, in the Enter the object name to select field,
enter the user ID.
d. Click Check Names.
e. Click OK.

If the server service is configured to run under the Local System account, you must
explicitly grant the Local System account access to the Tivoli Storage Manager
database. To grant database access to the Local System account, complete the
following steps:
1. Log on with the user ID that was used to create the Tivoli Storage Manager
server database. This is the user ID that was used to run the DSMSERV FORMAT
utility to initialize the server database. Alternatively, if you configured the
server with the dsmicfgx configuration wizard, this is the user ID that was used
to create the instance.
2. Open a DB2 command window by taking one of the following actions:
v If the Tivoli Storage Manager server is installed on Windows Server 2008 or
Windows Server 2008 R2, click Start > All Programs > IBM DB2 DB2TSM1
> DB2 Command Window - Administrator.
v If the Tivoli Storage Manager server is installed on Windows Server 2012,
open the Start window, and click DB2 Command Window - Administrator.
3. In the DB2 command window, enter the following commands:
set DB2INSTANCE=server1
db2 connect to TSMDB1
db2 grant dbadm with dataaccess with accessctrl on database to user system
db2 grant secadm on database to user system

Tip: When the server service is configured to run under the Local System
account, the database can be accessed by any administrator on the system. In
addition, any administrator who can log on to the system can run the Tivoli
Storage Manager server.

Procedure

To start the server as a Windows service, complete the following steps:

1. Log on to the Tivoli Storage Manager server under an account in the
Administrators group.
2. From the Windows Start menu, click Run, type [Link], and click OK.
3. In the Services window, select the server instance that you want to start, and
click Start.

What to do next

Because the server service can issue requests that require action, it is important to
monitor server activity with the Operations Center or the administrative client.

To view start and stop completion messages that are logged in the Windows
application log, use the Event Viewer tool in the Administrative Tools folder.

Chapter 18. Managing server operations 605

 Starting the server in the foreground
To directly interact with a Tivoli Storage Manager server, start the server in the
foreground. For example, if you want to enter commands, start the server in the
foreground.

Procedure
1. Change to the directory where the server is installed. For example, change to
the c:\program files\tivoli\tsm\server directory.
2. Enter the following command:
dsmserv -k instance_name

where instance_name specifies the Tivoli Storage Manager server instance.

Running multiple server instances on a single system

To run multiple server instances on a single system, set up separate database and
recovery log directories for each server instance. Also, set up an instance directory
for each server instance to contain the server options file and other files that are
required to run the instance.

About this task

The server instances can be owned by the same or different user accounts.

The files for one instance of the server are stored separately from the files that are
used by another server instance on the same system, and separately from the
server program files. Database and recovery log files are stored in separate
directories, not in the server instance directory.

As part of server configuration, you create a directory to store the files for the
server instance. The following files are stored in the server instance directory:
v The server options file, [Link]
v The server key database file, [Link], and the .arm files, which are used by
clients and other servers to import the Secure Sockets Layer certificates of the
server
v The device configuration file, if the DEVCONFIG server option does not specify a
fully qualified name
v The volume history file, if the VOLUMEHISTORY server option does not specify a
fully qualified name
v The volumes for storage pools of the FILE device type, if the directory for the
device class is not fully specified, or not fully qualified
v The user exits
v The trace output, if it is not fully qualified
v The log file for first-failure data capture, [Link]

Procedure
1. Set up database and recovery log directories for each server instance.
2. Set up an instance directory for each server instance. For example, to run two
server instances, tsminst1 and tsminst2, create instance directories such as
/tsminst1 and /tsminst2. In each directory, place the [Link] file for that
server. Each [Link] file must specify a different port for the server to use.
Ensure that all required files are included in the instance directory.

606 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 3. To manage the system memory that is used by each server on a system, use the
DBMEMPERCENT server option. Limit the percentage of system memory that can be
used by the database manager of each server. If all servers are equally
important, use the same value for each server. If one server is a production
server and other servers are test servers, set the value for the production server
to a higher value than the test servers.
4. Optional: Set up the server instances to start automatically.
For instructions, see “Starting the server as a Windows service” on page 604.

Halting the server

You can halt the server without warning if an unplanned operating system
problem requires the server to be stopped.

Task Required Privilege Class

Start, halt, and restart the server System or operator

When you halt the server, all processes are abruptly stopped and client sessions are
canceled, even if they are not complete. Any in-progress transactions are rolled
back when the server is restarted. Administrator activity is not possible.

If possible, halt the server only after current administrative and client node
sessions are finished or canceled. To shut down the server without severely
impacting administrative and client node activity with the server, you must
complete the following steps:
1. Disable the server to prevent new client node sessions from starting by issuing
the DISABLE SESSIONS command. This command does not cancel sessions
currently in progress or system processes like migration and reclamation.
2. Notify any existing administrative and client node sessions that you plan to
shut down the server. The server does not provide a network notification
facility; you must use external means to notify users.
3. Cancel any existing administrative or client node sessions by issuing the
CANCEL SESSION command and the associated session number. To obtain
session numbers and determine whether any sessions are running, use the
QUERY SESSION command. If a session is running, a table is displayed
showing the session number on the leftmost of the screen.
4. Find out whether any other processes are running, such as server migration or
inventory expiration, by using the QUERY PROCESS command. If a database
backup process is running, allow it to complete before you halt the server. If
other types of processes are running, cancel them by using the CANCEL
PROCESS command.

Note: If the process you want to cancel is waiting for a tape volume to be
mounted (for example, a process that is initiated by EXPORT, IMPORT, or
MOVE DATA commands), the mount request is automatically canceled. If a
volume associated with the process is being mounted by an automated library,
the cancel might not take effect until the mount is complete.
5. Halt the server to shut down all server operations by using the HALT
command.

Note:
1. The HALT command can be replicated by using the ALIASHALT server option.
By using the server option, you can define a term other than HALT that

Chapter 18. Managing server operations 607

 completes the same function. The HALT command functions, however the
server option provides an extra method for issuing the HALT command.
2. In order for the administrative client to recognize an alias for the HALT
command, the client must be started with the CHECKALIASHALT option
specified. See the Administrator's Reference for more information.

Moving the Tivoli Storage Manager server to another system

You can move your Tivoli Storage Manager server from one computer to another.

Before you begin

The following prerequisites back up the database from one server and restore it to
another server.
v The same operating system must be running on both servers.
v The sequential storage pool that you use to back up the server database must be
accessible from both servers. Only manual and SCSI library types are supported
for the restore operation.
v The restore operation must be done by a Tivoli Storage Manager server at a code
level that is the same as the code level on the server that was backed up.

Procedure

To move the database to another system, complete the following steps:

1. Install Tivoli Storage Manager on the target server. Ensure that you install the
instance as if you were creating a new server. See the Installation Guide for
details.
2. Back up the database to sequential media. For example, issue the following
command:
backup db devclass=lto4 type=full

The sequential storage pool that you use to back up the server database must
be accessible from both servers.
3. Halt the server.
4. Move any libraries and devices from the original server to the new server, or
ensure that they are accessible through a storage area network.
5. Move copies of the volume history file, device configuration file, and server
options file to the target server.
6. Restore the backed up database on the target server. Ensure that you issue the
following commands as the instance user. For example:
v To maintain the current directory structure on the target server, issue this
command:
dsmserv restore db
v To change the current directory structure on the target server, create a file; for
[Link]. List the directories that must be restored on separate
lines. Issue this command:
dsmserv restore db on=[Link]
7. Start the target server.
Related tasks:
“Moving the database and recovery log on a server” on page 679

608 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Date and time on the server
The date and time on the server must be correct. If the server detects an invalid
date or time, server sessions become disabled.

Every time the server is started and for each hour thereafter, a date and time check
occurs. An invalid date can be one of the following:
v Earlier than the server installation date and time.
v More than one hour earlier than the last time the date was checked.
v More than 30 days later than the last time the date was checked.

An error message (ANR0110E) is displayed and expiration, migration, reclamation,

and volume history deletion operations are not allowed. You may either change
the system date if it is in error, or issue the ACCEPT DATE command to force the
server to accept the current system date as valid. Use the ENABLE SESSIONS
command after you issue the ACCEPT DATE command to re-enable the server for
client node activity.

Managing server processes

When a user or administrator issues a IBM Tivoli Storage Manager command or
uses a graphical user interface to perform an operation, the server starts a process.
Some examples of an operation are registering a client node, deleting a
management class, or canceling a client session.

Task Required Privilege Class

Display information about a server Any administrator
background process
Cancel a server process System

Most processes occur quickly and are run in the foreground, but others that take
longer to complete run as background processes.

The server runs the following operations as background processes:

v Auditing an automated library
v Auditing licenses
v Auditing a volume
v Backing up the database
v Backing up a storage pool
v Checking volumes in and out of an automated library
v Deleting a file space
v Deleting a storage volume
v Expiring the inventory
v Exporting or importing data
v Generating a backup set
v Migrating files from one storage pool to the next storage pool
v Moving data from a storage volume
v Reclaiming space from tape storage volumes
v Restoring a storage pool
v Restoring a volume

Chapter 18. Managing server operations 609

 Note: To prevent contention for the same tapes, the server does not allow a
reclamation process to start if a DELETE FILESPACE process is active. The server
checks every hour for whether the DELETE FILESPACE process has completed so
that the reclamation process can start. After the DELETE FILESPACE process has
completed, reclamation begins within one hour.

The server assigns each background process an ID number and displays the
process ID when the operation starts. This process ID number is used for tracking
purposes. For example, if you issue an EXPORT NODE command, the server
displays a message similar to the following:
EXPORT NODE started as Process 10

Some of these processes can also be run in the foreground by using the WAIT=YES
parameter when you issue the command from an administrative client. See
Administrator's Reference for details.

Requesting information about server processes

You can request information about server background processes. If you know the
process ID number, you can use the number to limit the search.

If you do not know the process ID, you can display information about all
background processes by entering:
query process

The following figure shows a server background process report after a DELETE
FILESPACE command was issued. The report displays a process ID number, a
description, and a completion status for each background process.

Process Process Description Status

Number
-------- ------------------------ ---------------------------------------------
2 DELETE FILESPACE Deleting filespace DRIVE_D for node CLIENT1:
172 files deleted.

Canceling server processes

You can cancel a server background process by specifying its ID number and
issuing the CANCEL PROCESS command.

About this task

To find the process number, issue the QUERY PROCESS command . For details,
see “Requesting information about server processes.”

When canceling processes, the following conditions apply:

v If a process is currently waiting for a tape volume to be mounted (for example, a
process initiated by EXPORT, IMPORT, or MOVE DATA commands), the mount
request is automatically canceled.
v If a volume associated with a process is currently being mounted by an
automated library, the cancel may not take effect until the mount is complete.

610 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 v If a process has a pending mount request, the process might not respond to a
CANCEL PROCESS command until the mount request has been answered,
cancelled, or timed out. (For example, reclamation automatically generates
mount requests as part of the process.)
To answer a mount request, issue the REPLY REQUEST command. To cancel a
mount request, issue the CANCEL REQUEST command.

Note:
1. To list open mount requests, issue the QUERY REQUEST command. You can
also query the activity log to determine if a given process has a pending
mount request.
2. A mount request indicates that a volume is needed for the current process.
However, the volume might not be available in the library. If the volume is
not available, the reason might be that you either issued the MOVE MEDIA
command or CHECKOUT LIBVOLUME command, or that you manually
removed the volume from the library.

Preempting operations
The server can preempt server or client operations for a higher priority operation
when a mount point is in use and no others are available, or access to a specific
volume is required. When an operation is preempted, it is canceled.

You can use the QUERY MOUNT command to see the status of the volume for the
mount point.

By default, preemption is enabled on the server. To disable preemption, specify

NOPREEMPT in the server options file. If you specify this option, the BACKUP DB
command, and the export and import commands are the only operations that can
preempt other operations.

Mount point preemption

If a high-priority operation requires a mount point that is in a specific device class
and all the mount points in the device class are in use, the high-priority operation
can preempt a mount point from a lower-priority operation.

Mount points can be preempted only when the device class of the operation
preempting and the operation that is being preempted is the same.

The following high-priority operations can preempt other operations for a mount
point.
v Back up database operations
v Retrieve, restore, or HSM recall operations that are initiated by clients
v Restore operations by using a remote data mover
v Export operations
v Import operations
v Backup set generation operations

The following server operations cannot preempt other operations or be preempted:

v Audit a volume
v Restore data from a copy or active-data pool
v Prepare a recovery plan file
v Store data by using a remote data mover

Chapter 18. Managing server operations 611

 The following operations can be preempted and are listed in order of priority, from
highest priority to lowest priority. The server selects the lowest priority operation
to preempt, for example, identify duplicates.
v Replicate nodes
v Back up to a copy storage pool
v Copy active data to an active data pool
v Move data on a storage pool volume
v Migrate data from disk to sequential media
v Migrate data from sequential media to sequential media
v Back up, archive, or HSM migrate operations that are initiated by clients
v Reclaims volumes in a sequential-access storage pool
v Identify duplicates

Volume access preemption

If a high-priority operation requires access to a specific volume and that volume is
in use, the high-priority operation can preempt the lower-priority operation for
that volume.

For example, if a restore request requires access to a volume in use by a

reclamation operation and a drive is available, the reclamation operation is
canceled.

The following high-priority operations can preempt operations for access to a

specific volume.
v Back up database operations
v Retrieve, restore, or HSM recall operations that are initiated by clients
v Restore operations by using a remote data mover
v Export operations
v Import operations
v Backup set generation operations

The following operations cannot preempt other operations or be preempted:

v Audit volume
v Restore data from a copy or active-data pool
v Prepare a recovery plan
v Store data by using a remote data mover

The following operations can be preempted, and are listed in order of priority,
from highest priority to lowest priority. The server selects the lowest priority
operation to preempt, for example, identify duplicates.
v Replicate nodes
v Back up to a copy storage pool
v Copy active data to an active data pool
v Move data on a storage pool volume
v Migrate data from disk to sequential media
v Migrate data from sequential media to sequential media
v Back up, archive, or HSM migrate data that is initiated by client
v Reclaim volumes in a sequential-access storage pool
v Identify duplicates

612 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Changing names in Tivoli Storage Manager
When you change the server name, it can affect communication from the server to
backup-archive nodes. When you change the host name for the server, it can affect
the connection between the database and the server.

Setting the server name

At installation, the server name is set to SERVER1 or on Windows, the machine
name. After installation, you can use the SET SERVERNAME command to change
the server name.

You can issue the QUERY STATUS command to see the name of the server.

To specify the server name you must have system privileges. For example, to
change the server name to WELLS_DESIGN_DEPT., enter the following:
set servername wells_design_dept.

You must set unique names on servers that communicate with each other. See
“Setting up communications among servers” on page 692 for details. On a network
where clients connect to multiple servers, it is preferable that all servers have
unique names.

Attention:
v If this is a source server for a virtual volume operation, changing its name can
impact its ability to access and manage the data it has stored on the
corresponding target server.
v To prevent problems related to volume ownership, do not change the name of a
server if it is a library client.

You can change the server name with the SET SERVERNAME command. But you
might have unfortunate results that, according to the platform, can vary. Some
examples to be aware of are:
v Passwords might be invalidated. For example, Windows clients use the server
name to identify which passwords belong to which servers. Changing the server
name after Windows backup-archive clients are connected forces clients to
re-enter the passwords.
v Device information might be affected.
v Registry information on Windows platforms might change.

Changing the host name for a Tivoli Storage Manager server

When you change the server host name, you can avoid errors at server startup by
ensuring that the database continues to function correctly.

Procedure

Use the following steps to change a host name when the Tivoli Storage Manager
server is installed.
1. Back up the Tivoli Storage Manager database.
2. Stop the Tivoli Storage Manager server.
3. Change the startup service of the Tivoli Storage Manager server to manual
startup:
a. In the Windows Services Management Console, select TSM Server service.

Chapter 18. Managing server operations 613

 b. Right-click the service and click Properties.
c. In the Startup Type field, select Manual.
4. Issue the following commands from the DB2 command prompt window to
update the DB2SYSTEM registry variable, turn off extended security, and
verify the settings:
db2set -g DB2SYSTEM=new_host_name
db2set -g DB2_EXTSECURITY=NO
db2set -all

Tip: The DB2_EXTSECURITY parameter is reset to YES when you restart the
system.
5. Check for the presence of the [Link] file. Depending on your version of
Windows, the [Link] file might be in one of the following directories:
v Windows 2008 or later:
C:\ProgramData\IBM\DB2\DB2TSM1\<DB2 Instance name>
v Other versions of Windows:
C:\Documents and Settings\All Users\Application Data\IBM\DB2\DB2TSM1\
<DB2 Instance name>

Tip: The [Link] file is a hidden file. Ensure that you show all files by
going to Windows Explorer and selecting Tools > Folder Options and
specifying to view hidden files.
If the [Link] file does not exist on your system, proceed to the next
step. If the file does exist, issue the following command to update the host
name:
db2nchg /n:0 /i:<instance> /h:<new host name>
6. Change the Windows host name, as described in the documentation for the
Windows system that you are using.
7. Restart the server.
8. Update the security settings by running the following command:
db2extsec -a new_host_name\DB2ADMNS -u new_host_name\DB2USERS
9. Start the Tivoli Storage Manager server.
10. Reset the startup service for Tivoli Storage Manager server to be automatic:
a. In the Windows Services Management Console, select TSM Server service.
b. Right-click the service and click Properties.
c. In the Startup Type field, select Automatic.

Add or update server options

If you have system privileges, you can add or update server options in the
[Link] file.

To add or update server options, use the SETOPT command to edit the [Link]
file.

For information about editing the server options file, refer to the Administrator's
Reference.

614 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 Adding or updating a server option without restarting the
server
A system administrator can add or update a limited number of server options
without stopping and restarting the server. The added or updated server option is
appended to the end of the server options file.

You can update existing server options by issuing the SETOPT command. For
example, to update the existing server option value for MAXSESSIONS to 20, enter:
setopt maxsessions 20

These server options can be added or updated:

v COMMTIMEOUT
v DATEFORMAT
v DNSLOOKUP
v EXPINTERVAL
v EXPQUIET
v IDLETIMEOUT
v MAXSESSIONS
v NUMBERFORMAT
v RESTOREINTERVAL
v TIMEFORMAT
v THROUGHPUTDATATHRESHOLD
v THROUGHPUTTIMETHRESHOLD

Note: SETOPT commands in a macro cannot be rolled back.

Deleting information about volume history

Periodically delete outdated information from the volume history file.

About this task

The contents of the volume history file are created by using the volume history
table in the server database. When opening a volume, the server might check the
table to determine whether the volume is already used. If the table is large, it can
take a long time to search. Other sessions or processes, such as backups and other
processes that use multiple sequential volumes, can be delayed due to locking.

For example, if you keep backups for seven days, information older than seven
days is not needed. If information about database backup volumes or export
volumes is deleted, the volumes return to scratch status. For scratch volumes of
device type FILE, the files are deleted. When information about storage pools
volumes is deleted, the volumes themselves are not affected.

Procedure

To delete volume history, issue the DELETE VOLHISTORY command. For example, to
delete volume history that is seven days old or older, issue the following
command:
delete volhistory type=all todate=today-8

Chapter 18. Managing server operations 615

 When deleting information about volume history, keep in mind the following
guidelines:
v Ensure that you delete volume history entries such as STGNEW, STGDELETE,
and STGREUSE that are older than the oldest database backup that is required
to perform a point-in-time database restore. If necessary, you can delete other
types of entries.
v Existing volume history files are not automatically updated with the DELETE
VOLHISTORY command.
v Do not delete information about sequential volumes until you no longer need
that information. For example, do not delete information about the reuse of
storage volumes unless you backed up the database after the time that was
specified for the delete operation.
v Do not delete the volume history for database backup or export volumes that
are stored in automated libraries unless you want to return the volumes to
scratch status. When the DELETE VOLHISTORY command removes information for
such volumes, the volumes automatically return to scratch status. The volumes
are then available for reuse by the server and the information stored on them
can be overwritten.
v To ensure that you have a backup from which to recover, you cannot remove the
most current database snapshot entry by deleting volume history. Even if a more
current, standard database backup exists, the latest database snapshot is not
deleted.
v To display volume history, issue the QUERY VOLHISTORY command. For example,
to display volume history up to yesterday, issue the following command:
query volhistory enddate=today-1

Results
DRM: DRM automatically expires database backup series and deletes the volume history
entries.

Getting help on commands and error messages

Any administrator can issue the HELP command to display information about
administrative commands and messages from the server and the administrative
command-line client.

About this task

You can issue the HELP command with no operands to display a menu of help
selections. You also can issue the HELP command with operands that specify help
menu numbers, commands, or message numbers.

To display the help menu, enter:

help

To display help information on the REMOVE commands, enter:

help remove

To display help information on a specific message, such as ANR0992I for example,

enter:
help 0992

616 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Additional information is also available in the online documentation.

Chapter 18. Managing server operations 617

618 IBM Tivoli Storage Manager for Windows: Administrator's Guide
Chapter 19. Automating server operations
You can schedule administrative commands to tune server operations and to start
functions that require significant server or system resources during times of low
usage. Automating these operations allows the administrator to ensure that server
resources are available when needed by clients.

About this task

An administrator can automate the process of issuing a sequence of commands by

storing the commands in a server script. From the command line, the administrator
can immediately process the script or schedule the script for processing.

Tivoli Storage Manager includes a central scheduling component that allows the
automatic processing of administrative commands during a specific time period
when the schedule is activated. Schedules that are started by the scheduler can run
in parallel. You can process scheduled commands sequentially by using scripts that
contain a sequence of commands with WAIT=YES. You can also use a scheduler
external to invoke the administrative client to start one or more administrative
commands.

Each scheduled administrative command is called an event. The server tracks and
records each scheduled event in the database. You can delete event records as
needed to recover database space.

Scripts can be scheduled. For example:

Procedure
1. Define a schedule named EXPPROC that invokes expiration processing every
night at 2:00 a.m. For example:
define schedule expproc type=administrative -
cmd=’expire inventory’ active=yes starttime=02:00

This schedule calls for a schedule window that:

v Begins on the date the schedule is defined (the default) at 2:00 a.m.
v Lasts for 1 hour (the default)
v Is repeated every day
v Takes effect immediately
v Stays in effect indefinitely (the default)
2. Because the EXPPROC schedule is to run daily, you can verify that the
automation is working as it should on the day after you define the schedule.
For example:
query event expproc type=administrative begindate=today-1

If the schedule ran successfully, the status is Completed.

© Copyright IBM Corp. 1993, 2013 619

 What to do next

See the following topics for more information:

Concepts:
“Automating a basic administrative command schedule”
“Tailoring schedules” on page 621
“Copying schedules” on page 624
“Deleting schedules” on page 625
“Managing scheduled event records” on page 625
“Tivoli Storage Manager server scripts” on page 626
“Using macros” on page 637

Automating a basic administrative command schedule

You can set up a basic administrative command schedule by using Tivoli Storage
Manager defaults.

Notes:
1. Scheduled administrative command output is directed to the activity log. This
output cannot be redirected. For information about the length of time activity
log information is retained in the database, see “Using the Tivoli Storage
Manager activity log” on page 799.
2. You cannot schedule MACRO or QUERY ACTLOG commands.

To later update or tailor your schedules, see “Tailoring schedules” on page 621.

Task Required Privilege Class

Define, update, copy, or delete System
administrative schedules
Display information about scheduled Any administrator
operations

Defining the schedule

You can use the DEFINE SCHEDULE command to create a new schedule for
processing an administrative command.

About this task

Include the following parameters when you define a schedule with the DEFINE
SCHEDULE command:
v Specify the administrative command to be issued (CMD= ).
v Specify whether the schedule is activated (ACTIVE= ).

Issue the following example command:

define schedule backup_archivepool type=administrative
cmd=’backup stgpool archivepool recoverypool’ active=yes

To receive the following results:

v The schedule created is BACKUP_ARCHIVEPOOL.

620 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 v The schedule is to process the administrative command:
backup stgpool archivepool recoverypool
This command specifies that primary storage pool ARCHIVEPOOL is backed up
to the copy storage pool RECOVERYPOOL.
v The schedule is currently active.
v Administrative command output is redirected to the activity log.
v The following defaults are in effect:
– The start date and time defaults to the current date and time.
– The length of the startup window is 1 hour.
– The priority for the schedule is 5. If schedules conflict, the schedule with the
highest priority (lowest number) is run first.
– The schedule never expires.

To change the defaults, see “Tailoring schedules.”

Verifying the schedule

You can verify the details of what you have scheduled by using the QUERY
SCHEDULE command. When you use the QUERY SCHEDULE command, you
must specify the TYPE=ADMINISTRATIVE parameter to view an administrative
command schedule.

About this task

The following figure shows an example of a report that is displayed after you
enter:
query schedule backup_archivepool type=administrative

* Schedule Name Start Date/Time Duration Period Day

- ---------------- -------------------- -------- ------ ---
BACKUP_ARCHIVE- 09/04/2002 [Link] 1 H 1 D Any
POOL

Note: The asterisk (*) in the first column specifies whether the corresponding
schedule has expired. If there is an asterisk in this column, the schedule has
expired.

You can check when the schedule is projected to run and whether it ran
successfully by using the QUERY EVENT command. For information about
querying events, see “Querying events” on page 625.

Tailoring schedules
To control your schedules more precisely, specify values for the schedule
parameters instead of accepting the default settings when you define or update
schedules.

You can specify the following values when you issue the DEFINE SCHEDULE or
UPDATE SCHEDULE command:
Schedule name
All schedules must have a unique name, which can be up to 30 characters.

Chapter 19. Automating server operations 621

 Schedule style
You can specify either classic or enhanced scheduling. With classic
scheduling, you can define the interval between the startup windows for a
schedule. With enhanced scheduling, you can choose the days of the week,
days of the month, weeks of the month, and months the startup window
can begin on.
Initial start date, initial start time, and start day
You can specify a past date, the current date, or a future date for the initial
start date for a schedule with the STARTDATE parameter.
You can specify a start time, such as 6 p.m. with the STARTTIME parameter.
For classic scheduling, use the DAYOFWEEK parameter to specify that the
startup window begins on a certain day of the week, over the weekend,
during the week, or on any day. If the start date and time specified fall on
a day that does not correspond to your value for the day of the week, the
start date and time are shifted forward in 24-hour increments until the day
of the week is satisfied. If you select a value for the day of the week other
than ANY, schedules may not process when you expect. This depends on
the values for PERIOD and PERUNITS. Issue the QUERY EVENT command to
project when schedules will process to ensure that you achieve the desired
result.
For enhanced scheduling, you can use the DAYOFWEEK parameter to specify
that the startup window begin on one or more days of the week, over the
weekend, during the week, or on any day. MONTH, DAYOFMONTH, and
WEEKOFMONTH can also determine a start date. When used with the
DAYOFMONTH parameter, DAYOFWEEK must have a value of ANY. If
DAYOFWEEK=WEEKDAY or DAYOFWEEK=WEEKEND, you must specify
a value of either FIRST or LAST for the parameter WEEKOFMONTH.
Duration of a startup window
You can specify the duration of a startup window, such as 12 hours, with
the DURATION and DURUNITS parameters. The server must start the
scheduled service within the specified duration, but does not necessarily
complete it within that period of time. If the schedule needs to be retried
for any reason, the retry attempt must begin before the startup window
elapses or the operation does not restart.
If the schedule does not start during the startup window, the server
records this as a missed event in the database. You can get an exception
report from the server to identify schedules that did not run. For more
information, see “Querying events” on page 625.
How often to run the scheduled service
With classic scheduling, you can set the schedule frequency based on a
period of hours, days, weeks, months, or years with the PERIOD and
PERUNITS parameters. To have weekly backups, for example, set the period
to one week with PERIOD=1 and PERUNITS=WEEKS.
With enhanced scheduling specified, you can set your schedule to run on
certain months, days of the month, days of the week, and weeks of the
month with the MONTH, DAYOFMONTH, DAYOFWEEK, and WEEKOFMONTH parameters,
respectively. For example, if you want your schedule to run on the first
and last day of January and June, specify the months of January and June
and choose the first and last days of the month with
MONTH=JANUARY,JUNE and DAYOFMONTH=1,-1. If you want your

622 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 schedule to run during the last week of November, for example, choose the
last week of the month and November with MONTH=NOVEMBER and
WEEKOFMONTH=LAST.
Expiration date
You can specify an expiration date for a schedule with the EXPIRATION
parameter if the services it initiates are required for only a specific period
of time. If you set an expiration date, the schedule is not used after that
date, but it still exists. You must delete the schedule to remove it from the
database.
Priority
You can assign a priority to schedules with the PRIORITY parameter. For
example, if you define two schedules and they have the same startup
window or windows overlap, the server runs the schedule with the highest
priority first. A schedule with a priority of 1 is started before a schedule
with a priority of 3.
If two schedules try to use the same resources, the schedule that first
initiated the process will be the one to continue processing. The second
schedule will start but will not successfully complete. Be sure to check the
activity log for details.
Administrative schedule name
If you are defining or updating an administrative command schedule, you
must specify the schedule name.
Type of schedule
If you are updating an administrative command schedule, you must
specify TYPE=ADMINISTRATIVE on the UPDATE command. If you are
defining a new administrative command schedule, this parameter is
assumed if the CMD parameter is specified.
Command
When you define an administrative command schedule, you must specify
the complete command that is processed with the schedule with the CMD
parameter. These commands are used to tune server operations or to start
functions that require significant server or system resources. The functions
include:
v Migration
v Reclamation
v Export and import
v Database backup
Whether or not the schedule is active
Administrative command schedules can be active or inactive when they
are defined or updated. Active schedules are processed when the specified
command window occurs. Inactive schedules are not processed until they
are made active by an UPDATE SCHEDULE command with the ACTIVE
parameter set to YES.

Chapter 19. Automating server operations 623

 Using classic and enhanced command schedules
Depending on what type of event you want you schedule, and how often, you can
schedule commands to run using classic or enhanced scheduling.

About this task

Classic Scheduling
To schedule the backup of the ARCHIVEPOOL primary storage pool
periodically, use classic scheduling. Enter the following command:
define schedule backup_archivepool type=administrative
cmd=’backup stgpool archivepool recoverypool’
active=yes startime=20:00 period=2

This command specifies that, starting today, the ARCHIVEPOOL primary

storage pool is to be backed up to the RECOVERYPOOL copy storage pool
every two days at 8 p.m.
To update the BACKUP_ARCHIVEPOOL schedule, enter:
update schedule backup_archivepool type=administrative
startime=20:00 period=3

Starting today, the BACKUP_ARCHIVEPOOL schedule begins the backup

every three days at 10 p.m.
Enhanced Scheduling
To schedule the backup of the CENTRALPOOL primary storage pool on
specific days of the month, use enhanced scheduling. Enter the following
command:
define schedule backup_centralpool type=administrative
cmd=’backup stgpool centralpool auxilarypool’
active=yes startime=22:00 schedstyle=enhanced dayofmonth=10,-1

This command specifies that the CENTRALPOOL primary storage pool is

to be backed up to the AUXILARYPOOL copy storage pool on the tenth
and last day of each month at 10 p.m.

To update the BACKUP_CENTRALPOOL schedule, enter:

update schedule backup_centralpool type=administrative
startime=19:00 dayofmonth=-2

Starting today, the BACKUP_CENTRALPOOL schedule will begin the

backup on the second-to-last day of the month at 7 p.m.

Copying schedules
You can create a new schedule by copying an existing administrative schedule.

About this task

When you copy a schedule, Tivoli Storage Manager copies the following
information:
v A description of the schedule
v All parameter values from the original schedule

624 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 Procedure

To copy the BACKUP_ARCHIVEPOOL administrative schedule and name the new

schedule BCKSCHED, issue the following command:
copy schedule backup_archivepool bcksched type=administrative

Deleting schedules
You can delete the ENGBKUP administrative schedule.

About this task

Procedure

To delete the administrative schedule ENGBKUP, issue the following command:

delete schedule engbkup type=administrative

Managing scheduled event records

Each scheduled administrative command operation is called an event.

Task Required Privilege Class

Display information about events Any administrator
Set the retention period for event records System
Delete event records System or unrestricted policy

All scheduled events, including their status, are tracked by the server. An event
record is created in the server database whenever processing of a scheduled
command is created or missed.

Querying events
To help manage schedules for administrative commands, you can request
information about scheduled and completed events. You can request general or
exception reporting queries.

About this task

v To get information about past and projected scheduled processes, use a general
query. If the time range you specify includes the future, the query output shows
which events should occur in the future based on current schedules.
v To get information about scheduled processes that did not complete successfully,
use exception reporting.

To minimize the processing time when querying events, minimize the time range.

To query an event for an administrative command schedule, you must specify the
TYPE=ADMINISTRATIVE parameter. Figure 78 on page 626 shows an example of
the results of the following command:
query event * type=administrative

Chapter 19. Automating server operations 625

 Scheduled Start Actual Start Schedule Name Status
-------------------- -------------------- ------------- ---------
09/04/2002 [Link] 09/04/2002 [Link] BACKUP_ARCHI- Completed
VEPOOL

Figure 78. Query results for an administrative schedule

Removing event records from the database

You can specify how long event records stay in the database before the server
deletes them. You can also manually remove event records from the database.

If you issue a query for events, past events may display even if the event records
have been deleted. The events displayed with a status of Uncertain indicate that
complete information is not available because the event records have been deleted.
To determine if event records have been deleted, check the message that is issued
after the DELETE EVENT command is processed.

Setting the event record retention period

You can specify the retention period for event records in the database. After the
retention period passes, the server automatically removes the event records from
the database. At installation, the retention period is set to 10 days.

Event records are automatically removed from the database after both of the
following conditions are met:
v The specified retention period has passed
v The startup window for the event has elapsed

You can change the retention period from the default of 10 days by using the SET
EVENTRETENTION command.

Deleting event records

Because event records are deleted automatically, you do not have to manually
delete them from the database. However, you may want to manually delete event
records to increase available database space.

About this task

Use the DELETE EVENT command manually remove event records. For example,
to delete all event records written prior to 11:59 p.m. on June 30, 2002, enter:
delete event type=administrative 06/30/2002 23:59

Tivoli Storage Manager server scripts

Tivoli Storage Manager provides for automation of common administrative tasks
with server scripts that are stored in the database.

Tivoli Storage Manager provides sample scripts in:

v [Link]

The sample scripts have an example order of execution for scheduling

administrative commands. If one of the specified commands in the script does not
process successfully, the remaining commands are not processed. For more
information, see “Using SELECT commands in Tivoli Storage Manager scripts” on
page 797.

626 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 The administrator can schedule the script for processing using the administrative
command scheduler on the server.

Tivoli Storage Manager scripts can include the following:

v Command parameter substitution.
v SQL SELECT statements that you specify when the script is processed.
v Command execution control, such as PARALLEL and SERIAL processing
options.
v Conditional logic flow statements. These logic flow statements include:
– The IF clause; this clause determines how processing should proceed based
on the current return code value.
– The EXIT statement; this statement ends script processing.
– The GOTO and LABEL statement; this statement directs logic flow to
continue processing with the line that starts with the label specified.
– Comment lines.

Defining a server script

You can define a server script line-by-line, create a file that contains the command
lines, or copy an existing script.

About this task

Task Required Privilege Class
Define a server script System, policy, storage, and operator

You can define a script with the DEFINE SCRIPT command. You can initially
define the first line of the script with this command. For example:
define script qaixc "select node_name from nodes where platform=’aix’"
desc=’Display AIX clients’

This example defines the script as QAIXC. When you run the script, all AIX clients
are displayed.

To define additional lines, use the UPDATE SCRIPT command. For example, you
want to add a QUERY SESSION command, enter:
update script qaixc "query session *"

You can specify a WAIT parameter with the DEFINE CLIENTACTION command.
This allows the client action to complete before processing the next step in a
command script or macro. To determine where a problem is within a command in
a script, use the ISSUE MESSAGE command.

Restriction: You cannot redirect the output of a command within a Tivoli Storage
Manager script. Instead, run the script and then specify command redirection. For
example, to direct the output of script1 to the c:\temp\[Link] directory, run the
script and specify command redirection as in the following example:
run script1 > c:\temp\[Link]

Refer to Administrator's Reference for information on the DEFINE CLIENTACTION

and ISSUE MESSAGE commands.

Chapter 19. Automating server operations 627

 For additional information about updating server scripts, or updating a command
line, see “Updating a script” on page 633.

Defining a server script using contents of another file

You can define a script whose command lines are read in from another file that
contains statements for the script to be defined.

About this task

For example, to define a script whose command lines are read in from the file
[Link], issue:
define script admin1 file=[Link]

The script is defined as ADMIN1, and the contents of the script have been read in
from the file [Link].

Note: The file must reside on the server, and be read by the server.

Creating a maintenance script

You can create a predefined or a custom maintenance script. Either kind can help
you protect your data by running maintenance commands on a schedule.
Custom maintenance script
The custom maintenance script is either created using the maintenance
script editor or is converted using a predefined script. The script editor is
designed for experienced Tivoli Storage Manager users who require more
flexibility when constructing maintenance scripts.
Predefined maintenance script
The predefined maintenance script is created using a wizard.

You must schedule the maintenance script to run. The script typically includes
commands to back up, copy, and delete data. You can automate your server
maintenance by creating a maintenance script, and running it when your server is
not in heavy use.

Creating a custom maintenance script:

A custom maintenance script can be created using the maintenance script editor or
by converting a predefined maintenance script.

Before you begin

When you click Server Maintenance in the navigation tree, a list of servers is
displayed in the Maintenance Script table with either None, Custom, or
Predefined noted in the Maintenance Script column.

If you want to convert a predefined maintenance script into a custom maintenance

script, select a server that has a predefined script and click Select Action >
Convert to Custom Maintenance Script. Your predefined maintenance script
converts into a custom script and opens in the maintenance script editor. You
cannot convert a custom script into a predefined maintenance script.

Procedure

Complete the following steps to create a custom maintenance script using the
maintenance script editor.

628 IBM Tivoli Storage Manager for Windows: Administrator's Guide

1. Select a server.
2. Click Select Action > Create Custom Maintenance Script.
3. Click Select an Action and construct your maintenance script by adding a
command to the script. The following actions are available:
v Back Up Server Database
v Back Up Storage Pool
v Copy Active Data to Active-data Pool
v Create Recovery Plan File
v Insert Comment
v Delete Volume History
v Delete Expired Data
v Migrate Stored Data
v Move Disaster Recovery Media
v Run Script Commands in Parallel
v Run Script Commands Serially
v Reclaim Primary Storage Pool
v Reclaim Copy Storage Pool
You can change the order of commands by selecting a command and using the
up or down arrow to position it.
4. When you are satisfied with your script, define the schedule and click OK.

What to do next

To edit your custom script after it is created and saved, click Server Maintenance
in the navigation tree, select the server with the custom script and click Select
Action > Modify Maintenance Script. Your custom maintenance script opens in
the script editor where you can add, remove, or change the order of the
commands.

Creating a predefined maintenance script:

You can produce a predefined maintenance script using the maintenance script
wizard.

Before you begin

When you click Server Maintenance in the navigation tree, a list of servers is
displayed in the Maintenance Script table with either None, Custom, or
Predefined noted in the Maintenance Script column.

Procedure

Complete the following steps to create a maintenance script using the maintenance
script wizard.
1. Select a server that requires a maintenance script to be defined (None is
specified in the Maintenance Script column).
2. Click Select Action > Create Maintenance Script.
3. Follow the steps in the wizard.

Chapter 19. Automating server operations 629

 What to do next

After completing the steps in the wizard, you can convert your predefined
maintenance script into a custom maintenance script. If you choose to convert your
script into a custom script, select the server and click Select Action > Convert to
Custom Maintenance Script. Your predefined maintenance script is converted and
opened in the maintenance script editor where you can modify the schedule and
the maintenance actions.

Running commands in parallel or serially

You have the options of running commands in a script serially, in parallel, or
serially and in parallel by using the SERIAL or PARALLEL script commands in the
COMMAND_LINE parameter of DEFINE and UPDATE SCRIPT. You can run multiple
commands in parallel and wait for them to complete before proceeding to the next
command.

About this task

Running commands serially in a script ensures that any preceding commands are
complete before proceeding and ensures that any following commands are run
serially. When a script starts, all commands are run serially until a PARALLEL
command is encountered. Multiple commands running in parallel and accessing
common resources, such as tape drives, can run serially.

Script return codes remain the same before and after a PARALLEL command is run.
When a SERIAL command is encountered, the script return code is set to the
maximum return code from any previous commands run in parallel.

When using server commands that support the WAIT parameter after a PARALLEL
command, the behavior is as follows:
v If you specify (or use the default) WAIT=NO, a script does not wait for the
completion of the command when a subsequent SERIAL command is
encountered. The return code from that command reflects processing only up to
the point that the command starts a background process. The final return code
from the command is not available to your script.
v If you specify WAIT=YES, your script waits for the completion of the command
when a subsequent SERIAL command is encountered. The return code from that
command reflects processing for the entire command.

In most cases, you can use WAIT=YES on commands that are run in parallel.

The following example illustrates how the PARALLEL command is used to back up,
migrate, and reclaim storage pools.
/*run multiple commands in parallel and wait for
them to complete before proceeding*/
PARALLEL
/*back up four storage pools simultaneously*/
BACKUP STGPOOL PRIMPOOL1 COPYPOOL1 WAIT=YES
BACKUP STGPOOL PRIMPOOL2 COPYPOOL2 WAIT=YES
BACKUP STGPOOL PRIMPOOL3 COPYPOOL3 WAIT=YES
BACKUP STGPOOL PRIMPOOL4 COPYPOOL4 WAIT=YES
/*wait for all previous commands to finish*/
SERIAL
/*after the backups complete, migrate stgpools
simultaneously*/
PARALLEL
MIGRATE STGPOOL PRIMPOOL1 DURATION=90 WAIT=YES
MIGRATE STGPOOL PRIMPOOL2 DURATION=90 WAIT=YES

630 IBM Tivoli Storage Manager for Windows: Administrator's Guide

MIGRATE STGPOOL PRIMPOOL3 DURATION=90 WAIT=YES
MIGRATE STGPOOL PRIMPOOL4 DURATION=90 WAIT=YES
/*wait for all previous commands to finish*/
SERIAL
/*after migration completes, relcaim storage
pools simultaneously*/
PARALLEL
RECLAIM STGPOOL PRIMPOOL1 DURATION=120 WAIT=YES
RECLAIM STGPOOL PRIMPOOL2 DURATION=120 WAIT=YES
RECLAIM STGPOOL PRIMPOOL3 DURATION=120 WAIT=YES
RECLAIM STGPOOL PRIMPOOL4 DURATION=120 WAIT=YES

Using continuation characters for long commands

You can continue long commands across multiple command lines by specifying the
continuation character (-) as the last character for a command that is continued.

About this task

The following example continues an SQL statement across multiple command

lines:
/*-----------------------------*/
/* Sample continuation example */
SELECT-
* FROM-
NODE WHERE-
PLATFORM=’win32’

When this command is processed, it runs the following:

select * from nodes where platform=’win32’

Using substitution variables

You can include substitution variables in a script. Substitution variables are
specified with a $ character followed by a number that represents the position of
the parameter when the script is processed.

About this task

The following example SQLSAMPLE script specifies substitution variables $1 and

$2:
/*----------------------------------------------*/
/* Sample substitution example */
/* ---------------------------------------------*/
SELECT-
$1 FROM-
NODES WHERE-
PLATFORM=’$2’

When you run the script you must specify two values, one for $1 and one for $2.
For example:
run sqlsample node_name aix

The command that is processed when the SQLSAMPLE script is run is:
select node_name from nodes where platform=’aix’

Chapter 19. Automating server operations 631

 Using logic flow statements in a script
You can use conditional logic flow statements based on return codes issued from
previous command processing. These logic statements allow you to process your
scripts based on the outcome of certain commands. You can use IF, EXIT, or GOTO
(label) statements.

As each command is processed in a script, the return code is saved for possible
evaluation before the next command is processed. The return code can be one of
three severities: OK, WARNING, or ERROR. Refer to Administrator's Reference for a
list of valid return codes and severity levels.

Specifying the IF clause:

You can use the IF clause at the beginning of a command line to determine how
processing of the script should proceed based on the current return code value. In
the IF clause you specify a return code symbolic value or severity.

About this task

The server initially sets the return code at the beginning of the script to RC_OK.
The return code is updated by each processed command. If the current return code
from the processed command is equal to any of the return codes or severities in
the IF clause, the remainder of the line is processed. If the current return code is
not equal to one of the listed values, the line is skipped.

The following script example backs up the BACKUPPOOL storage pool only if
there are no sessions currently accessing the server. The backup proceeds only if a
return code of RC_NOTFOUND is received:
/* Backup storage pools if clients are not accessing the server */
select * from sessions
/* There are no sessions if rc_notfound is received */
if(rc_notfound) backup stg backuppool copypool

The following script example backs up the BACKUPPOOL storage pool if a return
code with a severity of warning is encountered:
/* Backup storage pools if clients are not accessing the server */
select * from sessions
/* There are no sessions if rc_notfound is received */
if(warning) backup stg backuppool copypool

Specifying the EXIT statement:

Use the EXIT statement to end script processing.

About this task

The following example uses the IF clause together with RC_OK to determine if
clients are accessing the server. If a RC_OK return code is received, this indicates
that client sessions are accessing the server. The script proceeds with the exit
statement, and the backup does not start.
/* Back up storage pools if clients are not accessing the server */
select * from sessions
/* There are sessions if rc_ok is received */
if(rc_ok) exit
backup stg backuppool copypool

632 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 Specifying the GOTO statement:

The GOTO statement is used in conjunction with a label statement. The label
statement is the target of the GOTO statement. The GOTO statement directs script
processing to the line that contains the label statement to resume processing from
that point.

About this task

The label statement always has a colon (:) after it and may be blank after the colon.
The following example uses the GOTO statement to back up the storage pool only
if there are no sessions currently accessing the server. In this example, the return
code of RC_OK indicates that clients are accessing the server. The GOTO statement
directs processing to the done: label which contains the EXIT statement that ends
the script processing:
/* Back up storage pools if clients are not accessing the server */
select * from sessions
/* There are sessions if rc_ok is received */
if(rc_ok) goto done
backup stg backuppool copypool
done:exit

Managing server scripts

You can update, copy, rename, query, delete, and run server scripts.

Task Required Privilege Class

Update, copy, rename, query, and delete a System, policy, storage, and operator
script
System, policy, storage, and operator
Run a script

Updating a script
You can update a script to change an existing command line or to add a new
command line to a script.

Appending a new command:

To append a command line to an existing script issue the UPDATE SCRIPT

command without the LINE= parameter. The appended command line is assigned
a line number of five greater than the last command line number in the command
line sequence. For example, if your script ends with line 010, the appended
command line is assigned a line number of 015.

About this task

The following is an example of the QSTATUS script. The script has lines 001, 005,
and 010 as follows:
001 /* This is the QSTATUS script */
005 QUERY STATUS
010 QUERY PROCESS

To append the QUERY SESSION command at the end of the script, issue the
following:
update script qstatus "query session"

Chapter 19. Automating server operations 633

 The QUERY SESSION command is assigned a command line number of 015 and
the updated script is as follows:
001 /* This is the QSTATUS script */
005 QUERY STATUS
010 QUERY PROCESS
015 QUERY SESSION

Replacing an existing command:

You can change an existing command line by specifying the LINE= parameter.

About this task

Line number 010 in the QSTATUS script contains a QUERY PROCESS command.
To replace the QUERY PROCESS command with the QUERY STGPOOL command,
specify the LINE= parameter as follows:
update script qstatus "query stgpool" line=10

The QSTATUS script is updated to the following:

001 /* This is the QSTATUS script */
005 QUERY STATUS
010 QUERY STGPOOL
015 QUERY SESSION

Adding a new command and line number:

You can change an existing script by adding new lines.

About this task

To add the SET REGISTRATION OPEN command as the new line 007 in the
QSTATUS script, issue the following:
update script qstatus "set registration open" line=7

The QSTATUS script is updated to the following:

001 /* This is the QSTATUS script */
005 QUERY STATUS
007 SET REGISTRATION OPEN
010 QUERY STGPOOL
015 QUERY SESSION

Copying a server script

You can copy an existing script to a new script with a different name.

About this task

For example, to copy the QSTATUS script to QUERY1 script, issue:

copy script qstatus query1

The QUERY1 command script now contains the same command lines as the
QSTATUS command script.

634 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Querying a server script
You can query a script to display information about the script. You can specify
wildcard characters to display all scripts with names that match a particular
pattern. When you query a script, you can direct the output to a file in a file
system that the server can access.

About this task

The various formats you can use to query scripts are as follows:

Format Description
Standard Displays the script name and description. This is the default.
Detailed Displays commands in the script and their line numbers, date of
last update, and update administrator for each command line in the
script.
Lines Displays the name of the script, the line numbers of the commands,
comment lines, and the commands.
File Outputs only the commands contained in the script without all
other attributes. You can use this format to direct the script to a file
so that it can be loaded into another server with the DEFINE script
command specifying the FILE= parameter.

To query a script in the standard format, issue the following:

query script *

The command gives results like the following:

Name Description
--------------- ------------------------------------------------------
QCOLS Display columns for a specified SQL table
QSAMPLE Sample SQL Query

For more information about querying a server script, refer to Administrator's

Reference.

Querying a server script to create another server script:

You can create additional server scripts by querying a script and specifying the
FORMAT=FILE and OUTPUTFILE parameters. You can use the resulting output as
input into another script without having to create a script line by line.

About this task

The following is an example of querying the SRTL2 script and directing the output
to [Link]:
query script srtl2 format=raw outputfile=[Link]

You can then edit the [Link] with an editor that is available to you on
your system. To create a new script using the edited output from your query, issue:
define script srtnew file=[Link]

Chapter 19. Automating server operations 635

 Renaming a server script
You can rename a script to a different name.

About this task

For example, to rename the QUERY1 script to QUERY5, issue:

rename script query1 query5

The QUERY1 script is now named QUERY5.

Deleting a command from a server script

You can delete an individual command line from a script. When you specify a line
number, only the corresponding command line is deleted from the script.

About this task

For example, to delete the 007 command line from the QSTATUS script, issue:
delete script qstatus line=7

Deleting a server script

To delete an entire script, issue the DELETE SCRIPT command.

About this task

For example, to delete the QSTATUS script, issue:

delete script qstatus

Running a server script

To process a script, issue the RUN command. You can run a script that contains
substitution variables by specifying them along with the RUN command.

About this task

Note: There is no Tivoli Storage Manager command that can cancel a script after it
starts. To stop a script, an administrator must halt the server.

You can preview the command lines of a script without actually executing the
commands by using the PREVIEW=YES parameter with the RUN command. If the
script contains substitution variables, the command lines are displayed with the
substituted variables. This is useful for evaluating a script before you run it.

For example, to process the QAIXC script previously defined, issue:

run qaixc

To process the following script that contains substitution variables:

/*----------------------------------------------*/
/* Sample continuation and substitution example */
/* ---------------------------------------------*/
SELECT-
$1 FROM-
NODES WHERE-
PLATFORM=’$2’

Enter:
run qaixc node_name aix

636 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 Where $1 is node_name and $2 is aix.

Using macros
Tivoli Storage Manager supports macros on the administrative client. A macro is a
file that contains one or more administrative client commands. You can only run a
macro from the administrative client in batch or interactive modes. Macros are
stored as a file on the administrative client. Macros are not distributed across
servers and cannot be scheduled on the server.

Macros can include the following:

v Administrative commands
For more information on administrative commands, see “Writing commands in a
macro.”
v Comments
For more information on comments, see “Writing comments in a macro” on page
638.
v Continuation characters
For more information on continuation characters, see “Using continuation
characters” on page 638.
v Variables
For more information on variables, see “Using substitution variables in a macro”
on page 639.

The name for a macro must follow the naming conventions of the administrative
client running on your operating system. For more information about file naming
conventions, refer to the Administrator's Reference.

In macros that contain several commands, use the COMMIT and ROLLBACK
commands to control command processing within the macro. For more information
about using these commands, see “Command processing in a macro” on page 640.

You can include the MACRO command within a macro file to invoke other macros
up to ten levels deep. A macro invoked from the Tivoli Storage Manager
administrative client command prompt is called a high-level macro. Any macros
invoked from within the high-level macro are called nested macros.

Writing commands in a macro

You can add commands to a macro.

About this task

The administrative client ignores any blank lines included in your macro.
However, a completely blank line terminates a command that is continued (with a
continuation character).

The following is an example of a macro called [Link] that registers and grants
authority to a new administrator:
register admin pease mypasswd -
contact=’david pease, x1234’
grant authority pease -
classes=policy,storage -
domains=domain1,domain2 -
stgpools=stgpool1,stgpool2

Chapter 19. Automating server operations 637

 This example uses continuation characters in the macro file. For more information
on continuation characters, see “Using continuation characters.”

After you create a macro file, you can update the information that it contains and
use it again. You can also copy the macro file, make changes to the copy, and then
run the copy. Refer to the Administrator's Reference for more information on how
commands are entered and the general rules for entering administrative
commands.

Writing comments in a macro

You can add comments to your macro file.

About this task

To write a comment:
v Write a slash and an asterisk (/*) to indicate the beginning of the comment.
v Write the comment.
v Write an asterisk and a slash (*/) to indicate the end of the comment.

You can put a comment on a line by itself, or you can put it on a line that contains
a command or part of a command.

For example, to use a comment to identify the purpose of a macro, write the
following:
/* [Link]-register new nodes */

Or, to write a comment to explain something about a command or part of a

command, write:
domain=domain1 /*assign node to domain1 */

Comments cannot be nested and cannot span lines. Every line of a comment must
contain the comment delimiters.

Using continuation characters

You can use continuation characters in a macro file. Continuation characters are
useful when you want to execute a command that is longer than your screen or
window width.

About this task

Attention: Without continuation characters, you can enter up to 256 characters.

With continuation characters, you can enter up to 1500 characters. In the MACRO
command, these maximums are after any substitution variables have been applied
(see “Using substitution variables in a macro” on page 639).

To use a continuation character, enter a dash or a back slash at the end of the line
that you want to continue. With continuation characters, you can do the following:
v Continue a command. For example:
register admin pease mypasswd -
contact="david, ext1234"

638 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 v Continue a list of values by entering a dash or a back slash, with no preceding
blank spaces, after the last comma of the list that you enter on the first line.
Then, enter the remaining items in the list on the next line with no preceding
blank spaces. For example:
stgpools=stg1,stg2,stg3,-
stg4,stg5,stg6
v Continue a string of values enclosed in quotation marks by entering the first
part of the string enclosed in quotation marks, followed by a dash or a back
slash at the end of the line. Then, enter the remainder of the string on the next
line enclosed in the same type of quotation marks. For example:
contact="david pease, bldg. 100, room 2b, san jose,"-
"ext. 1234, alternate contact-norm pass,ext 2345"
Tivoli Storage Manager concatenates the two strings with no intervening blanks.
You must use only this method to continue a quoted string of values across more
than one line.

Using substitution variables in a macro

You can use substitution variables in a macro to supply values for commands
when you run the macro. When you use substitution variables, you can use a
macro again and again, whenever you need to perform the same task for different
objects or with different parameter values.

About this task

A substitution variable consists of a percent sign (%), followed by a number that

indicates the number of the substitution variable. When you run the file with the
MACRO command, you must specify values for the variables.

For example, to create a macro named [Link] to register new nodes, write it
as follows:
/* register new nodes */
register node %1 %2 - /* userid password */
contact=%3 - /* ’name, phone number’ */
domain=%4 /* policy domain */

Then, when you run the macro, you enter the values you want to pass to the
server to process the command.

For example, to register the node named DAVID with a password of DAVIDPW,
with his name and phone number included as contact information, and assign him
to the DOMAIN1 policy domain, enter:
macro [Link] david davidpw "david pease, x1234" domain1

If your system uses the percent sign as a wildcard character, the administrative
client interprets a pattern-matching expression in a macro where the percent sign is
immediately followed by a numeric digit as a substitution variable.

You cannot enclose a substitution variable in quotation marks. However, a value

you supply as a substitution for the variable can be a quoted string.

Chapter 19. Automating server operations 639

 Running a macro
Use the MACRO command when you want to run a macro. You can enter the
MACRO command in batch or interactive mode.

About this task

If the macro does not contain substitution variables (such as the [Link] macro
described in the “Writing commands in a macro” on page 637), run the macro by
entering the MACRO command with the name of the macro file. For example:
macro [Link]

If the macro contains substitution variables (such as the [Link] macro

described in “Using substitution variables in a macro” on page 639), include the
values that you want to supply after the name of the macro. Each value is
delimited by a space. For example:
macro [Link] pease mypasswd "david pease, x1234" domain1

If you enter fewer values than there are substitution variables in the macro, the
administrative client replaces the remaining variables with null strings.

If you want to omit one or more values between values, enter a null string ("") for
each omitted value. For example, if you omit the contact information in the
previous example, you must enter:
macro [Link] pease mypasswd "" domain1

Command processing in a macro

When you issue a MACRO command, the server processes all commands in the
macro file in order, including commands contained in any nested macros. The
server commits all commands in a macro after successfully completing processing
for the highest-level macro.

If an error occurs in any command in the macro or in any nested macro, the server
terminates processing and rolls back any changes caused by all previous
commands.

If you specify the ITEMCOMMIT option when you enter the DSMADMC
command, the server commits each command in a script or a macro individually,
after successfully completing processing for each command. If an error occurs, the
server continues processing and only rolls back changes caused by the failed
command.

You can control precisely when commands are committed with the COMMIT
command. If an error occurs while processing the commands in a macro, the server
terminates processing of the macro and rolls back any uncommitted changes.
Uncommitted changes are commands that have been processed since the last
COMMIT. Make sure that your administrative client session is not running with the
ITEMCOMMIT option if you want to control command processing with the
COMMIT command.

Note: Commands that start background processes cannot be rolled back. For a list
of commands that can generate background processes, see “Managing server
processes” on page 609.

640 IBM Tivoli Storage Manager for Windows: Administrator's Guide

You can test a macro before implementing it by using the ROLLBACK command.
You can enter the commands (except the COMMIT command) you want to issue in
the macro, and enter ROLLBACK as the last command. Then, you can run the
macro to verify that all the commands process successfully. Any changes to the
database caused by the commands are rolled back by the ROLLBACK command
you have included at the end. Remember to remove the ROLLBACK command
before you make the macro available for actual use. Also, make sure your
administrative client session is not running with the ITEMCOMMIT option if you
want to control command processing with the ROLLBACK command.

If you have a series of commands that process successfully via the command line,
but are unsuccessful when issued within a macro, there are probably dependencies
between commands. It is possible that a command issued within a macro cannot
be processed successfully until a previous command that is issued within the same
macro is committed. Either of the following actions allow successful processing of
these commands within a macro:
v Insert a COMMIT command before the command dependent on a previous
command. For example, if COMMAND C is dependent upon COMMAND B,
you would insert a COMMIT command before COMMAND C. An example of
this macro is:
command a
command b
commit
command c/
v Start the administrative client session using the ITEMCOMMIT option. This
causes each command within a macro to be committed before the next command
is processed.

Chapter 19. Automating server operations 641

642 IBM Tivoli Storage Manager for Windows: Administrator's Guide
Chapter 20. Managing the database and recovery log
The IBM Tivoli Storage Manager database contains information that is needed for
server operations and information about client data that has been backed up,
archived, and space-managed. The recovery log contains information about
database updates that have not yet been committed.

About this task

The following sections provide detailed concept and task information about the
database and recovery log.

Concepts:
“Database and recovery log overview”

Tasks:
“Estimating database space requirements” on page 652
“Estimating recovery log space requirements” on page 656
“Monitoring the database and recovery log” on page 671
“Increasing the size of the database” on page 672
“Reducing the size of the database” on page 676
“Increasing the size of the active log” on page 678
“Backing up the server database” on page 870
“Restoring the database” on page 900
“Moving the database and recovery log on a server” on page 679
“Adding optional logs after server initialization” on page 684
“Transaction processing” on page 685

Database and recovery log overview

The Tivoli Storage Manager administrative interfaces work with the database and
recovery log. The skills of a database administrator are not required to manage
them.

Tivoli Storage Manager version 6.3 is installed with the IBM DB2 database
application. Users who are experienced DB2 administrators can choose to perform
advanced SQL queries and use DB2 tools to monitor the database. However, do
not use DB2 tools to change DB2 configuration settings from those settings that are
preset by Tivoli Storage Manager. Do not alter the DB2 environment for Tivoli
Storage Manager in other ways, such as with other products. The Tivoli Storage
Manager Version 6.3 server was built and tested with the data definition language
(DDL) and database configuration that Tivoli Storage Manager deploys.

© Copyright IBM Corp. 1993, 2013 643

 Attention: Making changes to the DDL or database configuration without using
Tivoli Storage Manager interfaces can adversely affect performance, damage or
destroy the server database, or cause data to become permanently lost.

Ensure that you do not do any of the following:

v Use database tools or interfaces other than those provided or documented by
Tivoli Storage Manager to change configuration settings from those that are set
by Tivoli Storage Manager at installation.
v Alter the DB2 environment in other ways. If you use database tools or interfaces
other than those provided or documented by Tivoli Storage Manager, you must
treat the server database as read-only.
v Use other interfaces to make changes to the server database.

Database: Overview
The database does not store client data; it points to the locations of the client files
in the storage pools. The Tivoli Storage Manager database contains information
about the Tivoli Storage Manager server. The database also contains information
about the data that is managed by the Tivoli Storage Manager server.

The database includes information about:

v Client nodes and administrators
v Policies and schedules
v Server settings
v Locations of client files on server storage
v Server operations (for example, activity logs and event records)
v Intermediate results for queries

The maximum size of the Tivoli Storage Manager database is 4 TB.

The database can be distributed across up to 128 directories. It is important that

the database is placed on fast, reliable disks that are configured for random access
I/O. Locating each directory on a different file system provides the best
performance because the data is striped across the directories. Enable read cache
for the database, and enable write cache if the disk subsystem supports it.

The database cannot be mirrored through Tivoli Storage Manager, but it can be
mirrored by using hardware mirroring, such as Redundant Array of Independent
Disks (RAID) 5.

Attention: If the database is unusable, the Tivoli Storage Manager server is

unavailable. You must backup the database to ensure that data that is managed by
the server can be recovered. Encrypt sensitive data by using the Tivoli Storage
Manager client or a storage device, unless the storage media is physically secured.
Security can be compromised even if data is not recovered. If a database is lost and
cannot be recovered, it might be difficult or impossible to recover data that is
managed by that server. Fragments of data or complete files might be read from
storage pool volumes that are not encrypted. See Part 6, “Protecting the server,” on
page 829 for steps that you can take to protect your database.

The database manager manages database volumes, and there is no need to format
them. Some advantages of the database manager are:
Automatic backups
When the server is started for the first time, a full backup begins

644 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 automatically. When the server is next started, the database manager
automatically backs up the database according to the following values set
by Tivoli Storage Manager:
v The active log space that is used since the last backup, which triggers a
full database backup
v The active log utilization ratio, which triggers an incremental database
backup
The database manager automatically backs up the database when the
archive log usage is 80% full.
Automatic statistics collection
Automatic statistics collection helps to improve database performance by
collecting up-to-date table statistics. The database manager determines
which statistics must be updated.
Automatic database reorganization
Reorganization of table data can be initiated by the server, or by DB2. If
server-initiated reorganization is enabled, based on table activity, the server
analyzes selected database tables and their indexes to determine when
reorganization is required. The database manager runs a reorganization
while server operations continue.

Attention: If reorganization by DB2 is enabled, DB2 controls the

reorganization process. Reorganization by DB2 is not recommended.
Multiple data streams for database backup and restore
Using a single data stream to back up databases of multiple terabytes can
take many hours. It can also affect the administrator’s ability to schedule
database backups effectively. The time to recover the server by using a
single data stream might not be enough to meet disaster recovery
objectives for the server. The Tivoli Storage Manager server provides a
multiple data stream capability for backups and restores.
SQL queries
The database makes more sophisticated SQL queries on the data possible.
To take advantage of these functions, you must use SQL to develop new
tools and create SQL statements.
Database audits
Database audits are run automatically, as needed, to ensure consistency. As
data is added to the server database, the database manager checks data
constraints and data types. Online integrity checks can prevent problems
for which offline audits had been needed in earlier releases.
Related tasks:
“Scheduling table and index reorganization” on page 676

Chapter 20. Managing the database and recovery log 645

 Connecting the server to the database with TCP/IP
The default configuration for the Tivoli Storage Manager V6.1 and V6.2 servers is
to use interprocess communications (IPC) to communicate with the database
manager. With Tivoli Storage Manager V6.3, the server can also connect to the
database manager by using TCP/IP.

Using TCP/IP to communicate with DB2 can greatly extend the number of
concurrent connections. The TCP/IP connection is part of the default configuration.
When the Tivoli Storage Manager V6.3 server is started for the first time, it
inspects the current configuration of the DB2 instance. It then makes any necessary
changes to ensure that both IPC and TCP/IP can be used to communicate with the
database manager. Any changes are made only as needed. For example, if the
TCP/IP node exists and has the correct configuration, it is not changed. If the node
was cataloged but has an incorrect IP address or port, it is deleted and replaced by
a node having the correct configuration.

When cataloging the remote database, the Tivoli Storage Manager server generates
a unique alias name based on the name of the local database. By default, a remote
database alias of TSMAL001 is created to go with the default database name of
TSMDB1.

Tip: Tivoli Storage Manager disables the TCP/IP connections if it cannot find an
alias in the range TSMAL001-TSMAL999 that is not already in use.

By default, the Tivoli Storage Manager server uses IPC to establish connections for
the first two connection pools, with a maximum of 480 connections for each pool.
After the first 960 connections are established, the Tivoli Storage Manager server
uses TCP/IP for any additional connections.

You can use the DBMTCPPORT server option to specify the port on which the TCP/IP
communication driver for the database manager waits for requests for client
sessions. The port number must be reserved for use by the database manager.

If Tivoli Storage Manager cannot connect to the database by using TCP/IP, it issues
an error message and halts. The administrator must determine the cause of the
problem and to correct it before restarting the server. The server verifies that it can
connect by using TCP/IP at startup even if it is configured to initially favor IPC
connections over TCP/IP connections.

Recovery log
The recovery log helps to ensure that a failure (such as a system power outage or
application error) does not leave the database in an inconsistent state. The recovery
log is essential when you restart the Tivoli Storage Manager or the database, and is
required if you must restore the database.

When you issue a command to make changes, the changes are committed to the
database to complete. A committed change is permanent and cannot be rolled
back. If a failure occurs, the changes that were made but not committed are rolled
back. Then all committed transactions, which might not have been physically
written to disk, are reapplied and committed again.

The recovery log consists of these logs:

v Active log
v Log mirror (optional)

646 IBM Tivoli Storage Manager for Windows: Administrator's Guide

v Archive log
v Archive failover log (optional)

During the installation process, you specify the directory location, the size of the
active log, and the location of the archive logs. You can also specify the directory
location of a log mirror if you want the additional protection of mirroring the
active log. The amount of space for the archive logs is not limited, which improves
the capacity of the server for concurrent operations compared to previous versions.

The space that you designate for the recovery log is managed automatically by the
database manager program. Space is used as needed, up to the capacity of the
defined log directories. You do not need to create and format volumes for the
recovery log.

Ensure that the recovery log has enough space. Monitor the space usage for the
recovery log to prevent problems.

Attention: To protect your data, locate the database directories and all the log
directories on separate physical disks.
Related concepts:
“Transaction processing” on page 685
“Active log”

Recovery log mode

The recovery log mode for the Tivoli Storage Manager is the roll-forward mode.
The recovery log stores data that is required to back up a restored database to the
most recent committed transaction. You must have the most recent recovery logs to
use the roll-forward mode.

Changes to the database are recorded in the recovery log to maintain a consistent
database image. You can restore the server to the latest time possible, by using the
active and archive log files, which are included in database backups.

To help ensure that the required log information is available for restoring the
database, you can specify that the active log is mirrored to another file system
location. For the best availability, locate the active log mirror on a different
physical device.

Active log
The active log files record transactions that are in progress on the server.

The active log stores all the transactions that have not yet been committed. The
active log always contains the most recent log records. If a failure occurs, the
changes that were made but not committed are rolled back, and all committed
transactions, which might not have been physically written to disk, are reapplied
and committed again.

The location and size of the active log are set during initial configuration of a new
or upgraded server. You can also set these values by specifying the
ACTIVELOGDIRECTORY and the ACTIVELOGSIZE parameters of the DSMSERV FORMAT or
DSMSERV LOADFORMAT utilities. Both the location and size can be changed later. To
change the size of the active log, see “Increasing the size of the active log” on page
678. To change the location of the active log directory, see “Moving only the active
log, archive log, or archive failover log” on page 681.

Chapter 20. Managing the database and recovery log 647

 For information about the space required for the active log directory, see “Active
and archive log space” on page 656.

Active log mirror

The active log mirror is a copy of the active log that can be used if the active log
files cannot be read. All changes made to the active log are also written to the log
mirror. There can be only one active log mirror.

Mirroring the active log can protect the database when a hardware failure occurs
on the device where the active log is stored. Mirroring the active log provides
another level of protection in addition to placing the active log on hardware that
has high-availability features. Creating a log mirror is optional but recommended.
Place the active log directory and the log mirror directory on different physical
devices. If you increase the size of the active log, the log mirror size is increased
automatically.

Mirroring the log can affect performance, because of the doubled I/O activity that
is required to maintain the mirror. The additional space that the log mirror requires
is another factor to consider.

You can create the log mirror during initial configuration of a new or upgraded
server. If you use the DSMSERV LOADFORMAT utility instead of the wizard to configure
the server, specify the MIRRORLOGDIRECTORY parameter. If the log mirror directory is
not created at that time, you can create it later by specifying the
MIRRORLOGDIRECTORY option in the server options file, [Link].

Archive log
The archive log contains copies of closed log files that had been in the active log.
The archive log is not needed for normal processing, but it is typically needed for
recovery of the database.

To provide roll-forward recovery of the database to the current point in time, all
logs since the last database backup must be available for the restore operation. The
archive log files are included in database backups and are used for roll-forward
recovery of the database to the current point-in-time. All logs since the last full
database backup must be available to the restore function. These log files are
stored in the archive log. The pruning of the archive log files is based on full
database backups. The archive log files that are included in a database backup are
automatically pruned after a full database backup cycle has been completed.

The archive log is not needed during normal processing, but it is typically needed
for recovery of the database. Archived log files are saved until they are included in
a full database backup. The amount of space for the archive log is not limited.

Archive log files are automatically deleted as part of the full backup processes and
must not be deleted manually. Monitor both the active and archive logs. If the
active log is close to filling, check the archive log. If the archive log is full or close
to full, run one or more full database backups.

If the file systems or drives where the archive log directory and the archive
failover log directory are located become full, the archived logs are stored in the
active log directory. Those archived logs are returned to the archive log directory
when the space problem is resolved, or when a full database backup is run.

648 IBM Tivoli Storage Manager for Windows: Administrator's Guide

You set the location of the archive log directory during initial configuration of a
new or upgraded server. You can also specify the ARCHLOGDIRECTORY parameter of
the DSMSERV FORMAT or DSMSERV LOADFORMAT utility. The location of the log can be
changed later.
Related concepts:
“Active and archive log space” on page 656
Related tasks:
“Moving only the active log, archive log, or archive failover log” on page 681

Archive failover log

The archive failover log, also called a secondary archive log, is the directory that
the server uses to store archive log files when the archive log directory is full. Its
use is optional but highly recommended.

Specifying an archive failover log directory can prevent problems that occur if the
archive log runs out of space. Place the archive log directory and the archive
failover log directory on different physical drives.

You can specify the location of the failover log directory during initial
configuration of a new or upgraded server. You can also specify its location with
the ARCHFAILOVERLOGDIRECTORY parameter of the DSMSERV FORMAT or DSMSERV
LOADFORMAT utility. If it is not created through the utilities, it can be created later by
specifying the ARCHFAILOVERLOGDIRECTORY option in the server options file,
[Link]. See “Adding optional logs after server initialization” on page 684 for
details.

For information about the space required for the log, see “Archive failover log
space” on page 669.

The role of the recovery log

When the logs that make up the recovery log are set up carefully, they work
together to ensure that data is not lost.

The active log files contain information about in-progress transactions. This
information is needed to restart the server and database after a disaster.
Transactions are stored in the log files of the active log, and a transaction can span
multiple log files.

When all transactions that are part of an active log file complete, that log file is
copied from the active log to the archive log. Transactions continue to be written to
the active log files while the completed active log files are copied to the archive
log. If a transaction spans all the active log files, and the files are filled before the
transaction is committed, the Tivoli Storage Manager server halts.

When an active log file is full, and there are no active transactions referring to it,
the file is copied to the archive log directory. An active log file cannot be deleted
until all transactions in the log file are either committed or discontinued.

If the archive log is full and there is no failover archive log, the log files remain in
the active log. If the active log then becomes full and there are in-progress
transactions, the Tivoli Storage Manager server halts. If there is an archive failover
log, it is used only if the archive log fills. It is important to monitor the archive log
directory to ensure that there is space in the active log.

Chapter 20. Managing the database and recovery log 649

 The Tivoli Storage Manager database manager can move active log files to the
failover archive log. The database manager automatically manages the space that is
available to the directories as database space. The database manager determines
when database backups are required and automatically initiates them.

When the database is backed up, the database manager deletes the archive log files
that are no longer needed for future database backups or restores.

The archive log is included in database backups and is used for roll-forward
recovery of the database. The archive log files that are included in a database
backup are automatically pruned after a full database backup cycle has completed.
Therefore, ensure that the archive log has enough space to store the log files for the
database backups.

Setting the user data limit for the database

You can set the user data limit and private memory for the database to ensure that
private memory is not exhausted. Exhausting private memory can result in errors,
less than optimal performance, and instability.

About this task

The user data limit that is displayed when you issue the ulimit -d command is the
soft user data limit. It is not necessary to set the hard user data limit for DB2. The
default soft user data limit is 128 MB. This is equivalent to the value of 262,144
512-byte units as set in /etc/security/limits folder, or 131,072 KB units as
displayed by the ulimit -d command. This setting limits private memory usage to
about one half of what is available in the 256 MB private memory segment
available for a 32-bit process on AIX.

Note: A DB2 server instance cannot make use of the Large Address Space or of
very large address space AIX 32-bit memory models due to shared memory
requirements. On some systems, for example those requiring large amounts of sort
memory for performance, it is best to increase the user data limit to allow DB2 to
allocate more than 128 MB of memory in a single process.

You can set the user data memory limit to "unlimited" (a value of "-1"). This setting
is not recommended for 32-bit DB2 because it allows the data region to overwrite
the stack, which grows downward from the top of the 256 MB private memory
segment. The result would typically be to cause the database to end abnormally. It
is, however, an acceptable setting for 64-bit DB2 because the data region and stack
are allocated in separate areas of the very large address space available to 64-bit
AIX processes.

Disk space requirements for the server database and recovery log
The drives or file systems on which you locate the database and log directories are
important to the proper operation of your IBM Tivoli Storage Manager server.
Placing each database and recovery log directory on a separate disk provides the
best performance and the best disaster protection.

For the optimal database performance, choose the fastest and most reliable disks
that are configured for random access I/O, such as Redundant Array of
Independent Disks (RAID) hardware. The internal disks included by default in

650 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 most servers and consumer grade Parallel Advanced Technology Attachment
(PATA) disks and Serial Advanced Technology Attachment (SATA) disks are too
slow.

To maintain database integrity, ensure that the storage hardware can withstand
failures such as power outages and controller failure. You can improve database
performance by using hardware that provides a fast, nonvolatile write cache for
both the database and logs. Put the database directories on fault tolerant storage
with high-availability features.

It is best to use multiple directories for the database, with four to eight directories
for a large Tivoli Storage Manager database. Locate each database directory on a
disk volume that uses separate physical disks from other database directories. The
Tivoli Storage Manager server database I/O workload is spread over all
directories, thus increasing the read and write I/O performance. Having many
small capacity physical disks is better than having a few large capacity physical
disks with the same rotation speed.

Locate the active log, mirror log, and archive log directories also on high-speed,
reliable disks. The failover archive log can be on slower disks, assuming that the
archive log is sufficiently large and that the failover log is used infrequently.

The access pattern for the active log is always sequential. Physical placement on
the disk is important. It is best to isolate the active log from the database and from
the disk storage pools. If they cannot be isolated, then place the active log with
storage pools and not with the database.

Enable read cache for the database and recovery log, and enable write cache if the
disk subsystems support it.

Restriction: You cannot use raw logical volumes for the database. To reuse space
on the disk where raw logical volumes were located for an earlier version of the
server, create file systems on the disk first.

Capacity planning
Capacity planning for Tivoli Storage Manager includes managing resources such as
the database and recovery log. To maximize resources as part of capacity planning,
you must estimate space requirements for the database and the recovery log.

Procedure

For information about the benefits of deduplication and guidance on how to make
effective use of the Tivoli Storage Manager deduplication feature, see Optimizing
Performance.

Chapter 20. Managing the database and recovery log 651

 Estimating database space requirements
To estimate space requirements for the database, you can use the maximum
number of files that can be in server storage at one time or you can use storage
pool capacity.

About this task

Consider using at least 25 GB for the initial database space. Provision file system
space appropriately. A database size of 25 GB is adequate for a test environment or
a library-manager-only environment. For a production server supporting client
workloads, the database size is expected to be larger. If you use random-access
disk (DISK) storage pools, more database and log storage space is needed than for
sequential-access storage pools.

The maximum size of the Tivoli Storage Manager database is 4 TB.

For information about sizing the database in a production environment that is

based on the number of files and on storage pool size, see the following topics.

Estimating database space requirements based on the number of

files
If you can estimate the maximum number of files that might be in server storage at
a time, you can use that number to estimate space requirements for the database.

About this task

To estimate space requirements that is based on the maximum number of files in

server storage, use the following guidelines:
v 600 - 1000 bytes for each stored version of a file, including image backups.

Restriction: The guideline does not include space that is used during data
deduplication.
v 100 - 200 bytes for each cached file, copy storage pool file, active-data pool file,
and deduplicated file.
v Additional space is required for database optimization to support varying
data-access patterns and to support server back-end processing of the data. The
amount of extra space is equal to 50% of the estimate for the total number of
bytes for file objects.

In the following example for a single client, the calculations are based on the
maximum values in the preceding guidelines. The examples do not take into
account that you might use file aggregation. In general, when you aggregate small
files, it reduces the amount of required database space. File aggregation does not
affect space-managed files.

Procedure
1. Calculate the number of file versions. Add each of the following values to
obtain the number of file versions:
a. Calculate the number of backed-up files. For example, as many as 500,000
client files might be backed up at a time. In this example, storage policies
are set to keep up to three copies of backed up files:
500,000 files * 3 copies = 1,500,000 files
b. Calculate the number of archive files. For example, as many as 100,000
client files might be archived copies.

652 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 c. Calculate the number of space-managed files. For example, as many as
200,000 client files might be migrated from client workstations.
Using 1000 bytes per file, the total amount of database space that is required
for the files that belong to the client is 1.8 GB:
(1,500,000 + 100,000 + 200,000) * 1000 = 1.8 GB
2. Calculate the number of cached files, copy storage-pool files, active-data pool
files, and deduplicated files:
a. Calculate the number of cached copies. For example, caching is enabled in a
5 GB disk storage pool. The high migration threshold of the pool is 90%
and the low migration threshold of the pool is 70%. Thus, 20% of the disk
pool, or 1 GB, is occupied by cached files.
If the average file size is about 10 KB, approximately 100,000 files are in
cache at any one time:
100,000 files * 200 bytes = 19 MB
b. Calculate the number of copy storage-pool files. All primary storage pools
are backed up to the copy storage pool:
(1,500,000 + 100,000 + 200,000) * 200 bytes = 343 MB
c. Calculate the number of active storage-pool files. All the active
client-backup data in primary storage pools is copied to the active-data
storage pool. Assume that 500,000 versions of the 1,500,000 backup files in
the primary storage pool are active:
500,000 * 200 bytes = 95 MB
d. Calculate the number of deduplicated files. Assume that a deduplicated
storage pool contains 50,000 files:
50,000 * 200 bytes = 10 MB
Based on the preceding calculations, about 0.5 GB of extra database space is
required for the client’s cached files, copy storage-pool files, active-data pool
files, and deduplicated files.
3. Calculate the amount of extra space that is required for database optimization.
To provide optimal data access and management by the server, extra database
space is required. The amount of extra database space is equal to 50% of the
total space requirements for file objects.
(1.8 + 0.5) * 50% = 1.2 GB
4. Calculate the total amount of database space that is required for the client. The
total is approximately 3.5 GB:
1.8 + 0.5 + 1.2 = 3.5 GB
5. Calculate the total amount of database space that is required for all clients. If
the client that was used in the preceding calculations is typical and you have
500 clients, for example, you can use the following calculation to estimate the
total amount of database space that is required for all clients:
500 * 3.5 = 1.7 TB

Results

Tip: In the preceding examples, the results are estimates. The actual size of the
database might differ from the estimate because of factors such as the number of
directories and the length of the path and file names. Periodically monitor your
database and adjust its size as necessary.

Chapter 20. Managing the database and recovery log 653

 What to do next

During normal operations, the Tivoli Storage Manager server might require
temporary database space. This space is needed for the following reasons:
v To hold the results of sorting or ordering that are not already being kept and
optimized in the database directly. The results are temporarily held in the
database for processing.
v To give administrative access to the database through one of the following
methods:
– A DB2 open database connectivity (ODBC) client
– An Oracle Java database connectivity (JDBC) client
– Structured Query Language (SQL) to the server from an administrative-client
command line

Consider using an extra 50 GB of temporary space for every 500 GB of space for
file objects and optimization. See the guidelines in the following table. In the
example that is used in the preceding step, a total of 1.7 TB of database space is
required for file objects and optimization for 500 clients. Based on that calculation,
200 GB is required for temporary space. The total amount of required database
space is 1.9 TB.

Database size Minimum temporary-space requirement

< 500 GB 50 GB
≥ 500 GB and < 1 TB 100 GB
≥ 1 TB and < 1.5 TB 150 GB
≥ 1.5 and < 2 TB 200 GB
≥ 2 and < 3 TB 250 - 300 GB
≥ 3 and < 4 TB 350 - 400 GB

Estimating database space requirements based on storage pool

capacity
To estimate database space requirements based on storage pool capacity, use a ratio
of 1 - 5%. For example, if you require 200 TB of storage pool capacity, the size of
your database is expected to be 2 - 10 TB. As a general rule, make your database
as large as possible to prevent running out of space. If you run out of database
space, server operations and client-store operations can fail.

Database space required when running database operations

The database manager of the Tivoli Storage Manager server manages and allocates
system memory and disk space for the database. The amount of database space
that the system requires depends on the amount of system memory that is
available and the server workload.

For example, expiration processing can use a large amount of database space. If
there is not enough system memory in the database to store the files identified for
expiration, some of the data is allocated to temporary disk space. During
expiration processing, if a node or file space is selected that is too large to process,
the database manager cannot sort the data.

To run database operations, consider adding more database space for the following
scenarios:

654 IBM Tivoli Storage Manager for Windows: Administrator's Guide

v The database has a small amount of space and the server operation that requires
temporary space uses the remaining free space.
v The file spaces are large, or the file spaces have a policy assigned to it that
creates many file versions.
v The Tivoli Storage Manager server must run with limited memory.
v An out of database space error is displayed when you deploy a Tivoli Storage
Manager V6 server.

Attention: Do not install or upgrade to a different version, release, or fix pack of

DB2 software because doing so can damage the database.

The database manager and temporary space

The Tivoli Storage Manager server database manager manages and allocates
system memory and disk space for the database. The amount of database space
you require depends on the amount of system memory available and the server
workload.

The database manager sorts data in a specific sequence, as per the SQL statement
that you issue to request the data. Depending on the workload on the server, and
if there is more data than the database manager can manage, the data (that is
ordered in sequence) is allocated to temporary disk space. Data is allocated to
temporary disk space when there is a large result set. The database manager
dynamically manages the memory used when data is allocated to temporary disk
space.

For example, expiration processing can produce a large result set. If there is not
enough system memory on the database to store the result set, some of the data is
allocated to temporary disk space. During expiration processing, if a node or file
space are selected that are too large to process, the database manager does not
have enough memory to sort the data.

To run database operations, consider adding more database space for the following
scenarios:
v The database has a small amount of space and the server operation that requires
temporary space uses the remaining free space.
v The file spaces are large, or the file spaces has a policy assigned to it which
creates many file versions.
v The Tivoli Storage Manager server must run with limited memory. The database
uses the Tivoli Storage Manager server main memory to run database
operations. However, if there is insufficient memory available, the Tivoli Storage
Manager server allocates temporary space on disk to the database. For example,
if 10G of memory is available and database operations require 12G of memory,
the database uses temporary space.
v An out of database space error is displayed when you deploy a Tivoli Storage
Manager V6 server. Monitor the server activity log for messages related to
database space.

Important: Do not change the DB2 software that is installed with the Tivoli
Storage Manager installation packages and fix packs. Do not install or upgrade to a
different version, release, or fix pack, of DB2 software to avoid damage to the
database.

Chapter 20. Managing the database and recovery log 655

 Estimating recovery log space requirements
In Tivoli Storage Manager, the term recovery log comprises the active log, the
archive log, the active log mirror, and the archive failover log. The amount of space
that you require for the recovery log depends on various factors, including, for
example, the amount of client activity with the server.

Active and archive log space

When you estimate space requirements for active and archive logs, include some
extra space for contingencies such as occasional heavy workloads and failovers.

In Tivoli Storage Manager servers V6.1 and later, the active log can be a maximum
size of 128 GB. The archive log size is limited to the size of the file system that it is
installed on.

Use the following general guidelines when you estimate the size of the active log:
v The suggested starting size for the active log is 16 GB.
v Ensure that the active log is at least large enough for the amount of concurrent
activity that the server typically handles. As a precaution, try to anticipate the
largest amount of work that the server manages at one time. Provision the active
log with extra space that can be used if needed. Consider using 20% of extra
space.
v Monitor used and available active log space. Adjust the size of the active log as
needed, depending upon factors such as client activity and the level of server
operations.
v Ensure that the directory that holds the active log is as large as, or larger than,
the size of the active log. A directory that is larger than the active log can
accommodate failovers, if they occur.
v Ensure that the file system that contains the active log directory has at least 8
GB of free space for temporary log movement requirements.

The suggested starting size for the archive log is 48 GB.

The archive log directory must be large enough to contain the log files that are
generated since the previous full backup. For example, if you perform a full
backup of the database every day, the archive log directory must be large enough
to hold the log files for all the client activity that occurs during 24 hours. To
recover space, the server deletes obsolete archive log files after a full backup of the
database. If the archive log directory becomes full and a directory for archive
failover logs does not exist, log files remain in the active log directory. This
condition can cause the active log directory to fill up and stop the server. When the
server restarts, some of the existing active-log space is released.

After the server is installed, you can monitor archive log utilization and the space
in the archive log directory. If the space in the archive log directory fills up, it can
cause the following problems:
v The server is unable to perform full database backups. Investigate and resolve
this problem.
v Other applications write to the archive log directory, exhausting the space that is
required by the archive log. Do not share archive log space with other
applications including other Tivoli Storage Manager servers. Ensure that each
server has a separate storage location that is owned and managed by that
specific server.

656 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 For guidance about the layout and tuning of the active log and archive log, see
Optimizing Performance.
Related tasks:
“Increasing the size of the active log” on page 678

Example: Estimating active and archive log sizes for basic client-store
operations:

Basic client-store operations include backup, archive, and space management. Log
space must be sufficient to handle all store transactions that are in progress at one
time.

To determine the sizes of the active and archive logs for basic client-store
operations, use the following calculation:
number of clients x files stored during each transaction
x log space needed for each file

This calculation is used in the example in the following table.

Table 52. Basic client-store operations
Example
Item values Description
Maximum number of client nodes 300 The number of client nodes that back up, archive, or migrate
that back up, archive, or migrate files files every night.
concurrently at any time
Files stored during each transaction 4096 The default value of the server option TXNGROUPMAX is 4096.
Log space that is required for each 3053 bytes The value of 3053 bytes for each file in a transaction
file represents the log bytes that are needed when backing up
files from a Windows client where the file names are 12 -
120 bytes.

This value is based on the results of tests performed under

laboratory conditions. The tests consisted of backup-archive
clients performing backup operations to a random-access
disk (DISK) storage pool. DISK pools result in more log use
than sequential-access storage pools. Consider a value larger
than 3053 bytes if the data being stored has file names that
are longer than 12 - 120 bytes.
1
Active log: Suggested size 19.5 GB Use the following calculation to determine the size of the
active log. One GB equals 1,073,741,824 bytes.

(300 clients x 4096 files stored during each

transaction x 3053 bytes for each file) ÷ 1,073,741,824
bytes = 3.5 GB

Increase that amount by the suggested starting size of 16

GB:

3.5 + 16 = 19.5 GB

Chapter 20. Managing the database and recovery log 657

Table 52. Basic client-store operations (continued)
Example
Item values Description
1
Archive log: Suggested size 58.5 GB Because of the requirement to be able to store archive logs
across three server database-backup cycles, multiply the
estimate for the active log by 3 to estimate the total archive
log requirement.

3.5 x 3 = 10.5 GB

Increase that amount by the suggested starting size of 48

GB:

10.5 + 48 = 58.5 GB
1
The example values in this table are used only to illustrate how the sizes for active logs and archive logs are
calculated. In a production environment that does not use deduplication, 16 GB is the suggested minimum size for
an active log. The suggested minimum size for an archive log in a production environment that does not use
deduplication is 48 GB. If you substitute values from your environment and the results are larger than 16 GB and 48
GB, use your results to size the active log and archive log.

Monitor your logs and adjust their size if necessary.

Example: Estimating active and archive log sizes for clients that use multiple
sessions:

If the client option RESOURCEUTILIZATION is set to a value that is greater than the
default, the concurrent workload for the server increases.

To determine the sizes of the active and archive logs when clients use multiple
sessions, use the following calculation:
number of clients x sessions for each client x files stored
during each transaction x log space needed for each file

This calculation is used in the example in the following table.

Table 53. Multiple client sessions
Item Example values Description
Maximum number of client 300 1000 The number of client nodes that back up, archive, or
nodes that back up, archive, migrate files every night.
or migrate files concurrently
at any time
Possible sessions for each 3 3 The setting of the client option RESOURCEUTILIZATION is
client larger than the default. Each client session runs a maximum
of three sessions in parallel.
Files stored during each 4096 4096 The default value of the server option TXNGROUPMAX is 4096.
transaction

658 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Table 53. Multiple client sessions (continued)
Item Example values Description
Log space that is required for 3053 3053 The value of 3053 bytes for each file in a transaction
each file represents the log bytes needed when backing up files from
a Windows client where the file names are 12 - 120 bytes.

This value is based on the results of tests performed under

laboratory conditions. Tests consisted of clients performing
backup operations to a random-access disk (DISK) storage
pool. DISK pools result in more log use than
sequential-access storage pools. Consider a value larger
than 3053 bytes if the data being stored has file names that
are longer than 12 - 120 bytes.
1 1
Active log: Suggested size 26.5 GB 51 GB The following calculation was used for 300 clients. One GB
equals 1,073,741,824 bytes.

(300 clients x 3 sessions for each client x 4096 files

stored during each transaction x 3053 bytes for each
file) ÷ 1,073,741,824 = 10.5 GB

Increase that amount by the suggested starting size of 16

GB:

10.5 + 16 = 26.5 GB

The following calculation was used for 1000 clients. One

GB equals 1,073,741,824 bytes.

(1000 clients x 3 sessions for each client x 4096

files store during each transaction x 3053 bytes for
each file) ÷ 1,073,741,824 = 35 GB

Increase that amount by the suggested starting size of 16

GB:

35 + 16 = 51 GB
1 1
Archive log: Suggested size 79.5 GB 153 GB Because of the requirement to be able to store archive logs
across three server-database backup cycles, the estimate for
the active log is multiplied by 3:

10.5 x 3 = 31.5 GB

35 x 3 = 105 GB

Increase those amounts by the suggested starting size of 48

GB:

31.5 + 48 = 79.5 GB

105 + 48 = 153 GB
1
The example values in this table are used only to illustrate how the sizes for active logs and archive logs are
calculated. In a production environment that does not use deduplication, 16 GB is the suggested minimum size for
an active log. The suggested minimum size for an archive log in a production environment that does not use
deduplication is 48 GB. If you substitute values from your environment and the results are larger than 16 GB and 48
GB, use your results to size the active log and archive log.

Monitor your active log and adjust its size if necessary.

Chapter 20. Managing the database and recovery log 659

 Example: Estimating active and archive log sizes for simultaneous write
operations:

If client backup operations use storage pools that are configured for simultaneous
write, the amount of log space that is required for each file increases.

The log space that is required for each file increases by about 200 bytes for each
copy storage pool that is used for a simultaneous write operation. In the example
in the following table, data is stored to two copy storage pools in addition to a
primary storage pool. The estimated log size increases by 400 bytes for each file. If
you use the suggested value of 3053 bytes of log space for each file, the total
number of required bytes is 3453.

This calculation is used in the example in the following table.

Table 54. Simultaneous write operations
Example
Item values Description
Maximum number of client nodes 300 The number of client nodes that back up, archive, or migrate
that back up, archive, or migrate files files every night.
concurrently at any time
Files stored during each transaction 4096 The default value of the server option TXNGROUPMAX is 4096.
Log space that is required for each 3453 bytes 3053 bytes plus 200 bytes for each copy storage pool.
file
The value of 3053 bytes for each file in a transaction
represents the log bytes that are needed when backing up
files from a Windows client where the file names are 12 - 120
bytes.

This value is based on the results of tests performed under

laboratory conditions. The tests consisted of backup-archive
clients performing backup operations to a random-access
disk (DISK) storage pool. DISK pools result in more log use
than sequential-access storage pools. Consider a value larger
than 3053 bytes if the data being stored has file names that
are longer than 12 - 120 bytes.
1
Active log: Suggested size 20 GB Use the following calculation to determine the size of the
active log. One GB equals 1,073,741,824 bytes.

(300 clients x 4096 files stored during each

transaction x 3453 bytes for each file) ÷ 1,073,741,824
bytes = 4.0 GB

Increase that amount by the suggested starting size of 16 GB:

4 + 16 = 20 GB
1
Archive log: Suggested size 60 GB Because of the requirement to be able to store archive logs
across three server database-backup cycles, multiply the
estimate for the active log by 3 to estimate the archive log
requirement:

4 GB x 3 = 12 GB

Increase that amount by the suggested starting size of 48 GB:

12 + 48 = 60 GB

660 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Table 54. Simultaneous write operations (continued)
Example
Item values Description
1
The example values in this table are used only to illustrate how the sizes for active logs and archive logs are
calculated. In a production environment that does not use deduplication, 16 GB is the suggested minimum size for
an active log. The suggested minimum size for an archive log in a production environment that does not use
deduplication is 48 GB. If you substitute values from your environment and the results are larger than 16 GB and 48
GB, use your results to size the active log and archive log.

Monitor your logs and adjust their size if necessary.

Example: Estimating active and archive log sizes for basic client store operations
and server operations:

Migration of data in server storage, identification processes for data deduplication,

reclamation, and expiration might run concurrently with client store operations.
Administrative tasks such as administrative commands or SQL queries from
administrative clients can also run concurrently with client store operations. Server
operations and administrative tasks that run concurrently can increase the active
log space that is required.

For example, migration of files from the random-access (DISK) storage pool to a
sequential-access disk (FILE) storage pool uses approximately 110 bytes of log
space for each file that is migrated. For example, suppose that you have 300
backup-archive clients and each one of them backs up 100,000 files every night.
The files are initially stored on DISK and then migrated to a FILE storage pool. To
estimate the amount of active log space that is required for the data migration, use
the following calculation. The number of clients in the calculation represents the
maximum number of client nodes that back up, archive, or migrate files
concurrently at any time.
300 clients x 100,000 files for each client x 110 bytes = 3.1 GB

Add this value to the estimate for the size of the active log that calculated for basic
client store operations.

Example: Estimating active and archive log sizes under conditions of extreme
variation:

Problems with running out of active log space can occur if you have many
transactions that complete quickly and some transactions that take much longer to
complete. A typical case occurs when many workstation or file-server backup
sessions are active and a few very large database server-backup sessions are active.
If this situation applies to your environment, you might need to increase the size
of the active log so that the work completes successfully.

Chapter 20. Managing the database and recovery log 661

 Example: Estimating archive log sizes with full database backups:

The Tivoli Storage Manager server deletes unnecessary files from the archive log
only when a full database backup occurs. Consequently, when you estimate the
space that is required for the archive log, you must also consider the frequency of
full database backups.

For example, if a full database backup occurs once a week, the archive log space
must be able to contain the information in the archive log for a full week.

The difference in archive log size for daily and full database backups is shown in
the example in the following table.
Table 55. Full database backups
Example
Item values Description
Maximum number of client nodes 300 The number of client nodes that back up, archive, or migrate
that back up, archive, or migrate files files every night.
concurrently at any time
Files stored during each transaction 4096 The default value of the server option TXNGROUPMAX is 4096.
Log space that is required for each 3453 bytes 3053 bytes for each file plus 200 bytes for each copy storage
file pool.

The value of 3053 bytes for each file in a transaction

represents the log bytes needed when backing up files from
a Windows client where the file names are 12 - 120 bytes.

This value is based on the results of tests performed under

laboratory conditions. Tests consisted of clients performing
backup operations to a random-access disk (DISK) storage
pool. DISK pools result in more log use than
sequential-access storage pools. Consider a value larger than
3053 bytes if the data being stored has file names that are
longer than 12 - 120 bytes.
1
Active log: Suggested size 20 GB Use the following calculation to determine the size of the
active log. One GB equals 1,073,741,824 bytes.

(300 clients x 4096 files per transaction x 3453 bytes

per file) ÷ 1,073,741,824 bytes = 4.0 GB

Increase that amount by the suggested starting size of 16

GB:

4 + 16 = 20 GB
1
Archive log: Suggested size with a 60 GB Because of the requirement to be able to store archive logs
full database backup every day across three backup cycles, multiply the estimate for the
active log by 3 to estimate the total archive log requirement:

4 GB x 3 = 12 GB

Increase that amount by the suggested starting size of 48

GB:

12 + 48 = 60 GB

662 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Table 55. Full database backups (continued)
Example
Item values Description
1
Archive log: Suggested size with a 132 GB Because of the requirement to be able to store archive logs
full database every week across three server database-backup cycles, multiply the
estimate for the active log by 3 to estimate the total archive
log requirement. Multiply the result by the number of days
between full database backups:

(4 GB x 3 ) x 7 = 84 GB

Increase that amount by the suggested starting size of 48

GB:

84 + 48 = 132 GB
1
The example values in this table are used only to illustrate how the sizes for active logs and archive logs are
calculated. In a production environment that does not use deduplication, 16 GB is the suggested minimum size for
an active log. The suggested starting size for an archive log in a production environment that does not use
deduplication is 48 GB. If you substitute values from your environment and the results are larger than 16 GB and 48
GB, use your results to size the active log and archive log.

Monitor your logs and adjust their size if necessary.

Example: Estimating active and archive log sizes for data deduplication
operations:

If you deduplicate data, you must consider its effects on space requirements for
active and archive logs.

The following factors affect requirements for active and archive log space:
The amount of deduplicated data
The effect of data deduplication on the active log and archive log space
depends on the percentage of data that is eligible for deduplication. If the
percentage of data that can be deduplicated is relatively high, more log
space is required.
The size and number of extents
Approximately 1,500 bytes of active log space are required for each extent
that is identified by a duplicate-identification process. For example, if
250,000 extents are identified by a duplicate-identification process, the
estimated size of the active log is 358 MB:
250,000 extents identified during each process x 1,500 bytes
for each extent = 358 MB

Consider the following scenario. Three hundred backup-archive clients

back up 100,000 files each night. This activity creates a workload of
30,000,000 files. The average number of extents for each file is two.
Therefore, the total number of extents is 60,000,000, and the space
requirement for the archive log is 84 GB:
60,000,000 extents x 1,500 bytes for each extent = 84 GB
A duplicate-identification process operates on aggregates of files. An
aggregate consists of files that are stored in a given transaction, as
specified by the TXNGROUPMAX server option. Suppose that the TXNGROUPMAX
server option is set to the default of 4096. If the average number of extents
for each file is two, the total number of extents in each aggregate is 8192,
and the space required for the active log is 12 MB:
Chapter 20. Managing the database and recovery log 663
 8192 extents in each aggregate x 1500 bytes for each extent =
12 MB
The timing and number of the duplicate-identification processes
The timing and number of duplicate-identification processes also affects the
size of the active log. Using the 12 MB active-log size that was calculated
in the preceding example, the concurrent load on the active log is 120 MB
if 10 duplicate-identification processes are running in parallel:
12 MB for each process x 10 processes = 120 MB
File size
Large files that are processed for duplicate identification can also affect the
size of the active log. For example, suppose that a backup-archive client
backs up an 80 GB, file-system image. This object can have a high number
of duplicate extents if, for example, the files included in the file system
image were backed up incrementally. For example, assume that a file
system image has 1.2 million duplicate extents. The 1.2 million extents in
this large file represent a single transaction for a duplicate-identification
process. The total space in the active log that is required for this single
object is 1.7 GB:
1,200,000 extents x 1,500 bytes for each extent = 1.7 GB

If other, smaller duplicate-identification processes occur at the same time as

the duplicate-identification process for a single large object, the active log
might not have enough space. For example, suppose that a storage pool is
enabled for deduplication. The storage pool has a mixture of data,
including many relatively small files that range from 10 KB to several
hundred KB. The storage pool also has few large objects that have a high
percentage of duplicate extents.
To take into account not only space requirements but also the timing and
duration of concurrent transactions, increase the estimated size of the
active log by a factor of two. For example, suppose that your calculations
for space requirements are 25 GB (23.3 GB + 1.7 GB for deduplication of a
large object). If deduplication processes are running concurrently, the
suggested size of the active log is 50 GB. The suggested size of the archive
log is 150 GB.
The examples in the following tables show calculations for active and
archive logs. The example in the first table uses an average size of 700 KB
for extents. The example in the second table uses an average size of 256
KB. As the examples show, the average deduplicate-extent size of 256 KB
indicates a larger estimated size for the active log. To minimize or prevent
operational problems for the server, use 256 KB to estimate the size of the
active log in your production environment.
Table 56. Average duplicate-extent size of 700 KB
Item Example values Description
Size of largest single 800 GB 4 TB The granularity of processing for deduplication is at the file
object to deduplicate level. Therefore, the largest single file to deduplicate
represents the largest transaction and a correspondingly
large load on the active and archive logs.
Average size of 700 KB 700 KB The deduplication algorithms use a variable block method.
extents Not all deduplicated extents for a given file are the same
size, so this calculation assumes an average size for extents.

664 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Table 56. Average duplicate-extent size of 700 KB (continued)
Item Example values Description
Extents for a given 1,198,372 6,135,667 bits Using the average extent size (700 KB), these calculations
file bits represent the total number of extents for a given object.

The following calculation was used for an 800 GB object:

(800 GB ÷ 700 KB) = 1,198,372 bits

The following calculation was used for a 4 TB object: (4 TB ÷

700 KB) = 6,135,667 bits
Active log: Suggested 1.7 GB 8.6 GB The estimated active log space that are needed for this
size that is required transaction.
for the deduplication
of a single large object
during a single
duplicate-
identification process
1 1
Active log: Suggested 66 GB 79.8 GB After considering other aspects of the workload on the
total size server in addition to deduplication, multiply the existing
estimate by a factor of two. In these examples, the active log
space required to deduplicate a single large object is
considered along with previous estimates for the required
active log size.

The following calculation was used for multiple transactions

and an 800 GB object:

(23.3 GB + 1.7 GB) x 2 = 50 GB

Increase that amount by the suggested starting size of 16 GB:

50 + 16 = 66 GB

The following calculation was used for multiple transactions

and a 4 TB object:

(23.3 GB + 8.6 GB) x 2 = 63.8 GB

Increase that amount by the suggested starting size of 16 GB:

63.8 + 16 = 79.8 GB
1 1
Archive log: 198 GB 239.4 GB Multiply the estimated size of the active log by a factor of 3.
Suggested size
The following calculation was used for multiple transactions
and an 800 GB object:

50 GB x 3 = 150 GB

Increase that amount by the suggested starting size of 48 GB:

150 + 48 = 198 GB

The following calculation was used for multiple transactions

and a 4 TB object:

63.8 GB x 3 = 191.4 GB

Increase that amount by the suggested starting size of 48 GB:

191.4 + 48 = 239.4 GB

Chapter 20. Managing the database and recovery log 665

Table 56. Average duplicate-extent size of 700 KB (continued)
Item Example values Description
1
The example values in this table are used only to illustrate how the sizes for active logs and archive logs are
calculated. In a production environment that uses deduplication, 32 GB is the suggested minimum size for an active
log. The suggested minimum size for an archive log in a production environment that uses deduplication is 96 GB.
If you substitute values from your environment and the results are larger than 32 GB and 96 GB, use your results to
size the active log and archive log.

Monitor your logs and adjust their size if necessary.

Table 57. Average duplicate-extent size of 256 KB

Item Example values Description
Size of largest single 800 GB 4 TB The granularity of processing for deduplication is at the file
object to deduplicate level. Therefore, the largest single file to deduplicate
represents the largest transaction and a correspondingly
large load on the active and archive logs.
Average size of 256 KB 256 KB The deduplication algorithms use a variable block method.
extents Not all deduplicated extents for a given file are the same
size, so this calculation assumes an average extent size.
Extents for a given 3,276,800 bits 16,777,216 Using the average extent size, these calculations represent
file bits the total number of extents for a given object.

The following calculation was used for multiple transactions

and an 800 GB object:

(800 GB ÷ 256 KB) = 3,276,800 bits

The following calculation was used for multiple transactions

and a 4 TB object:

(4 TB ÷ 256 KB) = 16,777,216 bits

Active log: Suggested 4.5 GB 23.4 GB The estimated size of the active log space that is required for
size that is required this transaction.
for the deduplication
of a single large object
during a single
duplicate-
identification process

666 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Table 57. Average duplicate-extent size of 256 KB (continued)
Item Example values Description
1 1
Active log: Suggested 71.6 GB 109.4 GB After considering other aspects of the workload on the
total size server in addition to deduplication, multiply the existing
estimate by a factor of 2. In these examples, the active log
space required to deduplicate a single large object is
considered along with previous estimates for the required
active log size.

The following calculation was used for multiple transactions

and an 800 GB object:

(23.3 GB + 4.5 GB) x 2 = 55.6 GB

Increase that amount by the suggested starting size of 16 GB:

55.6 + 16 = 71.6 GB

The following calculation was used for multiple transactions

and a 4 TB object:

(23.3 GB + 23.4 GB) x 2 = 93.4 GB

Increase that amount by the suggested starting size of 16 GB:

93.4 + 16 = 109.4 GB
1 1
Archive log: 214.8 GB 328.2 GB The estimated size of the active log multiplied by a factor of
Suggested size 3.

The following calculation was used for an 800 GB object:

55.6 GB x 3 = 166.8 GB

Increase that amount by the suggested starting size of 48 GB:

166.8 + 48 = 214.8 GB

The following calculation was used for a 4 TB object:

93.4 GB x 3 = 280.2 GB

Increase that amount by the suggested starting size of 48 GB:

280.2 + 48 = 328.2 GB
1
The example values in this table are used only to illustrate how the sizes for active logs and archive logs are
calculated. In a production environment that uses deduplication, 32 GB is the suggested minimum size for an active
log. The suggested minimum size for an archive log in a production environment that uses deduplication is 96 GB.
If you substitute values from your environment and the results are larger than 32 GB and 96 GB, use your results to
size the active log and archive log.

Monitor your logs and adjust their size if necessary.

Chapter 20. Managing the database and recovery log 667

 Archive log space for online database reorganization
Tivoli Storage Manager manages table and index reorganization. The Tivoli Storage
Manager server requires additional archive log space if an online table must be
reorganized.

The Tivoli Storage Manager logs reorganization activity. No information is lost if

there is a database error that requires crash recovery.

The space required for an online table reorganization is determined by the

following factors:
v The number of rows to be reorganized
v The number of indexes
v The size of the index keys
v The current organization of the table

Typically, every row in a table is moved twice during an online table

reorganization. For each index, each table row must update the index key to reflect
the new location. After all accesses to the old location are complete, the index key
is updated again to remove references to the old location. When the row is moved
back, updates to the index key are performed again. This activity is logged to
make online table reorganization fully recoverable. There is a minimum of two
data log records (each including the row data) and four index log records (each
including the key data) for each row (assuming one index).

Clustering indexes are prone to filling up the index pages, causing index splits and
merges that must also be logged. A number of the tables implemented by the
server have more than one index. A table that has four indexes would require 16
index log records for each row that is moved for the reorganization.

The server monitors characteristics of the database, the active log, and the archive
log to determine if a database backup is needed. For example, during an online
table reorganization, if the file system for the archive log space begins to fill up,
the server triggers a database backup. When a database backup is started, any
online table reorganization in progress is paused so that the database backup can
operate without contending for resources with the reorganization.

Active log mirror space

The active log can be mirrored so that the mirrored copy can be used if the active
log files cannot be read. There can be only one active log mirror.

Creating a log mirror is a suggested option. If you increase the size of the active
log, the log mirror size is increased automatically. Mirroring the log can affect
performance because of the doubled I/O activity that is required to maintain the
mirror. The additional space that the log mirror requires is another factor to
consider when deciding whether to create a log mirror.

If the mirror log directory becomes full, the server issues error messages to the
activity log and to the [Link]. Server activity continues.

668 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 Archive failover log space
The archive failover log is used by the server if the archive log directory runs out
of space.

Specifying an archive failover log directory is optional, but it can prevent problems
that occur if the archive log runs out of space. If both the archive log directory and
the drive or file system where the archive failover log directory is located become
full, the data remains in the active log directory. This condition can cause the
active log to fill up, which causes the server to halt. If you use an archive failover
log directory, place the archive log directory and the archive failover log directory
on different physical drives.

Important: Maintain adequate space for the archive log directory, and consider
using an archive failover log directory. For example, suppose the drive or file
system where the archive log directory is located becomes full and the archive
failover log directory does not exist or is full. If this situation occurs, the log files
that are ready to be moved to the archive log remain in the active log directory. If
the active log becomes full, the server stops.

By monitoring the usage of the archive failover log, you can determine whether
additional space is needed for the archive log. The goal is to minimize the need to
use the archive failover log by ensuring that the archive log has adequate space.

The locations of the archive log and the archive failover log are set during initial
configuration. If you use the DSMSERV LOADFORMAT utility instead of the wizard to
configure the server, you specify the ARCHLOGDIRECTORY parameter for the archive
log directory. In addition, you specify the ARCHFAILOVERLOGDIRECTORY parameter for
the archive failover log directory. If the archive failover log is not created at initial
configuration, you can create it by specifying the ARCHFAILOVERLOGDIRECTORY option
in the server options file.

Monitoring space utilization for the database and recovery

logs
To determine the amount of used and available active log space, you issue the
QUERY LOG command. To monitor space utilization in the database and recovery
logs, you can also check the activity log for messages.

Active log

If the amount of available active log space is too low, the following messages are
displayed in the activity log:
ANR4531I: IC_AUTOBACKUP_LOG_USED_SINCE_LAST_BACKUP_TRIGGER
This message is displayed when the active log space exceeds the maximum
specified size. The Tivoli Storage Manager server starts a full database
backup.
To change the maximum log size, halt the server. Open the [Link]
file, and specify a new value for the ACTIVELOGSIZE option. When you are
finished, restart the server.
ANR0297I: IC_BACKUP_NEEDED_LOG_USED_SINCE_LAST_BACKUP
This message is displayed when the active log space exceeds the maximum
specified size. You must back up the database manually.

Chapter 20. Managing the database and recovery log 669

 To change the maximum log size, halt the server. Open the [Link]
file, and specify a new value for the ACTIVELOGSIZE option. When you are
finished, restart the server.
ANR4529I: IC_AUTOBACKUP_LOG_UTILIZATION_TRIGGER
The ratio of used active-log space to available active-log space exceeds the
log utilization threshold. If at least one full database backup has occurred,
the Tivoli Storage Manager server starts an incremental database backup.
Otherwise, the server starts a full database backup.
ANR0295I: IC_BACKUP_NEEDED_LOG_UTILIZATION
The ratio of used active-log space to available active-log space exceeds the
log utilization threshold. You must back up the database manually.

Archive log

If the amount of available archive log space is too low, the following message is
displayed in the activity log:
ANR0299I: IC_BACKUP_NEEDED_ARCHLOG_USED
The ratio of used archive-log space to available archive-log space exceeds
the log utilization threshold. The Tivoli Storage Manager server starts a full
automatic database backup.

Database

If the amount of space available for database activities is too low, the following
messages are displayed in the activity log:
ANR2992W: IC_LOG_FILE_SYSTEM_UTILIZATION_WARNING_2
The used database space exceeds the threshold for database space
utilization. To increase the space for the database, use the EXTEND DBSPACE
command, the EXTEND DBSPACE command, or the DSMSERV FORMAT
utility with the DBDIR parameter.
ANR1546W: FILESYSTEM_DBPATH_LESS_1GB
The available space in the directory where the server database files are
located is less than 1 GB.
When a Tivoli Storage Manager server is created with the DSMSERV
FORMAT utility or with the configuration wizard, a server database and
recovery log are also created. In addition, files are created to hold database
information used by the database manager. The path specified in this
message indicates the location of the database information used by the
database manager. If space is unavailable in the path, the server can no
longer function.
You must add space to the file system or make space available on the file
system or disk.

670 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Monitoring the database and recovery log
Monitor the database, log space, and file systems where the directories are located
to ensure that space is always available.

About this task

You can monitor the database and recovery log space whether the server is online
or offline.

Procedure
v When the Tivoli Storage Manager server is online, you can issue the QUERY
DBSPACE command to view the total space, used space, and free space for the file
systems or drives where your database located. To view the same information
when the server is offline, issue the DSMSERV DISPLAY DBSPACE command. The
following example shows the output of this command:
Location: d:\tsm\db001
Total Space (MB): 46,080.00
Used Space (MB): 20,993.12
Free Space (MB): 25,086.88

Location: e:\tsm\db002
Total Space (MB): 46,080.00
Used Space (MB): 20,992.15
Free Space (MB): 25,087.85

Location: f:\tsm\db003
Total Space (MB): 46,080.00
Used Space (MB): 20,993.16
Free Space (MB): 25,086.84

Location: g:\tsm\db004
Total Space (MB): 46,080.00
Used Space (MB): 20,992.51
Free Space (MB): 25,087.49

v To view more detailed information about the database when the server is online,
issue the QUERY DB command. The following example shows the output of this
command if you specify FORMAT=DETAILED:
Database Name: TSMDB1
Total Size of File System (MB): 184,320
Space Used by Database (MB): 83,936
Free Space Available (MB): 100,349
Total Pages: 6,139,995
Usable Pages: 6,139,451
Used Pages: 6,135,323
Free Pages: 4,128
Buffer Pool Hit Ratio: 100.0
Total Buffer Requests: 97,694,823,985
Sort Overflows: 0
Package Cache Hit Ratio: 100.0
Last Database Reorganization: 06/25/2009 [Link]
Full Device Class Name: LTO1_CLASS
Incrementals Since Last Full: 0
Last Complete Backup Date/Time: 06/06/2009 [Link]

v When the Tivoli Storage Manager server is online, issue the QUERY LOG
FORMAT=DETAILED command to display the total space, used space, and free space
for the active log, and the locations of all the logs. To display the same
information when the Tivoli Storage Manager server is offline, issue the DSMSERV
DISPLAY LOG command. The following example shows the output of this
command:

Chapter 20. Managing the database and recovery log 671

 Total Space(MB): 38,912
Used Space(MB): 401.34
Free Space(MB): 38,358.65
Active Log Directory: h:\tsm\activelog
Archive Log Directory: k:\tsm\archivelog
Mirror Log Directory: i:\tsm\mirrorlog
Archive Failover Log Directory: j:\tsm\archfailoverlog

v You can view information about the database on the server console and in the
activity log. You can set the level of database information by using the SET
DBREPORTMODE command. Specify that no diagnostic information is displayed
(NONE), that all diagnostic information is displayed (FULL), or that the only
events that are displayed are those that are exceptions and might represent
errors (PARTIAL). The default is PARTIAL.

Increasing the size of the database

You can increase the size of the Tivoli Storage Manager database by creating
directories and adding them to the database.

About this task

The Tivoli Storage Manager server can use all of the space that is available to the
drives or file systems where the database directories are located. To ensure that
database space is always available, monitor the space in use by the server and the
file systems where the directories are located.

The maximum size of the Tivoli Storage Manager database is 4 TB.

Use the QUERY DBSPACE command to display the number of free pages in the table
space and the free space that is available to the database. If the number of free
pages is low and there is plenty of free space available, the database allocates more
space. However, if free space in drives or file systems is low, it might not be
possible to expand the database.

You must format the database with Tivoli Storage Manager V6.2 server or later. If
you are using a Tivoli Storage Manager V6.1 server or earlier, you must manually
upgrade table spaces to the DB2 9.7 version of table spaces. The instructions for
manually upgrading the table spaces Tivoli Storage Manager V6.1 server and
earlier is different for each version level. Contact IBM Software Support for the
instructions for your version of Tivoli Storage Manager.

If you want to increase space for the database, you can create new directories and
add them by using the EXTEND DBSPACE command. You can also use the DSMSERV
EXTEND DBSPACE utility when the server is offline. When you issue these commands
with the default parameter values, data is redistributed across the new database
directories and storage space in the old directories is reclaimed. This action makes
the new directories available for use immediately and parallel I/O performance is
improved.

To redistribute data to new directories, the storage paths must be able to host the
directories and data. Make sure that sufficient disk space is available for the
operation. New directories must be empty.

For example, assume that your table space has four directories over two existing
storage paths (two directories per path), and you add another storage path to the

672 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 database. When data is redistributed for the table space, two new directories of the
same size as the existing directories are created on the new storage path. If your
table space has a size of approximately 100 GB, then you need 50 GB of free space
on the new storage path for redistribution to be successful. After redistribution is
complete and space is reclaimed, the total used space is still the same amount of
100 GB, until you begin adding more data.

The following image shows the example table space before and after data is
redistributed.

Table space with data redistributed

Existing table space - 100 GB across new directories - 150 GB

directories directories

storage path storage path

directories directories

storage path storage path

new directories
new
storage path

Figure 79. Redistributing data for table spaces

The process of redistributing data and reclaiming space uses considerable

resources. Ensure that you plan ahead when you want to add space to the
database. For planning requirements, review the EXTEND DBSPACE command.

If you do not want to redistribute data at the same time that you add directories,
you can set the RECLAIMSTORAGE parameter in the EXTEND DBSPACE command to No.
You can perform the tasks to redistribute data and reclaim space after the database
size is increased, but the steps must be done manually.

Procedure

To add space to the database, complete the following steps:

1. Create one or more directories for the database on separate drives or file
systems.
Review best practices for creating database directories in the Optimizing
Performance guide.
2. Issue the EXTEND DBSPACE command to add the directory or directories to the
database. The directories must be accessible to the user ID of the database
manager. By default, data is redistributed across all database directories and
space is reclaimed.

Chapter 20. Managing the database and recovery log 673

 Remember: The time that is needed to complete redistribution of data and
reclaiming of space is variable, depending on the size of your database. Make
sure that you plan adequately.
3. Halt and restart the server to fully use the new directories.

Example

Windows: For example, to add a directory and a drive to the storage space for the
database and then redistribute data and reclaim space, issue the following
command:
extend dbspace h:\tsmdb005,I:

To increase the size of the database without redistributing data and reclaiming
space, issue the following command:
extend dbspace h:\tsmdb005,I: reclaim=no

Manually redistributing data and reclaiming space after you

add directories to the database
If you add directories to the Tivoli Storage Manager database with the EXTEND
DBSPACE command and choose not to redistribute data and reclaim space at the
same time, you can complete these tasks later by using DB2 commands.

Before you begin

Restriction: The redistribution process, also known as rebalancing, only works

with DB2 version 9.7 or later table spaces, which are created when you format a
new Tivoli Storage Manager version 6.2 or later server. If you upgraded or restored
your Tivoli Storage Manager server from version 6.1, and you want to rebalance
your table spaces, contact IBM Software Support for instructions.

Redistribution uses considerable resources. Review the following guidelines before

you start the procedure:
v Run the process when the server is not handling a heavy workload.
v To redistribute data to new directories, the storage paths must be able to host
the directories and data. Make sure that sufficient disk space is available for the
operation.
v The time that is required to redistribute data and reclaim space might vary. File
system layout, the ratio of new paths to existing storage paths, server hardware,
and concurrent operations are all factors in determining the time requirement.
Start the process with one small and one medium-sized table space and then try
a larger table space. Use your results as a reference to estimate the time that is
needed to process remaining table spaces.
v Do not interrupt the process. If you try to stop it, for example, by halting the
process that is completing the work, you must stop and restart the DB2 server.
When the server is restarted, it will go into crash recovery mode, which takes
several minutes, after which the process resumes.

Procedure

Complete the following steps to redistribute data and then reclaim space for each
table space. For the best performance, rebalance a table space and then reduce the
size for that table space. While the size for the first table space is being reduced,
you can start rebalancing on the second table space, and so on.

674 IBM Tivoli Storage Manager for Windows: Administrator's Guide

1. Open the DB2 command line processor and issue the following command:
db2 connect to tsmdb1
2. List DB2 table spaces by issuing the following command. To display details
about each table space, including the total size of a table space and how many
bytes are used in each file system where the table space is located, include show
detail.
db2 list tablespaces show detail

You only need to redistribute data on Database Managed Space (DMS) table
spaces. The following example output shows where the table space type is
identified:
Tablespaces for Current Database

Tablespace ID = 0
Name = SYSCATSPACE
Type = Database managed space <---DMS table space.
Contents = All permanent data. Regular table space.
State = 0x0000
Detailed explanation:
Normal
3. Use the list that you obtained in Step 2 to identify each DMS table space. For
each DMS table space, issue the following command to start redistribution of
data for the first DMS space:
db2 alter tablespace tablespace_name rebalance
4. Monitor the data redistribution progress by issuing the following command:
db2list utilities show detail

If the rebalance process is running, the command output shows Type =

REBALANCE, and also indicates how many extents are moved and how many
remain to be moved.
The following example output shows where these details are displayed:
ID = 6219
Type = REBALANCE <--- Data is being redistributed.
Database Name = AX4
Partition Number = 0
Description = Tablespace ID: 37
Start Time = 04/27/2009 [Link].932471
State = Executing
Invocation Type = User
Throttling:
Priority = Unthrottled
Progress Monitoring:
Estimated Percentage Complete = 15
Total Work = 22366 extents <--- Total extents to be moved.
Completed Work = 3318 extents <--- Total extents moved.
Start Time = 04/27/2009 21:37

The value in the Completed Work field should increase as the redistribution
progresses. The db2diag log also records status about the process, including
start and complete time and what percentage of the process is complete at a
certain time.
5. After the redistribution process is completed, reduce the size for each table
space. During and after the operation, table spaces have a much larger total
size because directories are added. Issue the following command:
db2 alter tablespace tablespace_name reduce max

Chapter 20. Managing the database and recovery log 675

Reducing the size of the database
If a significant amount of data has been deleted from the database, consider
reducing the database size.

Procedure
1. Create a file that contains a list of directories that represent the new directories.
For example, [Link].
2. Run a full database backup by issuing the following command:
backup db devclass=tapeclass type=full
3. Halt the server.
4. Remove the database instance by issuing the following command:
dsmserv removedb TSMDB1
5. Restore the database specifying the file that contains the directories to be used.
For example, you can issue the following command:
dsmserv restore db todate=today on=[Link]
6. Restart the server.

Scheduling table and index reorganization

Unexpected database growth, unexpected active and archive log space
requirements, and reduced server performance can occur over time if tables or the
indexes associated with tables are not reorganized. Tables and indexes are
reorganized by default by Tivoli Storage Manager. If automatic reorganization is
affecting server performance, you can schedule reorganization manually.

Before you begin

Reorganization of table data can be initiated by the Tivoli Storage Manager server
or by DB2. If server-initiated reorganization is enabled. The server analyzes
selected database tables and indexes based on table activity, and determines when
reorganization is required. The database manager runs a reorganization while
server operations continue. If reorganization by DB2 is enabled, DB2 controls the
reorganization process. Reorganization by DB2 is not recommended.

Server-initiated table reorganization is enabled by default. If you are deduplicating

data, consider enabling server-initiated index reorganization. However, table and
index reorganization requires locks on the database that might interfere with server
operations. Reorganization also requires significant processor, active log, and
archive log resources. If reorganization is causing performance problems, you can
manually schedule table and index reorganizations.

The best time to start a reorganization is when server activity is low and when
access to the database is optimal. Schedule table reorganization for databases on
servers that are not running deduplication. Schedule table and index
reorganization on servers that are running deduplication.

Important: Ensure that the system on which the Tivoli Storage Manager server is
running has sufficient memory and processor resources. To assess how busy the
system is over time, use operating system tools to assess the load on the system.
You can also review the [Link] file and the server activity log. If the system
does not have sufficient resources, reorganization processing might be incomplete,
or it might degrade or destabilize the system.

676 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Restrictions to table and index reorganization
If you manually schedule a table or index reorganization, be aware of the
restrictions that can affect performance.

Table reorganization

Under certain conditions, table reorganization can pause if reorganization activity

exceeds the time specified by the REORGDURATION server option. Table reorganization
can also pause to make active log space available. If table reorganization is paused
for either reason, reorganization starts again if the interval is not exceeded.
Otherwise, reorganization remains paused until the following day at the next
scheduled reorganization.

Do not schedule database backups during a table reorganization interval. Database

backup takes precedence over table reorganization:
v If a database backup is running, table reorganization cannot start until the
database backup completes. After the database backup completes, table
reorganization can start if the start time is within the interval specified by the
REORGDURATION server option.
v If a running table reorganization is paused, the server checks to determine
whether a database backup is required. If a backup is required, it starts
automatically.

Index reorganization

If a database backup is running, index reorganization cannot start until the

database backup completes. After the database backup completes, index
reorganization can start if the start time is within the interval specified by the
REORGDURATION server option.

Unlike table reorganization, index reorganization cannot be paused. If an index

reorganization is running, the server cannot start a full database backup. If an
index is large, reorganization can take a long time to complete and can generate a
large amount of data in the active log and the archive log. To reduce the possibility
of having to manually start a full database backup, take one or more of the
following steps:
v Consider allocating relatively large amounts of archive log space that the server
can use during index reorganization.
v To ensure that running index reorganizations complete, do not schedule a full
database backup for least an hour after the interval during which an index
reorganization can start.
v Ensure that the value of the REORGDURATION server option limits when index
reorganizations can start.
Continue to monitor database and log space usage during index reorganization. If
the server starts to run out of archive log space, begin a full database backup.

Chapter 20. Managing the database and recovery log 677

 Scheduling table or index reorganization
To schedule table or index reorganization, you specify the earliest time and an
interval after that time during which reorganization can start.

Procedure

To schedule table or index reorganization, complete the following steps:

1. Enable table or index reorganization by setting one or both of the following
server options to yes in the [Link] file.
v ALLOWREORGTABLE
v ALLOWREORGINDEX
2. Specify the time that the server can start table or index reorganization by
setting the REORGBEGINTIME server option. Specify the time by using a 24-hour
format. The best time to start a reorganization is when server activity is low
and when access to the database is optimal. Schedule table reorganization for
databases on servers that are not running data deduplication. Schedule table
and index reorganization on servers that are running data deduplication.
3. Specify the interval during which table or index reorganization must start by
setting the REORGDURATION server option. Specify the time as a number of hours.

Results

If you set only the REORGBEGINTIME option, reorganization is enabled for an entire
day. If you do not specify the REORGBEGINTIME option, but you specify a value for
the REORGDURATION option, the reorganization interval starts at 6:00 a.m. and runs
for the specified number of hours.

If your system experiences performance problems, modify the values of the

REORGBEGINTIME option and the REORGDURATION option. When you modify the
values, you ensure that table or index reorganization is not running during periods
of peak server activity.

Increasing the size of the active log

If the log is running out of space, the current transaction is rolled back, and the
server issues an error message and halts. You cannot restart the server until the
active log size is increased.

Procedure

To increase the size of the active log while the server is halted, complete the
following steps:
1. Issue the DSMSERV DISPLAY LOG offline utility to display the size of the active
log.
2. Ensure that the location for the active log has enough space for the increased
log size. If a log mirror exists, its location must also have enough space for the
increased log size.
3. Halt the server.
4. In the [Link] file, update the ACTIVELOGSIZE option to the new maximum
size of the active log, in megabytes. For example, to change the active log to its
maximum size of 128 GB, enter the following server option:
activelogsize 131072

678 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 5. If you plan to use a new active log directory, update the directory name
specified in the ACTIVELOGDIRECTORY server option. The new directory must be
empty and must be accessible to the user ID of the database manager.
6. Restart the server.

Results

Log files of 512 MB are automatically defined until the size specified in the
ACTIVELOGSIZE option is reached. If a log mirror exists, log files are also defined
automatically.

Reducing the size of the active log

You can reduce the size of the active log during the startup of the Tivoli Storage
Manager server.

Procedure

If you have too much active log space, you can reduce the size of the active log by
completing the following steps:
1. Stop the Tivoli Storage Manager server.
2. In the [Link] file, change the ACTIVELOGSIZE option to the new size of the
active log, in megabytes. For example, to reduce the active log by 8 GB, enter
the following server option:
dsmserv activelogsize 8000
3. Restart the server.

Results

When you reduce the size of the active log, you must restart the Tivoli Storage
Manager server twice. The first restart changes the DB2 parameters. The second
restart removes the log files that are no longer required on the disk.

Moving the database and recovery log on a server

You can move the database and logs to different locations on the server.

About this task

You might want to move the database and logs to take advantage of a larger or
faster disk. You have the following options:
v “Moving both the database and recovery log” on page 680
v “Moving only the database” on page 680
v “Moving only the active log, archive log, or archive failover log” on page 681

For information about moving a Tivoli Storage Manager server to another machine,
see “Moving the Tivoli Storage Manager server to another system” on page 608

Chapter 20. Managing the database and recovery log 679

 Moving both the database and recovery log
You can move the database, active log, and archive logs that are on the same file
system to various directories on different file systems for better protection.

Procedure
1. Back up the database by issuing the following command:
backup db type=full devclass=files
2. Halt the server.
3. Create directories for the database, active logs, and archive logs. The directories
must be accessible to the user ID of the database manager. For example:
mkdir l:\tsm\db005
mkdir m:\tsm\db006
mkdir n:\tsm\db007
mkdir o:\tsm\db008
mkdir p:\tsm\activelog
mkdir q:\tsm\archivelog
4. Create a file that lists the locations of the database directories. This file is used
if the database must be restored. Enter each location on a separate line. For
example, these are the contents of the [Link] file:
l:\tsm\db005
m:\tsm\db006
n:\tsm\db007
o:\tsm\db008
5. Remove the database instance by issuing the following command:
dsmserv removedb TSMDB1
6. Issue the DSMSERV RESTORE DB utility to move the database and create the new
active log. For example:
dsmserv restore db todate=today on=[Link]
activelogdir=p:\tsm\activelog
7. Restart the server.
8. Move the archive logs from the old directory to the new directory. Ensure that
you preserve any subdirectory structure. Use and verify a recursive copy
directory command:
xcopy /s k:\tsm\archivelog\* q:\tsm\archivelog

Moving only the database

You can move only the database to new directories.

Procedure

To move the database from one location on the server to another location, follow
this procedure:
1. Back up the database by issuing the following command:
backup db type=full devclass=files
2. Halt the server.
3. Create directories for the database. The directories must be accessible to the
user ID of the database manager. For example:
mkdir l:\tsm\db005
mkdir m:\tsm\db006
mkdir n:\tsm\db007
mkdir o:\tsm\db008

680 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 4. Create a file that lists the locations of the database directories. This file will be
used if the database must be restored. Enter each location on a separate line.
For example, here are the contents of the [Link] file:
l:\tsm\db005
m:\tsm\db006
n:\tsm\db007
o:\tsm\db008
5. Remove the database instance by issuing the following command:
dsmserv removedb TSMDB1
6. Issue the DSMSERV RESTORE DB utility to move the database to the new
directories. For example:
dsmserv restore db todate=today on=[Link]
7. Start the server.

Moving only the active log, archive log, or archive failover log
You can move only the active log, archive log, or archive failover log from one
directory to another. You can move the logs as part of performance tuning,
replacing a hard disk drive, or changing to a file system with more space.

Procedure
1. Stop the server.
2. Create a directory by using the following commands for the log that you want
to move.
Table 58. Commands for new directories
Log name Command
Active log mkdir p:\tsm\activelog
Archive log mkdir q:\tsm\archivelog
Archive failover mkdir e:\tsmserver1\archfaillog2
log

3. Update the [Link] option file for the new directory by using one of the
following commands.
Table 59. Server option example to update the [Link] option file
Log name Server option example
Active log activelogdirectory p:\tsm\activelog
Archive log archlogdirectory q:\tsm\archivelog
Archive failover archfailoverlogdirectory
log e:\tsmserver1\archfaillog2

4. Restart the server.

5. Move or copy the logs from the old directory to the new directory.
Table 60. Commands to move or copy the logs
Log name Command
Active log No action is necessary. The active logs are automatically moved by the
database manager from the old directory to the new directory.

Chapter 20. Managing the database and recovery log 681

 Table 60. Commands to move or copy the logs (continued)
Log name Command
Archive log Preserve the subdirectory structure. Use and verify a recursive copy
directory command. For example:
xcopy /s k:\tsm\archivelog\*
q:\tsm\archivelog
Archive failover Preserve the subdirectory structure. Use and verify a recursive copy
log directory command. For example:
xcopy /s c:\tsmserver1\archfaillog\*
f:\tsmserver1\archfaillog2

6. Optional: Remove the old directory by using the rmdir command.

Specifying alternative locations for database log files

A database backup includes a database image and associated recovery-log files.
During a database restore operation, the recovery-log files are temporarily stored in
a subdirectory called RstDbLog. Log files that are created during day-to-day
operations and that cannot be stored in the active log directory or in the archive
log directory can be temporarily stored in a subdirectory called failarch.

Procedure

To specify alternative locations for the database log files, complete the following
steps:
1. To specify the location of subdirectories RstDbLog and failarch, use the
ARCHFAILOVERLOGDIRECTORY server option. The Tivoli Storage Manager server
creates the RstDbLog and failarch subdirectories in the directory that is
specified by the server option.

Restriction: If you do not specify the location of the subdirectories, the Tivoli
Storage Manager server automatically creates the two subdirectories under the
archive log directory

If the archive log directory becomes full, it can limit the amount of space that is
available for archived log files. If you must use the archive log directory, you
can increase its size to accommodate both the RstDbLog and failarch
directories.
2. Use a file system that is different from the file system that is specified by the
ACTIVELOGDIRECTORY and ARCHLOGDIRECTORY parameters.

Tip: If you do not set the ARCHFAILOVERLOGDIRECTORY option, the Tivoli Storage
Manager server creates the RstDbLog and failarch subdirectories automatically
in the directory that is specified for the ARCHLOGDIRECTORY parameter on the
DSMSERV FORMAT or DSMSERV LOADFORMAT command. You must specify the
ARCHLOGDIRECTORY parameter for these commands.
3. For a database restore operation, you can specify the location of the RstDbLog
subdirectory, but not the failarch subdirectory, by using the RECOVERYDIR
parameter on the DSMSERV RESTORE DB command. Consider allocating a
relatively large amount of temporary disk space for the restore operation.
Because database restore operations occur relatively infrequently, the RstDbLog
subdirectory can contain many logs from backup volumes that are stored in
preparation for pending roll-forward-restore processing.

682 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Specifying an alternative location with the
ARCHFAILOVERLOGDIRECTORY server option or parameter
If you specify the ARCHFAILOVERLOGDIRECTORY server option or the
ARCHFAILOVERLOGDIRECTORY parameter in the DSMSERV FORMAT or DSMSERV LOADFORMAT
command, the server creates the RstDbLog and failarch subdirectories in the
archive failover-log directory.

About this task

The server also updates the DB2 parameter OVERFLOWLOGPATH that points to the
RstDbLog subdirectory and the DB2 parameter FAILARCHPATH, that points to the
failarch subdirectory. For details about these parameters, see the DB2 information
center at [Link]

For example, suppose that you specify archlogfailover as the value of the
ARCHFAILOVERLOGDIRECTORY parameter on the DSMSERV FORMAT command:
dsmserv format
dbdir=a:\db001
activelogdirectory=b:\activelog
archlogdirectory=c:\archlog
archfailoverlogdirectory=d:\archlogfailover

The server creates the subdirectories RstDbLog and failarch in the parent directory
archlogfailover. The server also updates the following DB2 parameters:
OVERFLOWLOGPATH=d:\archlogfailover\RstDbLog
FAILARCHPATH=d:\archlogfailover\failarch

The server also updates the value of the ARCHFAILOVERLOGDIRECTORY option in the
server options file, [Link]:
ARCHFAILOVERLOGDIRECTORY d:\archlogfailover

Specifying an alternative location with the ARCHLOGDIRECTORY

server option or parameter
If you do not set the ARCHFAILOVERLOGDIRECTORY parameter or server option, the
server automatically creates the subdirectories RstDbLog and failarch in the
directory that is specified for the ARCHLOGDIRECTORY parameter. In addition, the
server updates the DB2 parameters OVERFLOWLOGPATH and FAILARCHPATH to point to
those subdirectories.

About this task

For details about these parameters, see the DB2 Information Center at
[Link]

For example, suppose that you specify a value of archlog for the ARCHLOGDIRECTORY
parameter in a DSMSERV FORMAT command. You do not specify the
ARCHFAILOVERLOGDIRECTORY parameter:
dsmserv format
dbdir=a:\db001
activelogdirectory=b:\activelog
archlogdirectory=c:\archlog

The Tivoli Storage Manager server creates the subdirectories RstDbLog and
failarch under the archlog parent directory. The server also updates the following
DB2 parameters:

Chapter 20. Managing the database and recovery log 683

 OVERFLOWLOGPATH=c:\archlog\RstDbLog
FAILARCHPATH=c:\archlog\failarch

The server also updates the value of the ARCHLOGDIRECTORY option in the server
options file, [Link]:
ARCHLOGDIRECTORY c:\archlog

Specifying the location of RstDbLog using the RECOVERYDIR

parameter
For a database restore operation, you can specify a temporary location for the
RstDbLog subdirectory by using the RECOVERYDIR parameter on the DSMSERV RESTORE
DB command. The Tivoli Storage Manager server creates the RstDbLog subdirectory
in the directory that is specified by that parameter. The server does not create the
failarch subdirectory.

About this task

The server also updates the DB2 parameter, OVERFLOWLOGPATH, that points to
RstDbLog. For details about this parameter, see the DB2 Information Center at
[Link]

Restriction: The location specified by the RECOVERYDIR parameter takes precedence

over the location specified by the ARCHFAILOVERLOGDIRECTORY or ARCHLOGDIRECTORY
server option or parameter.

For example, for a point-in-time database restore, you can issue the following
command:
dsmserv restore db
todate=5/12/2011
totime=14:45
recoverydir=e:\recovery

The server creates the RstDbLog subdirectory in the parent recovery directory. The
server also updates the OVERFLOWLOGPATH parameter:
OVERFLOWLOGPATH=e:\recovery\RstDbLog

After the database is restored, the RstDbLog subdirectory reverts to its location as
specified by the server option ARCHFAILOVERLOGDIRECTORY or ARCHLOGDIRECTORY in
the server options file, [Link].

Adding optional logs after server initialization

You can specify an archive failover log and a mirror log if they were not created
during server initialization.

Procedure

Complete the following procedure to add one or both of the logs.

1. Create the directories for the logs. The directories must be accessible to the user
ID of the database manager.
2. Halt the server.
3. Add the log server option or options to the server options file, [Link]. For
example:
archfailoverlogdirectory h:\tsm\archfailoverlog
mirrorlogdirectory n:\tsm\mirrorlog

684 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 4. Save the server options file.
5. Halt and restart the server.

Transaction processing
A transaction is the unit of work exchanged between the client and server.

The log records for a given transaction are moved into stable storage when the
transaction is committed. The database information that is stored on disk remains
consistent because the server ensures that the recovery log records, which represent
the updates to these database pages, are written to disk.

During restart-recovery, the server uses the active and archive log information to
maintain the consistency of the server by redoing and, if necessary, undoing
ongoing transactions from the time that the server was halted. The transaction is
then committed to the database.

Transaction commit is a function of all the log records for that transaction being
written to the recovery log. This function ensures that the necessary redo and undo
information is available to replay these transaction changes against the database
information.

Files moved as a group between client and server

The client program can move multiple files or directories between the client and
server before it commits the data to server storage.

A transaction that contains multiple files or directories is called a transaction group.

Using the TXNGROUPMAX server option, you can specify the number of files or
directories that are contained within a transaction group. You can use the
TXNGROUPMAX option to increase performance when Tivoli Storage Manager writes to
tape. This performance increase can be considerable when you transfer multiple
small files.

If you increase the value of TXNGROUPMAX by a large amount, monitor the effects on
the recovery log. A larger value for the TXNGROUPMAX option can have the following
impact:
v Affect the performance of client backup, archive, restore, and retrieve operations.
v Increase utilization of the recovery log, as well as increase the length of time for
a transaction to commit.
Also consider the number of concurrent sessions to be run. It might be possible to
run with a higher TXNGROUPMAX value with a few clients running. However, if there
are hundreds of clients running concurrently, you might need to reduce the
TXNGROUPMAX to help manage the recovery log usage and support this number of
concurrent clients. If the performance effects are severe, they might affect server
operations. See “Monitoring the database and recovery log” on page 671 for more
information.

The following examples show how the TXNGROUPMAX option can affect performance
throughput for operations to tape and the recovery log.
v The TXNGROUPMAX option is set to 20. The MAXSESSIONS option, which specifies the
maximum number of concurrent client/server sessions, is set to 5. Five
concurrent sessions are processing, and each file in the transaction requires 10
logged database operations. This would be a concurrent load of:
20*10*5=1000

Chapter 20. Managing the database and recovery log 685

 This represents 1000 log records in the recovery log. Each time a transaction
commits the data, the server can free 200 log records.
v The TXNGROUPMAX option is set to 2000. The MAXSESSIONS option is set to 5. Five
concurrent sessions are processing, and each file in the transaction requires 10
logged database operations, resulting in a concurrent load of:
2000*10*5=100 000

This represents 100 000 log records in the recovery log. Each time a transaction
commits the data, the server can free 20 000 log records.

Remember: Over time and as transactions end, the recovery log can release the
space that is used by the oldest transactions. These transactions complete, and the
log space usage increases.

Based on the previous two examples, five concurrent transactions with a

TXNGROUPMAX setting of 2000 consume much more space in the recovery log. This
increase in log space usage also increases the risk of running out of recovery log
space.

The following table shows a comparison of the examples of the preceding

TXNGROUPMAX settings. This example becomes more significant if a given log record
takes 100 bytes.
Table 61. Example of log bytes that are consumed by five concurrent sessions
TXNGROUPMAX Setting Number of Log Bytes Consumed
TXNGROUPMAX=20 100,000
TXNGROUPMAX=2000 10,000,000

You can use several server options to tune server performance and reduce the risk
of running out of recovery log space:
v Use the THROUGHPUTTIMETHRESHOLD and THROUGHPUTDATATHRESHOLD options with
the TXNGROUPMAX option to prevent a slower performing node from holding a
transaction open for extended periods.
v Increase the size of the recovery log when you increase the TXNGROUPMAX setting.

Evaluate the performance and characteristics of each node before increasing the
TXNGROUPMAX setting. Nodes that have only a few larger objects to transfer do not
benefit as much as nodes that have multiple, smaller objects to transfer. For
example, a file server benefits more from a higher TXNGROUPMAX setting than does a
database server that has one or two large objects. Other node operations can
consume the recovery log at a faster rate. Be careful when increasing the
TXNGROUPMAX settings for nodes that often perform high log-usage operations. The
raw or physical performance of the disk drives that are holding the database and
recovery log can become an issue with an increased TXNGROUPMAX setting. The
drives must handle higher transfer rates to handle the increased load on the
recovery log and database.

You can set the TXNGROUPMAX option as a global server option value, or you can set
it for a single node. For optimal performance, specify a lower TXNGROUPMAX value
(between 4 and 512). Select higher values for individual nodes that can benefit
from the increased transaction size.

Refer to the REGISTER NODE command and the server options in the Administrator's
Reference.
686 IBM Tivoli Storage Manager for Windows: Administrator's Guide
Chapter 21. Managing a network of Tivoli Storage Manager
servers
You might have several Tivoli Storage Manager servers in your network, at the
same or different locations. Tivoli Storage Manager provides functions to help you
configure, manage, and monitor the servers connected to a network.

About this task

An administrator working at one Tivoli Storage Manager server can work with
Tivoli Storage Manager servers at other locations around the world.

See the following topics:

Concepts:
“Concepts for managing server networks”
“Enterprise configuration” on page 688

Tasks:
“Setting up communications among servers” on page 692
“Setting up communications for enterprise configuration and enterprise event logging” on
page 692
“Setting up communications for command routing with multiple source servers” on page
697
“Completing tasks on multiple servers” on page 725
“Using virtual volumes to store data on another server” on page 732

Concepts for managing server networks

In a network of Tivoli Storage Manager servers, a server can play several different
roles. For example, a server can send volumes to be archived on another server
and also receive routed commands from a different server.

To manage a network of servers, you can use the following Tivoli Storage Manager
capabilities:
v Configure and manage multiple servers with enterprise configuration.
Distribute a consistent configuration for Tivoli Storage Manager servers through
a configuration manager to managed servers. By having consistent
configurations, you can simplify the management of a large number of servers
and clients.
v Perform tasks on multiple servers by using command routing, enterprise logon,
and enterprise console.
v Send server and client events to another server for logging.
v Monitor many servers and clients from a single server.
v Store data on another server by using virtual volumes.

© Copyright IBM Corp. 1993, 2013 687

 In the descriptions for working with a network of servers, when a server sends
data, that server is sometimes referred to as a source server, and when a server
receives data, it is sometimes referred to as a target server. In other words, one
Tivoli Storage Manager server may be both a source and a target server. At the
same time, any Tivoli Storage Manager server can still provide backup, archive,
and space management services to clients.

For details, see “Licensing IBM Tivoli Storage Manager” on page 595.

Enterprise configuration
The Tivoli Storage Manager enterprise configuration functions make it easier to
consistently set up and manage a network of Tivoli Storage Manager servers. You
can set up configurations on one server and distribute the configurations to other
servers. You can make changes to configurations and have the changes
automatically distributed.

Figure 80 on page 689 illustrates a simple configuration. To use enterprise

configuration, select the Tivoli Storage Manager server that is to act as the
configuration manager. You might want to dedicate a new server for this purpose. At
the configuration manager, define the details of the server configurations that you
want to distribute. For example:
v Set up backup and archive policies and client option sets
v Designate one or more administrators to have access to the servers, and control
their authority levels
v Define the servers that you want the configuration manager to manage or
communicate with, and you set up communications among the servers
In one or more profiles, point to the definitions of the configuration information
that you want to use to manage other servers.

On each server that is to receive the configuration information, identify the server
as a managed server by defining a subscription to one or more profiles owned by the
configuration manager. All the definitions associated with the profiles are then
copied into the managed server's database. Things defined to the managed server
in this way are managed objects that cannot be changed by the managed server.
From then on, the managed server gets any changes to the managed objects from
the configuration manager via the profiles. Managed servers receive changes to
configuration information at time intervals set by the servers, or by command.

For details, see “Setting up enterprise configurations” on page 701.

688 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 Policies

Client Option Sets

Configuration Administrators
Manager Profiles
Schedules

Servers, Server Groups

Scripts

Subscriptions to profiles

Managed
servers
Managed
objects

Figure 80. Enterprise configuration

Command routing
Use the command-line interface to route commands to other servers.

The other servers must be defined to the server to which you are connected. You
must also be registered on the other servers as an administrator with the
administrative authority that is required for the command. To make routing
commands easier, you can define a server group that has servers as members.
Commands that you route to a server group are sent to all servers in the group.

For details, see “Setting up server groups” on page 728 and “Routing commands”
on page 726.

Central monitoring for the Tivoli Storage Manager server

Tivoli Storage Manager provides you with several ways to centrally monitor the
activities of a server network.

The following methods are ways in which you can centrally monitor activities:
v Enterprise event logging, in which events are sent from one or more of servers
to be logged at an event server.
For a description of the function, see “Enterprise event logging: logging events
to another server” on page 820. For information about communications setup,
see “Setting up communications for enterprise configuration and enterprise
event logging” on page 692.
v Use the Operations Center to view server status and alerts. See Chapter 17,
“Managing the storage environment with the Operations Center,” on page 587
for more information.
v Allowing designated administrators to log in to any of the servers in the
network with a single user ID and password.

Chapter 21. Managing a network of Tivoli Storage Manager servers 689

 See “Managing multiple servers through a web-based interface” on page 725.
v Routing query commands to one or more of the servers in the network.
For a description of the function, see “Routing commands to one or more
servers” on page 726. For information about communications setup, see “Setting
up communications for enterprise configuration and enterprise event logging”
on page 692.

Data storage on another server

Tivoli Storage Manager lets one server store data in and retrieve data from the
storage pool of another server. This data, stored as virtual volumes, can include
database and storage pool backups, disaster recovery plan files, and data that is
directly backed up, archived, or space managed from client nodes.

The data can also be a recovery plan file created by using disaster recovery
manager (DRM). The source server is a client of the target server, and the data for
the source server is managed only by the source server. In other words, the source
server controls the expiration and deletion of the files that comprise the virtual
volumes on the target server.

To use virtual volumes to store database and storage pool backups and recovery
plan files, you must have the disaster recovery manager function. For details, see
“Licensing IBM Tivoli Storage Manager” on page 595.

For information about using virtual volumes with DRM, see Chapter 32, “Disaster
recovery manager,” on page 1005.

Examples: management of multiple Tivoli Storage Manager

servers
The functions for managing multiple servers can be applied in many ways.

Here are two scenarios to give you some ideas about how you can use the
functions:
v Setting up and managing Tivoli Storage Manager servers primarily from one
location. For example, an administrator at one location controls and monitors
servers at several locations.
v Setting up a group of Tivoli Storage Manager servers from one location, and
then managing the servers from any of the servers. For example, several
administrators are responsible for maintaining a group of servers. One
administrator defines the configuration information on one server for
distributing to servers in the network. Administrators on the individual servers
in the network manage and monitor the servers.

Example: management of multiple servers from one location

Enterprise management lets you set up and manage the servers in your network
from one location, the enterprise console.

For example, suppose that you are an administrator who is responsible for Tivoli
Storage Manager servers at your own location, plus servers at branch office
locations. Servers at each location have similar storage resources and client
requirements. You can set up the environment as follows:
v Set up an existing or new Tivoli Storage Manager server as a configuration
manager.

690 IBM Tivoli Storage Manager for Windows: Administrator's Guide

v Set up communications so that a configuration manager can send commands to
its managed servers.
v Define the configuration you want to distribute by defining policy domains,
schedules, and so on. Associate the configuration information with profiles.
v Have the managed servers subscribe to profiles.
v Activate policies and set up storage pools as needed on the managed servers.
v Set up enterprise monitoring by setting up one server as an event server. The
event server can be the same server as the configuration manager or a different
server.

After you complete the setup, you can manage many servers as if there was just
one. You can perform any of the following tasks:
v Have administrators that can manage the group of servers from anywhere in the
network by using the enterprise console, an interface available through a Web
browser.
v Have consistent policies, schedules, and client option sets on all servers.
v Make changes to configurations and have the changes automatically distributed
to all servers. Allow local administrators to monitor and tune their own servers.
v Perform tasks on any server or all servers by using command routing from the
enterprise console.
v Back up the databases of the managed servers on the automated tape library
that is attached to the server that is the configuration manager. You use virtual
volumes to accomplish this.
v Log on to individual servers from the enterprise console without having to
re-enter your password, if your administrator ID and password are the same on
each server.

Example: management of multiple servers from any server

Enterprise management lets you manage the servers in your network from many
locations.

For example, suppose that you are an administrator responsible for servers located
in different departments on a college campus. The servers have some requirements
in common, but also have many unique client requirements. You can set up the
environment as follows:
v Set up an existing or new Tivoli Storage Manager server as a configuration
manager.
v Set up communications so that commands can be sent from any server to any
other server.
v Define any configuration that you want to distribute by defining policy
domains, schedules, and so on, on the configuration manager. Associate the
configuration information with profiles.
v Have the managed servers subscribe to profiles as needed.
v Activate policies and set up storage pools as needed on the managed servers.
v Set up enterprise monitoring by setting up one server as an event server. The
event server can be the same server as the configuration manager or a different
server.

After setting up in this way, you can manage the servers from any server. You can
do any of the following tasks:
v Use enterprise console to monitor all the servers in your network.

Chapter 21. Managing a network of Tivoli Storage Manager servers 691

 v Perform tasks on any or all servers using the enterprise console and command
routing.
v Manage the group of servers from anywhere in the network. Allow local
administrators to monitor and tune their own servers.

Enterprise-administration planning
To take full advantage of the functions of enterprise administration, you should
decide on the servers you want to include in the enterprise network, the server
from which you want to manage the network, and other important issues.

Consider the following items when planning for Enterprise Administration:

v The servers you want to include in the enterprise network. The servers must
have unique names.
v The server or servers from which you want to manage the network. Servers can
have multiple roles. For example, one server can act as a server for
backup-archive clients, as the configuration manager, and as the event server.
You can also set up separate servers to fill each of these roles.
v Whether you want administrators to have the ability to route commands to
other servers. If you want administrators to route commands, decide on the
servers from which and to which commands will be routed.
v The administrator activities you want to be centrally managed.
v The authority level of the administrators and the servers to which they should
have access.

Setting up communications among servers

You can set up communications for enterprise configuration, enterprise event
logging, and command routing. When you set up communications among servers
for any purpose, ensure that servers have unique names.

About this task

Communication setup for server-to-server virtual volumes is described in “Setting

up source and target servers for virtual volumes” on page 734. See “Setting the
server name” on page 613 for more information before using the SET
SERVERNAME command.

Setting up communications for enterprise configuration and

enterprise event logging
The communication setup for enterprise configuration and enterprise event
logging, which is through TCP/IPv4 or IPv6, is identical.

About this task

The examples shown here apply to both functions. If you are set up for one, you
are set up for the other. However, be aware that the configuration manager and
event server are not defined simply by setting up communications. You must
identify a server as a configuration manager (SET CONFIGMANAGER command)
or an event server (DEFINE EVENTSERVER command). Furthermore, a
configuration manager and an event server can be the same server or different
servers.

692 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Enterprise configuration
Each managed server must be defined to the configuration manager, and
the configuration manager must be defined to each managed server.
Enterprise event logging
Each server sending events to an event server must be defined to the event
server, and the event server must be defined to each source server.

The following examples of setting up communications could be used to create

these configurations:
v A server named HEADQUARTERS as a configuration manager and two servers,
MUNICH and STRASBOURG, as managed servers.
v HEADQUARTERS as an event server and MUNICH and STRASBOURG as
source servers.
For a pair of servers to communicate with each other, each server must be defined
to the other. For example, if a configuration manager manages three managed
servers, there are three server pairs. You can issue separate definitions from each
server in each pair, or you can “cross define” a pair in a single operation. Cross
definition can be useful in large or complex networks. The following scenarios and
accompanying figures illustrate the two methods.

Using separate definitions – Follow this sequence:

Procedure
1. On MUNICH: Specify the server name and password of MUNICH.
On STRASBOURG: Specify the server name and password of STRASBOURG.
On HEADQUARTERS: Specify the server name and password of
HEADQUARTERS.
2. On HEADQUARTERS: Define MUNICH (whose password is BERYL and
whose address is [Link]:1919) and STRASBOURG (whose password is
FLUORITE and whose address is [Link]:1715).
On MUNICH and STRASBOURG: Define HEADQUARTERS (whose
password is AMETHYST and whose address is [Link]:1823).

Results

Figure 81 on page 694 shows the servers and the commands issued on each:

Chapter 21. Managing a network of Tivoli Storage Manager servers 693

 set servername headquarters
Headquarters set serverpassword amethyst

define server munich

serverpassword=beryl
hladdress=[Link]
lladdress=1919

define server strasbourg

serverpassword=fluorite
hladdress=[Link]
lladdress=1715

Munich Strasbourg

set servername munich set servername strasbourg

set serverpassword beryl set serverpassword flourite

define server headquarters define server headquarters

serverpassword=amethyst serverpassword=amethyst
hladdress=[Link] hladdress=[Link]
lladdress=1823 lladdress=1823

Figure 81. Communication configuration with separate server definitions

Using Cross Definitions – Follow this sequence:

1. On MUNICH: Specify the server name, password, and high and low level
addresses of MUNICH. Specify that cross define is permitted.
On STRASBOURG: Specify the server name, password, and high and low
level addresses of STRASBOURG. Specify that cross define is permitted.
On HEADQUARTERS: Specify the server name, password, and high and low
level addresses of HEADQUARTERS.
2. On HEADQUARTERS: Define MUNICH and STRASBOURG, specifying that
cross define should be done.

Figure 82 on page 695 shows the servers and the commands issued on each:

694 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 set servername headquarters
set serverpassword amethyst
Headquarters set serverhladdress [Link]
set serverlladdress 1823

define server munich crossdefine=yes

serverpassword=beryl hladdress=[Link]
lladdress=1919

define server strasbourg crossdefine=yes

serverpassword=fluorite hladdress=[Link]
lladdress=1715

Munich Strasbourg

set servername munich set servername strasbourg

set serverpassword beryl set serverpassword fluorite
set serverhladdress [Link] set serverhladdress [Link]
set serverlladdress 1919 set serverlladdress 1715
set crossdefine on set crossdefine on

Figure 82. Communication configuration with cross definition

Note: Issuing the SET SERVERNAME command can affect scheduled backups
until a password is re-entered. Windows clients use the server name to identify
which passwords belong to which servers. Changing the server name after the
clients are connected forces the clients to re-enter the passwords. On a network
where clients connect to multiple servers, it is recommended that all of the servers
have unique names. See the Administrator's Reference for more details.

Communication security
Security for this communication configuration is enforced through the exchange of
passwords (which are encrypted) and, in the case of enterprise configuration only,
verification keys.

Communication among servers, which is through TCP/IP, requires that the servers
verify server passwords (and verification keys). For example, assume that
HEADQUARTERS begins a session with MUNICH:
1. HEADQUARTERS, the source server, identifies itself by sending its name to
MUNICH.
2. The two servers exchange verification keys (enterprise configuration only).
3. HEADQUARTERS sends its password to MUNICH, which verifies it against
the password stored in its database.
4. If MUNICH verifies the password, it sends its password to HEADQUARTERS,
which, in turn, performs password verification.

Chapter 21. Managing a network of Tivoli Storage Manager servers 695

 Note: If another server named MUNICH tries to contact HEADQUARTERS for
enterprise configuration, the attempt fails. This is because the verification key will
not match. If MUNICH was moved or restored, you can issue the UPDATE
SERVER command with the FORCERESYNC parameter to override the condition.

Setting up communications for command routing

When setting up communications for command routing, you must define the target
servers to the source servers, and the same administrator must be registered on all
servers. Using enterprise configuration, you can easily distribute the administrator
information to all the servers.

About this task

Note: You must be registered as an administrator with the same name and
password on the source server and all target servers. The privilege classes do not
need to be the same on all servers. However, to successfully route a command to
another server, an administrator must have the minimum required privilege class
for that command on the server from which the command is being issued.

For command routing in which one server will always be the sender, you would
only define the target servers to the source server. If commands can be routed from
any server to any other server, each server must be defined to all the others.

Setting up communications for command routing with one

source server
The process of setting up communications for command routing can, at times, be a
challenge.

About this task

The example that is provided shows you how you can set up communications for
administrator HQ on the server HEADQUARTERS who routes commands to the
servers MUNICH and STRASBOURG. Administrator HQ has the password
SECRET and has system privilege class.

Procedure

The procedure for setting up communications for command routing with one
source server is shown in the following steps.
v On HEADQUARTERS, register administrator HQ and specify the server names
and addresses of MUNICH and STRASBOURG.
register admin hq secret
grant authority hq classes=system

define server munich serverpassword=beryl hladdress=[Link] lladdress=1919

define server strasbourg serverpassword=fluorite hladdress=[Link]
lladdress=1715

Note: Command routing uses the ID and password of the Administrator. It does
not use the password or server password that is set in the server definition.
v On MUNICH and STRASBOURG, register administrator HQ with the specific
privilege class on each server.
register admin hq secret
grant authority hq classes=system

696 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 Note: If your server network is using enterprise configuration, you can
automate the preceding operations. You can distribute the administrator and
server lists to MUNICH and STRASBOURG. In addition, all server definitions
and server groups are distributed by default to a managed server when it first
subscribes to any profile on a configuration manager. Therefore, it receives all
the server definitions that exist on the configuration manager, thus enabling
command routing among the servers.

Setting up communications for command routing with multiple

source servers
When setting up communications for command routing, you must define all the
servers to each other.

About this task

The examples provided below show you how to set up communications if the
administrator, HQ, can route commands from any of the three servers to any of the
other servers. You can separately define each server to each of the other servers, or
you can “cross define” the servers. In cross definition, defining MUNICH to
HEADQUARTERS also results in automatically defining HEADQUARTERS to
MUNICH.

Creating separate definitions:

When you are setting up communications for command routing, you can define
each server to each of the other servers.

Procedure

To create separate definitions, complete the following steps:

1. On MUNICH, specify the server name and password of MUNICH. Register
administrator HQ and grant HQ system authority.
On STRASBOURG, specify the server name and password of STRASBOURG.
Register administrator HQ and grant HQ system authority.
On HEADQUARTERS, specify the server name and password of
HEADQUARTERS. Register administrator HQ and grant HQ system authority.
2. On HEADQUARTERS, define MUNICH (whose password is BERYL and whose
address is [Link]:1919) and STRASBOURG (whose password is FLUORITE
and whose address is [Link]:1715).
On MUNICH, define HEADQUARTERS (whose password is AMETHYST and
whose address is [Link]:1823) and STRASBOURG.
On STRASBOURG, define HEADQUARTERS and MUNICH.

Results

Figure 83 on page 698 shows the servers and the commands that are issued on
each.

Chapter 21. Managing a network of Tivoli Storage Manager servers 697

 set servername headquarters
set serverpassword amethyst
Headquarters register admin hq secret
grant authority hq classes=system

define server munich

serverpassword=beryl
hladdress=[Link]
lladdress=1919

define server strasbourg

serverpassword=fluorite
hladdress=[Link]
lladdress=1715

Munich Strasbourg

set servername munich set servername strasbourg

set serverpassword beryl set serverpassword fluorite

register admin hq secret register admin hq secret

grant authority hq classes=system grant authority hq classes=system

define server headquarters define server headquarters

serverpassword=amethyst serverpassword=amethyst
hladdress=[Link] hladdress=[Link]
lladdress=1823 lladdress=1823

define server strasbourg define server munich

serverpassword=fluorite serverpassword=beryl
hladdress=[Link] hladdress=[Link]
lladdress=1715 lladdress=1919

Figure 83. Communication configuration with separate server definitions

Creating cross definitions:

When you are setting up communications for command routing, you can
cross-define the other servers.

Procedure

To create cross definitions, complete the following steps:

1. On MUNICH, specify the server name, password, and high and low-level
addresses of MUNICH. Specify that cross define is permitted. Register
administrator HQ and grant HQ system authority.
On STRASBOURG, specify the server name, password, and high and low-level
addresses of STRASBOURG. Specify that cross define is permitted. Register
administrator HQ and grant HQ system authority.

698 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 On HEADQUARTERS, specify the server name, password, and high and
low-level addresses of HEADQUARTERS. Register administrator HQ and grant
HQ system authority.
2. On HEADQUARTERS, define MUNICH and STRASBOURG, specifying that
cross define is to be completed.
3. On MUNICH, define STRASBOURG, specifying that cross define is to be
completed.

Results

Note: If your server network is using enterprise configuration, you can automate
the preceding operations. You can distribute the administrator lists and server lists
to MUNICH and STRASBOURG. In addition, all server definitions and server
groups are distributed by default to a managed server when it first subscribes to
any profile on a configuration manager. Therefore, it receives all the server
definitions that exist on the configuration manager, thus enabling command
routing among the servers.

Figure 84 on page 700 shows the servers and the commands that are issued on
each.

Chapter 21. Managing a network of Tivoli Storage Manager servers 699

 set servername headquarters
set serverpassword amethyst
set serverhladdress [Link]
set serverlladdress 1823
Headquarters
register admin hq secret
grant authority hq classes=system

define server munich crossdefine=yes

serverpassword=beryl
hladdress=[Link]
lladdress=1919

define server strasbourg crossdefine=yes

serverpassword=fluorite
hladdress=[Link]
lladdress=1715

Munich Strasbourg

set servername munich set servername strasbourg

set serverpassword beryl set serverpassword fluorite
set serverhladdress [Link] set serverhladdress [Link]
set serverlladdress 1919 set serverlladdress 1715
set crossdefine on set crossdefine on

register admin hq secret register admin hq secret

grant authority hq classes=system grant authority hq classes=system

define server strasbourg crossdefine=yes

serverpassword=fluorite
hladdress=[Link]
lladdress=1715

Figure 84. Communication configuration with cross definitions

Updating and deleting servers

You can update and delete server definitions for server-to-server virtual volumes,
enterprise configuration, and enterprise event logging.

About this task

You can update a server definition by issuing the UPDATE SERVER command.
v For server-to-server virtual volumes:
– If you update the node name, you must also update the password.
– If you update the password but not the node name, the node name defaults
to the server name specified by the SET SERVERNAME command.
v For enterprise configuration and enterprise event logging: If you update the
server password, it must match the password specified by the SET
SERVERPASSWORD command at the target server.

700 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 v For enterprise configuration: When a server is first defined at a managed server,
that definition cannot be replaced by a server definition from a configuration
manager. This prevents the definition at the managed server from being
inadvertently replaced. Such a replacement could disrupt functions that require
communication among servers, for example command routing or virtual
volumes.
To allow replacement, update the definition at the managed server by issuing
the UPDATE SERVER command with the ALLOWREPLACE=YES parameter.
When a configuration manager distributes a server definition, the definition
always includes the ALLOWREPLACE=YES parameter.

You can delete a server definition by issuing the DELETE SERVER command. For
example, to delete the server named NEWYORK, enter the following:
delete server newyork

The deleted server is also deleted from any server groups of which it is a member.

You cannot delete a server if any of the following conditions are true:
v The server is defined as an event server.
You must first issue the DELETE EVENTSERVER command.
v The server is a target server for virtual volumes.
A target server is named in a DEFINE DEVCLASS (DEVTYPE=SERVER)
command. You must first change the server name in the device class or delete
the device class.
v The server is named in a device class definition whose device type is SERVER.
v The server has paths defined to a file drive.
v The server has an open connection to or from another server.
You can find an open connection to a server by issuing the QUERY SESSION
command.

See “Setting up server groups” on page 728 for information about server groups.

Setting up enterprise configurations

With profiles, you can designate the configuration information that is distributed to
managed servers. Then you can set up other servers as managed servers. The
managed servers receive configuration information through subscriptions to
profiles on the configuration manager.

About this task

Each managed server stores the distributed information as managed objects in its
database. Managed servers receive periodic updates of the configuration
information from the configuration manager, or an administrator can trigger an
update by command.

If you use an LDAP directory server to authenticate passwords, any target servers
must be configured for LDAP passwords. Data that is replicated from a node that
authenticates with an LDAP directory server is inaccessible if the target server is
not properly configured. If your target server is not configured, replicated data
from an LDAP node can still go there. But the target server must be configured to
use LDAP in order for you to access the data.

Chapter 21. Managing a network of Tivoli Storage Manager servers 701

 You can distribute the following configuration information from a configuration
manager to managed servers:
v Administrators, including authorities for them
v Policy objects, including policy domains, and the policy sets, management
classes, copy groups and client schedules associated with them.
v Administrative command schedules
v Tivoli Storage Manager server scripts
v Client option sets
v Server definitions
v Server groups

“Enterprise configuration scenario” gives you an overview of the steps to take for
one possible implementation of enterprise configuration. Sections that follow give
more details on each step. For details on the attributes that are distributed with
these objects, see “Associating configuration information with a profile” on page
707. After you set up server communication as described in “Setting up
communications for enterprise configuration and enterprise event logging” on page
692, you set up the configuration manager and its profiles.

Enterprise configuration scenario

To illustrate how you might use the enterprise configuration functions, suppose
that your enterprise has offices around the world, with one or more Tivoli Storage
Manager servers at each location. To make managing these servers easier, you want
to control the configuration of all Tivoli Storage Manager servers from one Tivoli
Storage Manager server in the headquarters office.

About this task

Figure 85 shows the hierarchy that you want to set up.

Headquarters

Configuration Manager

Managed
servers
London Munich New York Santiago Delhi Tokyo

Strasbourg Mexico City Beijing

Europe Americas Asia

Figure 85. A scenario for implementing enterprise configuration

You want to set up a configuration manager named HEADQUARTERS. Managed

servers have the names of cities where they are located. You have three groups of
702 IBM Tivoli Storage Manager for Windows: Administrator's Guide
 managed servers, one in the Americas, one in Europe, and one in Asia. Each of the
servers supports backup and archive services for client machines in that office. For
client backup operations, you want to use the default policy that stores backups on
disk. Each server has an automated tape library configured to work with Tivoli
Storage Manager, and you want to use the tape library at each location for client
archive operations and for Tivoli Storage Manager server database backups. You
want to be able to monitor activities on all servers. You also want to designate
some other users as administrators who can work with these servers.

The following sections give you an overview of the steps to take to complete this
setup. For details on each step, see the section referenced.

Setting up a configuration manager

You are required to issue specific commands to set up one Tivoli Storage Manager
server as a configuration manager.

About this task

Figure 86 illustrates the commands that you must issue to set up one Tivoli Storage
Manager server as a configuration manager. The following procedure gives you an
overview of the steps required to set up a server as a configuration manager.

Headquarters
set configmanager on
define profile
Configuration define profassocation
Manager

Figure 86. Setting up a configuration manager

Procedure
1. Decide whether to use the existing Tivoli Storage Manager server in the
headquarters office as the configuration manager or to install a new Tivoli
Storage Manager server on a system.
2. Set up the communications among the servers.
3. Identify the server as a configuration manager.
Use the following command:
set configmanager on
This command automatically creates a profile named DEFAULT_PROFILE. The
default profile includes all the server and server group definitions on the
configuration manager. As you define new servers and server groups, they are
also associated with the default profile.
4. Create the configuration to distribute.
The tasks that might be involved include:
v Register administrators and grant authorities to those that you want to be
able to work with all the servers.
v Define policy objects and client schedules
v Define administrative schedules
v Define Tivoli Storage Manager server scripts
v Define client option sets
v Define servers

Chapter 21. Managing a network of Tivoli Storage Manager servers 703

 v Define server groups
Example 1: You need a shorthand way to send commands to different groups
of managed servers. You can define server groups. For example, you can define
a server group named AMERICAS for the servers in the offices in North
America and South America.
Example 2: You want each managed server to back up its database and storage
pools regularly. One way to do this is to set up Tivoli Storage Manager server
scripts and schedules to automatically run these scripts everyday. You can do
the following:
v Verify or define server scripts that perform these operations.
v Verify or define administrative command schedules that run these scripts.
Example 3: You want clients to back up data to the default disk storage pool,
BACKUPPOOL, on each server. But you want clients to archive data directly to
the tape library attached to each server. You can do the following:
v In the policy domain that you will point to in the profile, update the archive
copy group so that TAPEPOOL is the name of the destination storage pool.
v On each server that is to be a managed server, ensure that you have a tape
storage pool named TAPEPOOL.

Note: You must set up the storage pool itself (and associated device class) on
each managed server, either locally or by using command routing. If a
managed server already has a storage pool associated with the automated
tape library, you can rename the pool to TAPEPOOL.
Example 4: You want to ensure that client data is consistently backed up and
managed on all servers. You want all clients to be able to store three backup
versions of their files. You can do the following:
v Verify or define client schedules in the policy domain so that clients are
backed up on a consistent schedule.
v In the policy domain that you will point to in the profile, update the backup
copy group so that three versions of backups are allowed.
v Define client option sets so that basic settings are consistent for clients as
they are added.
5. Define one or more profiles.
For example, you can define one profile named ALLOFFICES that points to all
the configuration information (policy domain, administrators, scripts, and so
on). You can also define profiles for each type of information, so that you have
one profile that points to policy domains, and another profile that points to
administrators, for example.
See “Setting up communications among servers” on page 692 for details. For
more information, see “Creating the default profile on a configuration
manager” on page 706. See “Defining a server group and members of a server
group” on page 729 for details. For details, see “Creating and changing
configuration profiles” on page 707.

704 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 Setting up a managed server
Setting up the managed server can be done by an administrator working at a
central location, or by administrators working at the servers that will be managed
servers.

About this task

Figure 87 shows the specific commands needed to set up one Tivoli Storage
Manager server as a managed server. The following procedure gives you an
overview of the steps required to set up a server as a managed server.

query profile
define subscription
set configrefresh
Managed
server
Munich

Figure 87. Setting up a managed server

A server becomes a managed server when that server first subscribes to a profile
on a configuration manager.

Procedure
1. Query the server to look for potential conflicts.
Look for definitions of objects on the managed server that have the same name
as those defined on the configuration manager. With some exceptions, these
objects will be overwritten when the managed server first subscribes to the
profile on the configuration manager.
If the managed server is a new server and you have not defined anything, the
only objects you will find are the defaults (for example, the STANDARD policy
domain).
2. Subscribe to one or more profiles.
A managed server can only subscribe to profiles on one configuration manager.
If you receive error messages during the configuration refresh, such as a local
object that could not be replaced, resolve the conflict and refresh the
configuration again. You can either wait for the automatic refresh period to be
reached, or kick off a refresh by issuing the SET CONFIGREFRESH command,
setting or resetting the interval.
3. If the profile included policy domain information, activate a policy set in the
policy domain, add or move clients to the domain, and associate any required
schedules with the clients.
You may receive warning messages about storage pools that do not exist, but
that are needed for the active policy set. Define any storage pools needed by
the active policy set, or rename existing storage pools.
4. If the profile included administrative schedules, make the schedules active.
Administrative schedules are not active when they are distributed by a
configuration manager. The schedules do not run on the managed server until
you make them active on the managed server. See “Tailoring schedules” on
page 621.

Chapter 21. Managing a network of Tivoli Storage Manager servers 705

 5. Set how often the managed server contacts the configuration manager to
update the configuration information associated with the profiles.
The initial setting for refreshing the configuration information is 60 minutes.
For more information, see the following topics:
v “Associating configuration information with a profile” on page 707
v “Defining storage pools” on page 245
v “Getting information about profiles” on page 714
v “Refreshing configuration information” on page 721
v “Renaming storage pools” on page 409
v “Subscribing to a profile” on page 717

Creating the default profile on a configuration manager

To set up one Tivoli Storage Manager server as the source for configuration
information for other servers, identify the server as a configuration manager. A
configuration manager can be an existing Tivoli Storage Manager server that
already provides services to clients, or can be a server dedicated to just providing
configuration information to other Tivoli Storage Manager servers.

About this task

Task Required Privilege Class
Set up a server as a configuration manager System

Issue the following command:

set configmanager on

When a server becomes a configuration manager, the server automatically creates a

default profile named DEFAULT_PROFILE. The default profile contains any
definitions of servers and server groups that exist on the configuration manager.
You can change or delete the profile named DEFAULT_PROFILE.

When a managed server first subscribes to a profile on a configuration manager,

the configuration manager automatically also subscribes the managed server to the
profile named DEFAULT_PROFILE, if it exists. The information distributed via this
profile gets refreshed in the same way as other profiles. This helps ensure that all
servers have a consistent set of server and server group definitions for all servers
in the network.

If you do not change the DEFAULT_PROFILE, whenever a managed server

subscribed to the DEFAULT_PROFILE profile refreshes configuration information,
the managed server receives definitions for all servers and server groups that exist
on the configuration manager at the time of the refresh. As servers and server
groups are added, deleted, or changed on the configuration manager, the changed
definitions are distributed to subscribing managed servers.

If the configuration manager is on a V6.3.3 or later server, password authentication

methods can cause conflicts between the configuration manager and the managed
servers. Administrator IDs with LDAP-authenticated passwords on a V6.3.3 or later
server cannot be accessed on a managed server that is on a pre-V6.3.3 server.

706 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Creating and changing configuration profiles
You create configuration profiles on a configuration manager, which distributes the
information that is associated with the profiles to any managed server that
subscribes to those profiles.

Procedure

Create a configuration profile by completing the following steps:

1. Defining the profile.
2. Associating the configuration information with the profile.

Results
After you define the profile and its associations, a managed server can subscribe to
the profile and obtain the configuration information.

After you define a profile and associate information with the profile, you can
change the information later. While you make changes, you can lock the profiles to
prevent managed servers from refreshing their configuration information. To
distribute the changed information that is associated with a profile, you can unlock
the profile, and either wait for each managed server to refresh its configuration to
get the changed information or notify each managed server to refresh its
configuration. The following sections provide information on each of these tasks.

Defining the profile

When you define the profile, you select the name and can include a description.

About this task

Task Required Privilege Class
Define profiles System

For example, to define a profile named ALLOFFICES, issue the following

command:
define profile alloffices
description=’Configuration to be used by all offices’

Associating configuration information with a profile

After you define a profile, you associate the configuration information that you
want to distribute via that profile.

About this task

Task Required Privilege Class
Define profile associations System

You can associate the following configuration information with a profile:

v Tivoli Storage Manager administrators, including their authorities.
v Policy domains.
v Servers definitions.
v Server groups.
v Administrative command schedules.

Chapter 21. Managing a network of Tivoli Storage Manager servers 707

 v Tivoli Storage Manager server scripts.
v Client option sets.

Before you can associate specific configuration information with a profile, the
definitions must exist on the configuration manager. For example, to associate a
policy domain named ENGDOMAIN with a profile, you must have already
defined the ENGDOMAIN policy domain on the configuration manager.

Suppose you want the ALLOFFICES profile to distribute policy information from
the STANDARD and ENGDOMAIN policy domains on the configuration manager.
Enter the following command:
define profassociation alloffices domains=standard,engdomain

You can make the association more dynamic by specifying the special character, *
(asterisk), by itself. When you specify the *, you can associate all existing objects
with a profile without specifically naming them. If you later add more objects of
the same type, the new objects are automatically distributed via the profile. For
example, suppose that you want the ADMINISTRATORS profile to distribute all
administrators registered to the configuration manager. Enter the following
commands on the configuration manager:
define profile administrators
description=’Profile to distribute administrators IDs’

define profassociation administrators admins=*

Whenever a managed server that is subscribed to the ADMINISTRATORS profile

refreshes configuration information, it receives definitions for all administrators
that exist on the configuration manager at the time of the refresh. As
administrators are added, deleted, or changed on the configuration manager, the
changed definitions are distributed to subscribing managed servers.

For more information, see the following topics:

v “Configuring information for administrative command schedules” on page 711
v “Configuring information for policy domains” on page 709
v “Configuring information for servers and server groups” on page 710
v “Configuration information for Tivoli Storage Manager administrators”
v “Tivoli Storage Manager server scripts” on page 626
v “Managing client option sets” on page 467

Configuration information for Tivoli Storage Manager

administrators
Be careful if you are distributing definitions of administrators that have the same
name as administrators already defined to managed servers. The configuration
refresh overwrites the administrator definition and authority defined on the
managed server.

If the authority level of an administrator is less on the configuration manager than

it was on the managed server, you might have problems with access to the
managed server after distributing the administrator definition.

The configuration manager does not distribute information about whether an

administrator is locked (preventing access to the server).

708 IBM Tivoli Storage Manager for Windows: Administrator's Guide

The administrator with the name SERVER_CONSOLE is never distributed from the
configuration manager to a managed server.

For administrator definitions that have node authority, the configuration manager
only distributes information such as password and contact information. Node
authority for the managed administrator can be controlled on the managed server
using the GRANT AUTHORITY and REVOKE AUTHORITY commands specifying
the CLASS=NODE parameter.

Configuring information for policy domains

When you point to a policy domain in a profile, the configuration information that
will be sent to the managed servers includes the policy domain itself and all policy
sets with their associated management classes, copy groups, and client schedules in
the domain.

About this task

A configuration manager does not distribute the following:

v An active policy set and any of its associated management classes, copy groups,
and client schedules. On each managed server, you must activate a policy set in
each managed policy domain.
v Associations between clients and schedules. To have clients in a managed policy
domain run client schedules, you must associate the clients with the schedules
on the managed server.
v Client actions, which are schedules created by using the DEFINE
CLIENTACTION command. On each managed server, you can define and delete
client actions, even if the corresponding domain is a managed object.
v Definitions for any storage pools identified as destinations in the policy.
Definitions of storage pools and device classes are not distributed by a
configuration manager.
Policy domains can refer to storage pool names in the management classes,
backup copy groups, and archive copy groups. As you set up the configuration
information, consider whether managed servers already have or can set up or
rename storage pools with these names.

A subscribing managed server may already have a policy domain with the same
name as the domain associated with the profile. The configuration refresh
overwrites the domain defined on the managed server unless client nodes are
already assigned to the domain. Once the domain becomes a managed object on
the managed server, you can associate clients with the managed domain. Future
configuration refreshes can then update the managed domain.

If nodes are assigned to a domain with the same name as a domain being
distributed, the domain is not replaced. This safeguard prevents inadvertent
replacement of policy that could lead to loss of data. To replace an existing policy
domain with a managed domain of the same name, perform the following steps on
the managed server:

Procedure
1. Copy the domain.
2. Move all clients assigned to the original domain to the copied domain.
3. Trigger a configuration refresh.
4. Activate the appropriate policy set in the new, managed policy domain.
5. Move all clients back to the original domain, which is now managed.
Chapter 21. Managing a network of Tivoli Storage Manager servers 709
 Configuring information for servers and server groups
The DEFAULT_PROFILE that is automatically created on a configuration manager
already points to all servers and server groups defined to that server. If you leave
the DEFAULT_PROFILE intact, you do not need to include servers or server
groups in any other profile.

About this task

Any servers and server groups that you define later are associated automatically
with the default profile and the configuration manager distributes the definitions at
the next refresh. For a server definition, the following attributes are distributed:
v Communication method
v TCP/IP address (high-level address), Version 4 or Version 6
v Port number (low-level address)
v Server password
v Server URL
v The description
When server definitions are distributed, the attribute for allowing replacement is
always set to YES. You can set other attributes, such as the server's node name, on
the managed server by updating the server definition.

A managed server may already have a server defined with the same name as a
server associated with the profile. The configuration refresh does not overwrite the
local definition unless the managed server allows replacement of that definition.
On a managed server, you allow a server definition to be replaced by updating the
local definition. For example:
update server santiago allowreplace=yes

This safeguard prevents disruption of existing functions that require

communication among servers (such as virtual volumes).

Table 62 summarizes what happens when servers or server groups being

distributed have the same names as servers or server groups on the managed
server.
Table 62. Results of configuration refresh with duplicate object names
Local definition (on Object with duplicate
managed server) name to be distributed Result of configuration refresh
Server Server The local server definition is
replaced by the distributed server
definition only if an administrator
for the managed server updated
the local definition to allow
replacement.
Server Server group The local server definition remains.
The server group definition is not
distributed.
Server group Server The local server group is deleted.
The server definition is distributed.
Server group Server group The local server group definition is
replaced by the distributed server
group definition.

710 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Configuring information for administrative command schedules
When the configuration manager distributes administrative schedules, the
schedules are not active on the managed server. An administrator on the managed
server must activate any managed schedules to have them run on the managed
server.

About this task

A configuration refresh does not replace or remove any local schedules that are
active on a managed server. However, a refresh can update an active schedule that
is already managed by a configuration manager.

Changing a profile
You can change a profile and its associated configuration information.

About this task

Task Required Privilege Class
Define profile associations System

Update profiles System

For example, if you want to add a policy domain named FILESERVERS to objects
already associated with the ALLOFFICES profile, enter the following command:
define profassociation alloffices domains=fileservers

You can also delete associated configuration information, which results in removal
of configuration from the managed server. Use the DELETE PROFASSOCIATION
command.

On a configuration manager, you cannot directly change the names of

administrators, scripts, and server groups associated with a profile. To change the
name of an administrator, script, or server group associated with a profile, delete
the object then define it again with a new name and associate it with the profile
again. During the next configuration refresh, each managed server makes the
corresponding changes in their databases.

You can change the description of the profile. Enter the following command:
update profile alloffices
description=’Configuration for all offices with file servers’

See “Removing configuration information from managed servers” on page 712 for
details.

Preventing access to profiles while you make changes

If you are making changes to a profile, you might want to prevent any subscribing
managed server from refreshing its configuration information until you are done.
You can lock the profile to prevent access to the profile by a managed server.

About this task

Locking prevents a managed server from getting information that is incomplete

because you are still making changes.

Chapter 21. Managing a network of Tivoli Storage Manager servers 711

 Task Required Privilege Class
Lock and unlock profiles System

For example, to lock the ALLOFFICES profile for two hours (120 minutes), enter
the following command:
lock profile alloffices 120

You can let the lock expire after two hours, or unlock the profile with the following
command:
unlock profile alloffices

Distributing changed configuration information

To distribute the changed profile, wait for each managed server to refresh its
configuration to get the changed information, or notify each managed server from
the configuration manager. Managed servers refresh profile information on a
configuration refresh period.

About this task

Task Required Privilege Class
Notify servers that subscribe to profiles to System
refresh configuration information

From the configuration manager, to notify all servers that are subscribers to the
ALLOFFICES profile, enter the following command:
notify subscribers profile=alloffices

The managed servers then refresh their configuration information, even if the time
period for refreshing the configuration has not passed.

See “Refreshing configuration information” on page 721 for how to set this period.

Removing configuration information from managed servers

To remove configuration information from managed servers, you can delete the
association of the object with the profile, or delete the object itself from the
configuration manager.

About this task

Task Required Privilege Class
Delete profile associations System

Note: To remove all configuration information that is defined in the database of a

managed server as a result of a profile subscription, you must delete the
subscription using the option to discard all managed objects.

On the configuration manager, you can delete the association of objects with a
profile. For example, you may want to remove some of the administrators that are
associated with the ADMINISTRATORS profile. With an earlier command, you had
included all administrators defined on the configuration manager (by specifying
ADMINS=*). To change the administrators included in the profile you must first
delete the association of all administrators, then associate just the administrators

712 IBM Tivoli Storage Manager for Windows: Administrator's Guide

that you want to include. Do the following:

Procedure
1. Before you make these changes, you may want to prevent any servers from
refreshing their configuration until you are done. Enter the following
command:
lock profile administrators
2. Now make the change by entering the following commands:
delete profassociation administrators admins=*

define profassociation administrators

admins=admin1,admin2,admin3,admin4
3. Unlock the profile:
unlock profile administrators
4. You may want to notify any managed server that subscribes to the profile so
that servers refresh their configuration information:
notify subscribers profile=administrators

Results

When you delete the association of an object with a profile, the configuration
manager no longer distributes that object via the profile. Any managed server
subscribing to the profile deletes the object from its database when it next contacts
the configuration manager to refresh configuration information. However, a
managed server does not delete the following objects:
v An object that is associated with another profile to which the server subscribes.
v A policy domain that has client nodes still assigned to it. To delete the domain,
you must assign the affected client nodes to another policy domain on the
managed server.
v An administrator that currently has a session open with the server.
v An administrator that is the last administrator with system authority on the
managed server.
Also the managed server does not change the authority of an administrator if
doing so would leave the managed server without any administrators having
the system privilege class.
You can avoid both problems by ensuring that you have locally defined at least
one administrator with system privilege on each managed server.
v An administrative schedule that is active. To remove an active schedule, you
must first make the schedule inactive on the managed server.
v A server definition for a server that currently has an open connection from the
managed server.
v A server definition that is specified in the definition of a device class that is a
SERVER device type.
v A server definition that is the definition for the event server for the managed
server.

If you no longer need an object defined on the configuration manager itself or on

any managed server, you can delete the object itself. Deleting the object itself from
the configuration manager has an effect similar to deleting the association of that
object with the profile. The configuration manager no longer distributes that object,
and a managed server attempts to delete the object from its database when it
refreshes configuration information.

Chapter 21. Managing a network of Tivoli Storage Manager servers 713

 See “Deleting subscriptions” on page 720.

Deleting profiles
You can delete a profile from a configuration manager. Before deleting a profile,
you should ensure that no managed server still has a subscription to the profile. If
the profile still has some subscribers, delete the subscriptions on each managed
server first.

About this task

Task Required Privilege Class
Delete profiles System

When you delete subscriptions, consider whether you want the managed objects to
be deleted on the managed server at the same time. For example, to delete the
subscription to profile ALLOFFICES from managed server SANTIAGO without
deleting the managed objects, log on to the SANTIAGO server and enter the
following command:
delete subscription alloffices

Then, on the configuration manager, enter the following command:

delete profile alloffices

Note: You can use command routing to issue the DELETE SUBSCRIPTION
command for all managed servers.

If you try to delete a profile, that still has subscriptions, the command fails unless
you force the operation:
delete profile alloffices force=yes

If you do force the operation, managed servers that still subscribe to the deleted
profile will later contact the configuration manager to try to get updates to the
deleted profile. The managed servers will continue to do this until their
subscriptions to the profile are deleted. A message will be issued on the managed
server alerting the administrator of this condition.

See “Deleting subscriptions” on page 720 for more details about deleting
subscriptions on a managed server.

Getting information about profiles

You can get information about configuration profiles defined on any configuration
manager, as long as that server is defined to the server with which you are
working.

About this task

Task Required Privilege Class
Request information about profiles Any administrator

For example, from a configuration manager, you can display information about
profiles defined on that server or on another configuration manager. From a
managed server, you can display information about any profiles on the
configuration manager to which the server subscribes. You can also get profile

714 IBM Tivoli Storage Manager for Windows: Administrator's Guide

information from any other configuration manager defined to the managed server,
even though the managed server does not subscribe to any of the profiles.

For example, to get information about all profiles on the HEADQUARTERS

configuration manager when logged on to another server, enter the following
command:
query profile server=headquarters

The following shows what the results might look like:

Configuration Profile name Locked?
manager
--------------- --------------- -------
HEADQUARTERS ADMINISTRATORS No
HEADQUARTERS DEFAULT_PROFILE No
HEADQUARTERS ENGINEERING No
HEADQUARTERS MARKETING No

You may need to get detailed information about profiles and the objects associated
with them, especially before subscribing to a profile. You can get the names of the
objects associated with a profile by entering the following command:
query profile server=headquarters format=detailed

The following shows what the results might look like:

Chapter 21. Managing a network of Tivoli Storage Manager servers 715

 Configuration manager: HEADQUARTERS
Profile name: ADMINISTRATORS
Locked?: No
Description:
Server administrators: ADMIN1 ADMIN2 ADMIN3 ADMIN4
Policy domains:
Administrative command schedules: ** all objects **
Server Command Scripts:
Client Option Sets:
Servers:
Server Groups:

Configuration manager: HEADQUARTERS

Profile name: DEFAULT_PROFILE
Locked?: No
Description:
Server administrators:
Policy domains:
Administrative command schedules:
Server Command Scripts:
Client Option Sets:
Servers: ** all objects **
Server Groups: ** all objects **

Configuration manager: HEADQUARTERS

Profile name: ENGINEERING
Locked?: No
Description:
Server administrators:
Policy domains: ENGDOMAIN
Administrative command schedules:
Server Command Scripts: QUERYALL
Client Option Sets: DESIGNER PROGRAMMER
Servers:
Server Groups:

Configuration manager: HEADQUARTERS

Profile name: MARKETING
Locked?: Yes
Description:
Server administrators:
Policy domains: MARKETDOM
Administrative command schedules:
Server Command Scripts: QUERYALL
Client Option Sets: BASIC
Servers:
Server Groups:

If the server from which you issue the query is already a managed server
(subscribed to one or more profiles on the configuration manager being queried),
by default the query returns profile information as it is known to the managed
server. Therefore the information is accurate as of the last configuration refresh
done by the managed server. You may want to ensure that you see the latest
version of profiles as they currently exist on the configuration manager. Enter the
following command:
query profile uselocal=no format=detailed

To get more than the names of the objects associated with a profile, you can do one
of the following:
v If command routing is set up between servers, you can route query commands
from the server to the configuration manager. For example, to get details on the
ENGDOMAIN policy domain on the HEADQUARTERS server, enter this
command:
headquarters: query domain engdomain format=detailed

716 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 You can also route commands from the configuration manager to another server
to get details about definitions that already exist.
v If command routing is not set up, log on to the configuration manager and enter
the query commands to get the information you need.

Subscribing to a profile
After an administrator at a configuration manager has created profiles and
associated objects with them, managed servers can subscribe to one or more of the
profiles.

About this task

Task Required Privilege Class
Define subscriptions to profiles System

Set the period for configuration refreshes System

Note:
v Unless otherwise noted, the commands in this section would be run on a
managed server:
v An administrator at the managed server could issue the commands.
v You could log in from the enterprise console and issue them.
v If command routing is set up, you could route them from the server that you are
logged in to.

After a managed server subscribes to a profile, the configuration manager sends

the object definitions associated with the profile to the managed server where they
are automatically stored in the database. Object definitions created this way in the
database of a managed server are called managed objects. With a few exceptions,
you cannot change managed objects on the managed server. The exceptions are
that you can change:
v The active status of a schedule
v The lock status of an administrator
v Which policy set is active in a policy domain
v The default management class of a policy set
v The attributes of a server definition that are related to the use of virtual volumes
(node name, password, and delete grace period)

Before a managed server subscribes to a profile, be aware that if you have defined
any object with the same name and type as an object associated with the profile
that you are subscribing to, those objects will be overwritten. You can check for
such occurrences by querying the profile before subscribing to it.

When a managed server first subscribes to a profile on a configuration manager, it

also automatically subscribes to DEFAULT_PROFILE, if a profile with this name is
defined on the configuration manager. Unless DEFAULT_PROFILE is modified on
the configuration manager, it contains all the server definitions and server groups
defined on the configuration manager. In this way, all the servers in your network
receive a consistent set of server and server group definitions.

Chapter 21. Managing a network of Tivoli Storage Manager servers 717

 Note: Although a managed server can subscribe to more than one profile on a
configuration manager, it cannot subscribe to profiles on more than one
configuration manager at a time.

Changes can be made to a profile, after a managed server subscribes to it. An

administrator on the configuration manager can notify your server of a change by
issuing the NOTIFY SUBSCRIBERS command. The configuration manager contacts
each managed server having a subscription to one of the specified profiles. When a
managed server is contacted, it begins refresh processing to get the configuration
updates from the configuration manager.

Subscription scenario
The scenario that is documented is a typical one, where a server subscribes to a
profile on a configuration manager, in this case HEADQUARTERS.

About this task

In this scenario an administrator for the HEADQUARTERS server has defined

three profiles, ADMINISTRATORS, ENGINEERING, and MARKETING, each with
its own set of associations. In addition, DEFAULT_PROFILE was automatically
defined and contains only the server and server group definitions defined on the
HEADQUARTERS server. An administrator for HEADQUARTERS has given you
the names of the profiles that you should be using. To subscribe to the
ADMINISTRATORS and ENGINEERING profiles and keep them current, perform
the following steps:

Procedure
1. Display the names of the objects in the profiles on HEADQUARTERS.
You might want to perform this step to see if the object names on the profiles
are used on your server for any objects of the same type. Issue this command:
query profile * server=headquarters format=detailed

You might want to get detailed information on some of the objects by issuing
specific query commands on either your server or the configuration manager.

Note: If any object name matches and you subscribe to a profile containing an
object with the matching name, the object on your server will be replaced, with
the following exceptions:
v A policy domain is not replaced if the domain has client nodes assigned to it.
v An administrator with system authority is not replaced by an administrator
with a lower authority level if the replacement would leave the server
without a system administrator.
v The definition of a server is not replaced unless the server definition on the
managed server allows replacement.
v A server with the same name as a server group is not replaced.
v A locally defined, active administrative schedule is not replaced
2. Subscribe to the ADMINISTRATORS and ENGINEERING profiles.
After the initial subscription, you do not have to specify the server name on the
DEFINE SUBSCRIPTION commands. If at least one profile subscription already
exists, any additional subscriptions are automatically directed to the same
configuration manager. Issue these commands:
define subscription administrators server=headquarters

define subscription engineering

718 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 The object definitions in these profiles are now stored on your database. In
addition to ADMINISTRATORS and ENGINEERING, the server is also
subscribed by default to DEFAULT_PROFILE. This means that all the server
and server group definitions on HEADQUARTERS are now also stored in your
database.
3. Set the time interval for obtaining refreshed configuration information from the
configuration manager.
If you do not perform this step, your server checks for updates to the profiles
at start up and every 60 minutes after that. Set up your server to check
HEADQUARTERS for updates once a day (every 1440 minutes). If there is an
update, HEADQUARTERS sends it to the managed server automatically when
the server checks for updates.
set configrefresh 1440

Results

Note: You can initiate a configuration refresh from a managed server at any time.
To initiate a refresh, simply reissue the SET CONFIGREFRESH with any value
greater than 0. The simplest approach is to use the current setting:
set configrefresh 1440

Querying subscriptions
From time to time you might want to view the profiles to which a server is
subscribed. You might also want to view the last time that the configuration
associated with that profile was successfully refreshed on your server.

About this task

Task Required Privilege Class
Request information about subscriptions Any administrator

Request information about profiles Any administrator

The QUERY SUBSCRIPTION command gives you this information. You can name
a specific profile or use a wildcard character to display all or a subset of profiles to
which the server is subscribed. For example, the following command displays
ADMINISTRATORS and any other profiles that begin with the string “ADMIN”:
query subscription admin*

Here is a sample of the output:

Configuration Profile name Last update
manager date/time
--------------- --------------- --------------------
HEADQUARTERS ADMINISTRATORS 06/04/2002 [Link]
HEADQUARTERS ADMINS_1 06/04/2002 [Link]
HEADQUARTERS ADMINS_2 06/04/2002 [Link]

To see what objects the ADMINISTRATORS profile contains, use the following
command:
query profile administrators uselocal=no format=detailed

You will see output similar to the following:

Chapter 21. Managing a network of Tivoli Storage Manager servers 719

 Configuration manager: HEADQUARTERS
Profile name: ADMINISTRATORS
Locked?: No
Description:
Server administrators: ADMIN1 ADMIN2 ADMIN3 ADMIN4
Policy domains:
Administrative command schedules: ** all objects **
Server Command Scripts:
Client Option Sets:
Servers:
Server Groups:

Managed objects are stored in the database of a managed server as a result of

subscriptions to profiles on a configuration manager. Any object that was created
or updated in the database of the managed server as a result of a subscription has
the string $$CONFIG_MANAGER$$ in place of the name of the administrator who
last changed the object. For example, if the policy domain named ENGDOMAIN is
a managed object and you enter this command on the managed server:
query domain engdomain format=detailed

You will see output similar to the following:

Policy Domain Name: ENGDOMAIN
Activated Policy Set:
Activation Date/Time:
Days Since Activation:
Activated Default Mgmt Class:
Number of Registered Nodes: 0
Description: Policy for design and software engineers
Backup Retention (Grace Period): 30
Archive Retention (Grace Period): 365
Last Update by (administrator): $$CONFIG_MANAGER$$
Last Update Date/Time: 06/04/2002 [Link]
Managing profile: ENGINEERING

The field Managing profile shows the profile to which the managed server
subscribes to get the definition of this object.

Deleting subscriptions
If you decide that a server no longer needs to subscribe to a profile, you can delete
the subscription.

About this task

Task Required Privilege Class
Delete subscriptions to profiles System

When you delete a subscription to a profile, you can choose to discard the objects
that came with the profile or keep them in your database. For example, to request
that your subscription to PROFILEC be deleted and to keep the objects that came
with that profile, issue the following command:
delete subscription profilec discardobjects=no

After the subscription is deleted on the managed server, the managed server issues
a configuration refresh request to inform the configuration manager that the
subscription is deleted. The configuration manager updates its database with the
new information.

720 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 When you choose to delete objects when deleting the subscription, the server may
not be able to delete some objects. For example, the server cannot delete a
managed policy domain if the domain still has client nodes registered to it. The
server skips objects it cannot delete, but does not delete the subscription itself. If
you take no action after an unsuccessful subscription deletion, at the next
configuration refresh the configuration manager will again send all the objects
associated with the subscription. To successfully delete the subscription, do one of
the following:
v Fix the reason that the objects were skipped. For example, reassign clients in the
managed policy domain to another policy domain. After handling the skipped
objects, delete the subscription again.
v Delete the subscription again, except this time do not discard the managed
objects. The server can then successfully delete the subscription. However, the
objects that were created because of the subscription remain.

Refreshing configuration information

On a configuration manager, an administrator can make changes to configuration
information that is associated with a profile. How quickly the changes get
distributed to a subscribing managed server depends on the configuration refresh
period set on the managed server and whether the administrator on the
configuration manager sent a notification.

About this task

Task Required Privilege Class
Set the period for configuration refreshes System (on the managed server)
Notify servers that subscribe to profiles to System (on the configuration manager)
refresh configuration information

By default, a managed server refreshes its configuration information every 60

minutes. To cause an immediate refresh, change this period. For example, to
immediately refresh the configuration and change the frequency of future refreshes
to once a day, enter the following command for the managed server:
set configrefresh 1440

By issuing this command with a value greater than zero, you cause the managed
server to immediately start the refresh process.

At the configuration manager, you can cause managed servers to refresh their
configuration information by notifying the servers. For example, to notify
subscribers to all profiles, enter the following command:
notify subscribers profile=*

The managed servers then start to refresh configuration information to which they
are subscribed through profiles.

A managed server automatically refreshes configuration information when it is

restarted.

Chapter 21. Managing a network of Tivoli Storage Manager servers 721

 Managing problems with configuration refresh
To monitor for any problems during a configuration refresh, watch the server
console or activity log of the managed server. One problem that may occur is that
the refresh process can skip objects. For example, a policy domain of the same
name as an existing policy domain on the managed server is not distributed if the
policy domain has client nodes assigned to it.

About this task

The configuration manager sends the objects that it can distribute to the managed
server. The configuration manager skips (does not send) objects that conflict with
local objects. If the configuration manager cannot send all objects that are
associated with the profile, the managed server does not record the configuration
refresh as complete. The objects that the configuration manager successfully sent
are left as local instead of managed objects in the database of the managed server.
The local objects left as a result of an unsuccessful configuration refresh become
managed objects at the next successful configuration refresh of the same profile
subscription.

See “Associating configuration information with a profile” on page 707 for details
on when objects cannot be distributed.

Returning managed objects to local control

You might want to return one or more managed objects (objects distributed by a
configuration manager via profiles) to local control on the managed servers. You
can accomplish this from the configuration manager or from the managed servers.

About this task

To do this from the configuration manager, you do not simply delete the
association of the object from the profile, because that would cause the object to be
deleted from subscribing managed servers. To ensure the object remains in the
databases of the managed servers as a locally managed object, you can copy the
current profile, make the deletion, and change the subscriptions of the managed
servers to the new profile.

For example, servers are currently subscribed to the ENGINEERING profile. The
ENGDOMAIN policy domain is associated with this profile. You want to return
control of the ENGDOMAIN policy domain to the managed servers. You can do
the following:

Procedure
1. Copy the ENGINEERING profile to a new profile, ENGINEERING_B:
copy profile engineering engineering_b
2. Delete the association of the ENGDOMAIN policy domain from
ENGINEERING_B:
delete profassociation engineering_b domains=engdomain
3. Use command routing to delete subscriptions to the ENGINEERING profile:
americas,europe,asia: delete subscription engineering
discardobjects=no
4. Delete the ENGINEERING profile:
delete profile engineering

722 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 5. Use command routing to define subscriptions to the new ENGINEERING_B
profile:
americas,europe,asia: define subscription engineering_b

Results

To return objects to local control when working on a managed server, you can
delete the subscription to one or more profiles. When you delete a subscription,
you can choose whether to delete the objects associated with the profile. To return
objects to local control, you do not delete the objects. For example, use the
following command on a managed server:
delete subscription engineering discardobjects=no

Setting up administrators for the servers

Include any administrators, in your profiles, to whom you want to give access to
all servers in the network. These administrators must then maintain their
passwords on the configuration manager.

About this task

To ensure passwords stay valid for as long as expected on all servers, set the
password expiration period to the same time on all servers. One way to do this is
to route a SET PASSEXP command from one server to all of the others.

Ensure that you have at least one administrator that is defined locally on each
managed server with system authority. This avoids an error on configuration
refresh when all administrators for a server would be removed as a result of a
change to a profile on the configuration manager.

Managing problems with synchronization of profiles

In rare situations when a managed server contacts a configuration manager to
refresh configuration information, the configuration manager might determine that
the profile information on the two servers is not synchronized

About this task

It might appear that the configuration information is more recent on the managed
server than on the configuration manager. This could occur in the following
situations:
v The database on the configuration manager has been restored to an earlier time
and now has configuration information from profiles that appear to be older
than what the managed server has obtained.
v On the configuration manager, an administrator deleted a profile, forcing the
deletion even though one or more managed servers still subscribed to the
profile. The administrator redefined the profile (using the same name) before the
managed server refreshed its configuration information.

If the configuration manager still has a record of the managed server's subscription
to the profile, the configuration manager does not send its profile information at
the next request for refreshed configuration information. The configuration
manager informs the managed server that the profiles are not synchronized. The
managed server then issues a message indicating this condition so that an

Chapter 21. Managing a network of Tivoli Storage Manager servers 723

 administrator can take appropriate action. The administrator can perform the
following steps:

Procedure
1. If the configuration manager's database has been restored to an earlier point in
time, the administrator may want to query the profile and associated objects on
the managed server and then manually update the configuration manager with
that information.
2. Use the DELETE SUBSCRIPTION command on the managed server to delete
subscriptions to the profile that is not synchronized. If desired, you can also
delete definitions of the associated objects, then define the subscription again.

Results

It is possible that the configuration manager may not have a record of the
managed server's subscription. In this case, no action is necessary. When the
managed server requests a refresh of configuration information, the configuration
manager sends current profile information and the managed server updates its
database with that information.

Switching a managed server to a different configuration manager

You might want to switch a managed server from one configuration manager to
another to organize your policy needs.

Procedure
Complete the following steps to switch a managed server.
1. Query profiles on the server that is the new configuration manager to compare
with current profiles to which the managed server subscribes.
2. On the managed server, delete all subscriptions to profiles on the current
configuration manager. Remember to delete the subscription to the profile
named DEFAULT_PROFILE. Consider whether to discard the managed objects
in the database when you delete the subscriptions.
Verify that all subscriptions are deleted by querying subscriptions.
3. Change server communications as needed. Define the server that is the new
configuration manager. You can delete the server that was formerly the
configuration manager.
4. On the managed server, define subscriptions to profiles on the new
configuration manager.

Deleting subscribers from a configuration manager

Under normal circumstances, you do not have to delete subscribers from a
configuration manager. You only have to delete a subscription to a profile on the
managed server (by using the DELETE SUBSCRIPTION command).

About this task

When you issue the DELETE SUBSCRIPTION command, the managed server
automatically notifies the configuration manager of the deletion by refreshing its
configuration information. As part of the refresh process, the configuration
manager is informed of the profiles to which the managed server subscribes and to
which it does not subscribe. If the configuration manager cannot be contacted

724 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 immediately for a refresh, the configuration manager will find out that the
subscription was deleted the next time the managed server refreshes configuration
information.

Deleting subscribers from a configuration manager is only necessary as a way to

clean up in certain unusual situations. For example, you may need to delete
subscribers if a managed server goes away completely or deletes its last
subscription without being able to notify the configuration manager. You then use
the DELETE SUBSCRIBER command to delete all subscriptions for that subscriber
(the managed server) from the configuration manager's database.

Renaming a managed server

You might want to rename a managed server to align your policy configuration.

Procedure

To rename a managed server, complete the following steps:

1. Change the name of the managed server by using command routing or by
logging on to the managed server. Use the enterprise console or use the SET
SERVERNAME command.
2. Change the communication setup.
a. On the configuration manager, delete the server definition with the old
name.
b. On the configuration manager, define the server with its new name.
3. On the managed server, refresh the configuration information. You can wait for
the configuration refresh period to pass, or you can reset the refresh period to
cause an immediate refresh.

Results

See “Setting the server name” on page 613 for more information before you use
theSET SERVERNAME command.

Completing tasks on multiple servers

To easily complete tasks with multiple servers, Tivoli Storage Manager provides
you with the Operations Center, command routing, and server group definitions
that you can use to simplify command routing.

Managing multiple servers through a web-based interface

Tivoli Storage Manager provides a web-based interface for managing and
monitoring multiple Tivoli Storage Manager servers.

About this task

Through the Operations Center, you can manage multiple Tivoli Storage Manager
servers, issue commands to those servers, and access web clients.

Chapter 21. Managing a network of Tivoli Storage Manager servers 725

 Routing commands
Command routing enables an administrator to send commands for processing to
one or more servers at the same time. The output is collected and displayed at the
server that issued the routed commands.

About this task

If you have set up your servers as described in “Setting up communications for

command routing” on page 696, you can route Tivoli Storage Manager
administrative commands to one or more servers. A system administrator can
configure and monitor many different servers from a central server by using
command routing.

You can route commands to one server, multiple servers, servers defined to a
named group, or a combination of these servers. A routed command cannot be
further routed to other servers; only one level of routing is allowed.

Each server that you identify as the target of a routed command must first be
defined with the DEFINE SERVER command. If a server has not been defined, that
server is skipped and the command routing proceeds to the next server in the
route list.

Tivoli Storage Manager does not run a routed command on the server from which
you issue the command unless you also specify that server. To be able to specify
the server on a routed command, you must define the server just as you did any
other server.

Commands cannot be routed from the SERVER_CONSOLE ID.

Routed commands run independently on each server to which you send them. The
success or failure of the command on one server does not affect the outcome on
any of the other servers to which the command was sent.

For more information on command routing and return codes generated by

command processing, refer to Administrator's Reference.

(see “Setting up server groups” on page 728)

Routing commands to one or more servers

You can route commands to one or more servers, and to server groups. To
successfully route commands to other servers, you must have the proper
administrative authority on all servers that receive the command for processing.

About this task

The return codes for command routing can be one of three severities: 0, ERROR, or
WARNING. See Administrator's Reference for a list of valid return codes and
severity levels.

726 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Routing commands to single servers:

To route a command to a single server, enter the defined server's name, a colon,
and then the command to be processed.

About this task

For example, to route a QUERY STGPOOL command to the server that is named
ADMIN1, enter:
admin1: query stgpool

The colon after the server name indicates the end of the routing information. This
is also called the server prefix. Another way to indicate the server routing
information is to use parentheses around the server name, as follows:
(admin1) query stgpool

Note: When writing scripts, you must use the parentheses for server routing
information.

To route a command to more than one server, separate the server names with a
comma. For example, to route a QUERY OCCUPANCY command to three servers named
ADMIN1, GEO2, and TRADE5 enter:
admin1,geo2,trade5: query occupancy

or
(admin1,geo2,trade5) query occupancy

The command QUERY OCCUPANCY is routed to servers ADMIN1, GEO2, and

TRADE5. If a server has not been defined with the DEFINE SERVER command,
that server is skipped and the command routing proceeds to the next server in the
route list.

The routed command output of each server is displayed in its entirety at the server
that initiated command routing. In the previous example, output for ADMIN1
would be displayed, followed by the output of GEO2, and then the output of
TRADE5.

Processing of a command on one server does not depend upon completion of the
command processing on any other servers in the route list. For example, if GEO2
server does not successfully complete the command, the TRADE5 server continues
processing the command independently.

Routing commands to server groups:

A server group is a named group of servers. After you set up the groups, you can
route commands to the groups.

About this task

To route a QUERY STGPOOL command to the server group WEST_COMPLEX,

enter:
west_complex: query stgpool

or
(west_complex) query stgpool

Chapter 21. Managing a network of Tivoli Storage Manager servers 727

 The QUERY STGPOOL command is sent for processing to servers BLD12 and
BLD13 which are members of group WEST_COMPLEX.

To route a QUERY STGPOOL command to two server groups WEST_COMPLEX

and NORTH_COMPLEX, enter:
west_complex,north_complex: query stgpool

or
(west_complex,north_complex) query stgpool

The QUERY STGPOOL command is sent for processing to servers BLD12 and
BLD13 which are members of group WEST_COMPLEX, and servers NE12 and
NW13 which are members of group NORTH_COMPLEX.

See “Setting up server groups” for how to set up a server group.

Routing commands to single servers and server groups:

You can route commands to multiple single servers and to server groups at the
same time.

About this task

For example, to route the QUERY DB command to servers HQSRV, REGSRV, and
groups WEST_COMPLEX and NORTH_COMPLEX, enter:
hqsrv,regsrv,west_complex,north_complex: query db

or
(hqsrv,regsrv,west_complex,north_complex) query db

The QUERY DB command is sent for processing to servers HQSRV, REGSRV, to

BLD12 and BLD13 (both members of WEST_COMPLEX), and to NE12 and NW12
(both members of NORTH_COMPLEX).

Duplicate references to servers are removed in processing. For example, if you

route a command to server BLD12 and to server group WEST_COMPLEX (which
includes BLD12), the command is sent only once to server BLD12.

Setting up server groups

You can make command routing more efficient by creating one or more server
groups and adding servers to them. You can then route commands to server
groups in addition to, or in place of, routing commands to single servers.

Procedure

To use server groups, you must complete the following tasks:

1. Define the server groups.
2. Add the servers as members of the appropriate group.

Results

After you have the server groups set up, you can manage the groups and group
members.

728 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Defining a server group and members of a server group
You can define groups of servers to which you can then route commands. The
commands are routed to all servers in the group.

About this task

Task Required Privilege Class
Define a server group System

Define a server group member System

Procedure

To route commands to a server group, you must complete the following steps:
1. Define the server with the DEFINE SERVER command if it is not already defined.
2. Define a new server group with the DEFINE SERVERGROUP command. Server
group names must be unique because both groups and server names are
allowed for the routing information.
3. Define servers as members of a server group with the DEFINE GRPMEMBER
command.

Results

The following example shows how to create a server group that is called
WEST_COMPLEX, and define servers BLD12 and BLD13 as members of the
WEST_COMPLEX group:
define servergroup west_complex
define grpmember west_complex bld12,bld13

(see “Setting up communications for command routing” on page 696)

Managing server groups

You can query, copy, rename, update, and delete server groups as necessary.

About this task

Task Required Privilege Class
Query a server group System

Copy a server group System

Rename a server group System

Update a server group description System

Delete a server group System

Chapter 21. Managing a network of Tivoli Storage Manager servers 729

 Querying a server group:

You can obtain information about server groups by issuing the QUERY SERVERGROUP
command.

Procedure

To query the WEST_COMPLEX server group, issue the following command:

query servergroup west_complex

Results

The following sample output shows the results of the QUERY SERVERGROUP
command.
Server Group Members Description Managing profile
---------------------------------------------------------------------------
WEST_COMPLEX BLD12, BLD13

Copying a server group:

You can copy a server group by issuing the COPY SERVERGROUP command.

Procedure

To copy the entire server group contents of WEST_COMPLEX to the NEWWEST

server group, issue the following command:
copy servergroup west_complex newwest

Results

This command creates the new group. If the new group exists, the command fails.

Renaming a server group:

You can rename a server group by issuing the RENAME SERVERGROUP command.

Procedure

For example, to rename an existing server group NORTH_COMPLEX to NORTH,

issue the following command:
rename servergroup north_complex north

Updating a server group description:

You can update a server group by issuing the UPDATE SERVERGROUP command.

Procedure

For example, to update the NORTH server group to modify its description, issue
the following command:
update servergroup north description="Northern marketing region"

730 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Deleting a server group:

You can delete a server group by issuing the DELETE SERVERGROUP command.

Procedure

To delete the WEST_COMPLEX server group from the Tivoli Storage Manager
server, issue the following command:
delete servergroup west_complex

Results

This command removes all members from the server group. The server definition
for each group member is not affected. If the deleted server group is a member of
other server groups, the deleted group is removed from the other groups.

Managing group members

You can move and delete group members from a previously defined group.

About this task

Task Required Privilege Class
Move a group member to another group System

Delete a group member

Moving a group member to another group:

You can move group members to another group by issuing the MOVE GRPMEMBER
command.

Procedure

To move group member TRADE5 from the NEWWEST group to the

NORTH_COMPLEX group, issue the following command:
move grpmember trade5 newwest north_complex

Deleting a group member from a group:

You can delete group members from a group by issuing the DELETE GROUPMEMBER
command.

Procedure

To delete group member BLD12 from the NEWWEST server group, issue the
following command:
delete grpmember newwest bld12

Results

When you delete a server, the deleted server is removed from any server groups of
which it was a member.

Chapter 21. Managing a network of Tivoli Storage Manager servers 731

 Querying server availability
You can test a connection from your local server to a specified server by issuing
the PING SERVER command.

Procedure

For example, to ping the server GEO2, issue the following command:
ping server geo2

The PING SERVER command uses the user ID and password of the administrative ID
that issued the command. If the administrator is not defined on the server that is
being pinged, the ping fails even if the server is running.

Using virtual volumes to store data on another server

You can store the results of database backups and other items on another server as
a virtual volume.

About this task

Tivoli Storage Manager allows a server (a source server) to store these items on
another server (a target server):
v Database backups
v Export operations
v Storage pool operations
v DRM PREPARE command
The data is stored as virtual volumes, which look like sequential media volumes on
the source server, but are stored as archive files on a target server. The following
list includes the virtual volumes that can store data.
v Database backups
v Storage pool backups
v Data that is backed up, archived, or space managed from client nodes
v Client data that is migrated from storage pools on the source server
v Any data that can be moved by EXPORT and IMPORT commands
v DRM plan files

The source server is a client of the target server, and the data for the source server
is managed only by the source server. In other words, the source server controls
the expiration and deletion of the files that comprise the virtual volumes on the
target server. You cannot use virtual volumes when the source server and the
target server are on the same Tivoli Storage Manager server.

At the target server, the virtual volumes from the source server are seen as archive
data. The source server is registered as a client node (of TYPE=SERVER) at the
target server and is assigned to a policy domain. The archive copy group of the
default management class of that domain specifies the storage pool for the data
from the source server.

Note: If the default management class does not include an archive copy group,
data cannot be stored on the target server.

You can benefit from the use of virtual volumes in the following ways:

732 IBM Tivoli Storage Manager for Windows: Administrator's Guide

v Smaller Tivoli Storage Manager source servers can use the storage pools and
tape devices of larger Tivoli Storage Manager servers.
v For incremental database backups, virtual volumes can decrease wasted space on
volumes and under-utilization of high-end tape drives.
v The source server can use the target server as an electronic vault for recovery
from a disaster.

Be aware of the following conditions when you use virtual volumes:

v When you copy or move data from a deduplicated storage pool to a
non-deduplicated storage pool that uses virtual volumes, the data is
reconstructed. However, after the data movement or copy operation, the amount
of data that is reported as moved or copied is the amount of deduplicated data.
For example, if a storage pool contains 20 GB of deduplicated data that
represents 50 GB of total file data. If the data is moved or copied, the server
reports that 20 GB was moved or copied, even though 50 GB of data was sent.
v If you use virtual volumes for database backups, you might have the following
situation: SERVER_A backs up its database to SERVER_B, and SERVER_B backs
up its database to SERVER_A. If databases are backed up in that manner, if both
servers are at the same location, and if a disaster occurs that location, you might
have no backups with which to restore your databases.
v You cannot use a Centera storage pool as the target for virtual volumes.
v Under certain circumstances, inconsistencies might arise among virtual volume
definitions on the source server and the archive files on the target server. You
can use the RECONCILE VOLUMES command to reconcile these inconsistencies.
v To enable data validation between a source and target server, issuing both the
DEFINE SERVER and REGISTER NODE commands. For more information, see
“Validating a node's data” on page 535 and Administrator's Reference.
v Storage space limitations on the target server affect the amount of data that you
can store on that server.

Note: When you issue a DEFINE SERVER command, the source server sends a
verification code to the target server. When the source server begins a session with
the target server, it also sends the verification code. If the code matches what was
previously stored on the target, the session is opened in read/write mode. If the
verification code is lost at the source server (for example, after a database restore),
you can reset the code by issuing the UPDATE SERVER command with the FORCESYNC
parameter set to YES.

For details, see “Reconciling virtual volumes and archive files” on page 739.
Related concepts:
“Performance limitations for virtual volume operations” on page 735
Related tasks:
“Setting up source and target servers for virtual volumes” on page 734

Chapter 21. Managing a network of Tivoli Storage Manager servers 733

 Setting up source and target servers for virtual volumes
In the source and target relationship, the source server is defined as a client node
of the target server. To set up this relationship, a number of steps must be
performed at the two servers.

About this task

In the following example (illustrated in Figure 88 on page 735), the source server is
named TUCSON and the target server is named MADERA.
v At Tucson site:
1. Define the target server:
– MADERA has a TCP/IP address of [Link]:1845
– Assign the password CALCITE to MADERA.
– Assign TUCSON as the node name by which the source server TUCSON
will be known by the target server. If no node name is assigned, the server
name of the source server is used. To see the server name, you can issue
the QUERY STATUS command.
2. Define a device class for the data to be sent to the target server. The device
type for this device class must be SERVER, and the definition must include
the name of the target server.
v At Madera site:
Register the source server as a client node. The target server can use an existing
policy domain and storage pool for the data from the source server. However,
you can define a separate management policy and storage pool for the source
server. Doing so can provide more control over storage pool resources.
1. Use the REGISTER NODE command to define the source server as a node of
TYPE=SERVER. The policy domain to which the node is assigned determines
where the data from the source server is stored. Data from the source server
is stored in the storage pool specified in the archive copy group of the
default management class of that domain.
2. You can set up a separate policy and storage pool for the source server.
a. Define a storage pool named SOURCEPOOL:
define stgpool sourcepool autotapeclass maxscratch=20
b. Copy an existing policy domain STANDARD to a new domain named
SOURCEDOMAIN:
copy domain standard sourcedomain
c. Assign SOURCEPOOL as the archive copy group destination in the
default management class of SOURCEDOMAIN:
update copygroup sourcedomain standard standard type=archive
destination=sourcepool
3. After issuing these commands, ensure that you assign the source server to
the new policy domain (UPDATE NODE) and activate the policy.

734 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 Source server Target server

TUCSON MADERA

define server madera register node tucson calcite

password=calcite domain=standard
hladdress=[Link] type=server
lladdress=1845
nodename=tucson

define devclass targetclass

devtype=server
servername=madera

Figure 88. Communication configuration for virtual volumes

Related tasks:
“Changing policy” on page 478

Performance limitations for virtual volume operations

Server-to-server virtual volume performance can vary depending on your
environment and other variables, and might not be optimal for all data recovery
needs.

Some of the factors that can affect volume performance when using virtual
volumes are:
v Distance between locations
v Network infrastructure and bandwidth between locations
v Network configuration
v Data size and distribution
v Data read and write patterns

Given this performance variability, testing must be performed in environments that

are representative of the final production configuration. In your testing, include
throughput evaluations for both data-write operations (storing data from the
source server to the target server), and for data-read operations (reading data from
the target server to the source server, such as restoring data for a client).

Use the server-to-server virtual volumes feature to share a single tape library with
multiple servers. Although there are other situations that can use this feature, such
as cross-server or off-site vaulting, this feature is not optimized for long distances.
For best results, use server-to-server virtual volumes for situations where the two
servers (source and target) are not communicating over long distances. For
example:
v The source server and target server reside within the same building.
v The source server and target server reside in different buildings but are typically
covering short geographic distances such as a mile or a couple miles.

Chapter 21. Managing a network of Tivoli Storage Manager servers 735

 v The source server and target server reside in the same metropolitan area and are
located in separate buildings 10 - 50 miles apart.

Although network technologies allow for server-to-server communications covering

greater distances than discussed here, be careful when implementing a
server-to-server virtual volume configuration. Network latency and other factors
can significantly affect performance for long-distance implementations and might
not meet recovery-time or recovery-point objectives. In these situations, consider
using other technologies such as Virtual Tape Library replication, Metro Mirror, or
other storage hardware approaches where tuning and bandwidth management are
available to mitigate the distances involved.

Avoid moving large amounts of data between the servers, which might slow down
communications significantly, depending on the network bandwidth and
availability.

Specify, in the device class definition (DEVTYPE=SERVER) how often, and how
long a time period you want the source server to attempt to contact the target
server. Keep in mind that frequent attempts to contact the target server over an
extended period can affect your communications.

To minimize mount wait times, set the total mount limit for all server definitions
that specify the target server to a value that does not exceed the mount total limit
at the target server. For example, a source server has two device classes, each
specifying a mount limit of 2. A target server has only two tape drives. In this case,
the source server mount requests might exceed the target server tape drives.

Performance can also vary by operation. For example, in some network

configurations data store operations perform better than data read operations
depending on how the TCPWINDOWSIZE option is specified. Test all data read
operations to verify that adequate data read retrieval rates can be achieved.

Performing operations at the source server

You can perform certain operations at the source server that cause data to be
stored in a storage pool at the target server.

About this task

These operations are:

v Database backups
v Storage pool backups
v Client data backup, archive, or migration
v Data migration from one storage pool to another
v Export of server information
v DRM prepare
The following sections describe how to perform these operations. In the examples,
the following is assumed:
v The definitions shown in the previous section have been done.
v An operational TCP/IP connection, Version 4 or Version 6, exists between both
servers.
v Both servers are running.

736 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Backing up the database
You can back up the database of a source server to a target server.

About this task

For example, to perform an incremental backup of the source server and send the
volumes to the target server, issue the following command:
backup db type=incremental devclass=targetclass

Expiration Processing of Database Backup Volumes and Recovery Plan Files

with Disaster Recovery Manager: If your server uses DRM, expiration processing
can delete volumes containing expired database backups and recovery plan files.
One or more database backup volumes may be deleted from the volume history
during expiration processing if the following conditions are true:
v The volume has a device type of SERVER
v The volume is not part of the most recent database backup series
v The last volume of the database backup series has exceeded the expiration value
specified with the SET DRMDBBACKUPEXPIREDAYS command

See “Moving copy storage pool and active-data pool volumes on-site” on page
1028 for more information.

Backing up a storage pool

You can back up a storage pool of a source server to a target server.

About this task

For example, a primary storage pool named TAPEPOOL is on the source server.
You can define a copy storage pool named TARGETCOPYPOOL, also on the
source server. TARGETCOPYPOOL must have an associated device class whose
device type is SERVER. When you back up TAPEPOOL to TARGETCOPYPOOL,
the backup is sent to the target server. To accomplish this, issue the following
commands:
define stgpool targetcopypool targetclass pooltype=copy
maxscratch=20
backup stgpool tapepool targetcopypool

Storing client data on a target server

You can configure your Tivoli Storage Manager system so that when client nodes
registered to the source server back up, archive, or migrate their data, that data is
sent to the target server. When clients restore, retrieve, or recall their data, the
source server gets the data from the target server.

About this task

To configure your system, ensure that the management policy for those nodes
specifies a storage pool that has a device class whose device type is SERVER. For
example, the following command defines the storage pool named TARGETPOOL.
define stgpool targetpool targetclass maxscratch=20
reclaim=100

Note: Reclamation of a storage pool automatically begins when the percentage of

reclaimable space, which is specified by the RECLAIM parameter, is reached.
Reclamation of a target storage pool can involve the movement of a great deal of
data from the target server to the source server and back to the target. If this

Chapter 21. Managing a network of Tivoli Storage Manager servers 737

 operation occurs automatically during peak operating periods, it could slow
network performance significantly. If you set the value to 100, reclamation will not
occur automatically.

For details about storage pool reclamation and how to begin it manually, see
“Reclaiming space in sequential-access storage pools” on page 367.

Migrate data from a source server storage pool to a target server

storage pool
You can set up your storage pool hierarchy so that client data is migrated from a
storage pool on the source server to the target server.

About this task

For example, storage pool TAPEPOOL is on the source server. The TAPEPOOL
definition specifies NEXTSTGPOOL=TARGETPOOL. TARGETPOOL has been
defined on the source server as a storage pool of device type SERVER. When data
is migrated from TAPEPOOL, it is sent to the target server.
define stgpool tapepool tapeclass nextstgpool=targetpool
maxscratch=20

Exporting server information to a target server

You can use any of the Tivoli Storage Manager EXPORT commands to export data
from one Tivoli Storage Manager source server to sequential media on a target
Tivoli Storage Manager server. You must specify a device class with a device type
specified as SERVER.

About this task

For example, to copy server information directly to a target server, issue the
following command:
export server devclass=targetclass

Importing server information from a target server:

If data has been exported from a source server to a target server, you can import
that data from the target server to a third server. The server that will import the
data uses the node ID and password of the source server to open a session with
the target server. That session is in read-only mode because the third server does
not have the proper verification code.

About this task

For example, to import server information from a target server, issue the following
command:
import server devclass=targetclass

738 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Reconciling virtual volumes and archive files
When you restore the database on the source or target server, you should reconcile
the virtual volumes on the source server and the archive files on the target server.
Reconciliation can also be used if you have any other reason to suspect
inconsistencies.

About this task

To complete reconciliation, issue the RECONCILE VOLUMES command specifying

a device class with the device type of SERVER. In the following example,
TARGETCLASS is a server device class:
reconcile volumes targetclass fix=yes

The reconciliation action is determined by the FIX parameter as shown in Table 63.
Table 63. FIX parameter reconciliation
FIX= At the Source At the Target Server Action
Server
No files exist
Files exist but are marked
Volumes exist for deletion Report error
Active files exist but
NO
attributes do not match
Active files exist Report error
Volumes do not
exist Files exist but are marked
None
for deletion
No files exist Report error

For storage pool volumes:

Mark volumes as unavailable
Files exist but marked for Report error
deletion
For storage pool volumes: If
attributes match, mark files
on the target server as active
again, mark volumes on the
source server as unavailable,
Volumes exist and recommend that an
AUDIT VOLUME be done to
further verify the data. If
YES
attributes do not match, mark
volumes as unavailable.
Active files exist but Report error
attributes do not match
For storage pool volumes:
Mark volumes as unavailable
and recommend that an
AUDIT VOLUME be done to
further verify the data.
Active files exist Mark files for deletion on the
Volumes do not target server.
exist Files exist but marked for None
deletion

Chapter 21. Managing a network of Tivoli Storage Manager servers 739

740 IBM Tivoli Storage Manager for Windows: Administrator's Guide
Chapter 22. Exporting and importing data
Tivoli Storage Manager provides an export and import facility that allows you to
copy all or part of a server (export) so that data can be transferred to another
server (import).

Two methods are available to perform the export and import operation:
v Export directly to another server on the network. This results in an immediate
import process without the need for compatible sequential device types between
the two servers.
v Export to sequential media. Later, you can use the media to import the
information to another server that has a compatible device type.

Task Required Privilege Class

Export and import operations System
Display information about export and Any administrator
import operations

This chapter takes you through the export and import tasks. See the following
sections:

Concepts:
“Reviewing data that can be exported and imported”
Tasks for Exporting Directly to Another Server:
“Exporting data directly to another server” on page 744
“Preparing to export to another server for immediate import” on page 748
“Monitoring the server-to-server export process” on page 750
Tasks for Exporting to Sequential Media:
“Exporting and importing data using sequential media volumes” on page 753
“Exporting tasks” on page 755
“Importing data from sequential media volumes” on page 758

Reviewing data that can be exported and imported

Administrators can export or import server control information and file data from
server storage.

See the following lists for definitions of these kinds of data:

v Server control information, which includes:
– Administrator definitions
– Client node definitions
– Policy and scheduling definitions
v File data from server storage, which includes file space definitions and
authorization rules. You can request that file data be exported in any of the
following groupings of files:

© Copyright IBM Corp. 1993, 2013 741

 – Active and inactive versions of backed up files, archive copies of files, and
space-managed files
– Active versions of backed up files, archive copies of files, and space-managed
files
– Active and inactive versions of backed up files
– Active versions of backed up files
– Archive copies of files
– Space-managed files

Exporting restrictions
The export function does have some limitations and restrictions. One restriction is
that you can export information from an earlier version and release of Tivoli
Storage Manager to a later version and release, but not from a later version and
release to an earlier version and release.

For example, you can export from a V6.1 server to a V6.2 server, but you cannot
export from V6.2 server to V6.1 server.

The following list shows the restrictions for exporting data:

v Export between servers that are at the same version and release but different fix
packs might be possible. However, functional changes that are made in some fix
packs might prevent a successful export operation. For example, you cannot
export from a V6.1.3 server to a V6.1.2 server, or from a V6.1.2 server to a V6.1.1
or earlier server.
v Data can be exported from a server with retention protection enabled. The data
is not protected by retention when it is imported to another server.
v You cannot export nodes of type NAS (network-attached storage). Export
processing excludes these nodes.
v Exporting data to a Centera device class or importing data from a Centera
device class is not supported. However, files that are stored in Centera storage
pools can be exported and files that must be imported can be stored on a
Centera storage device.
v The EXPORT NODE and EXPORT SERVER commands do not export data from a shred
pool unless you explicitly permit it by setting the ALLOWSHREDDABLE
parameter to YES. If this value is specified, and the exported data includes data
from shred pools, that data cannot be shredded. The server does not issue a
warning if the export operation includes data from shred pools.
Related concepts:
“Securing sensitive client data” on page 539

Deciding what information to export

Your decision on what information to export depends upon why you are exporting
that information.

See the possible causes for exporting information below:

v To balance the workload across servers. For example, when many client nodes
access the same server, users contend for communication paths, server resources,
and tape mounts during a restore or retrieve operation.
To relieve a server of some work load and improve its performance, you may
want to take one or all of the following actions:
– Move a group of client nodes to a target server

742 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 – Move policy definitions associated with these client nodes
– Move administrator definitions for administrators who manage these client
nodes
To copy information to a second server (the target server), use the EXPORT NODE,
EXPORT POLICY, and EXPORT ADMIN commands.
When you complete the import, you can delete file spaces, client nodes, policy
objects, scheduling objects and administrators from the source server. This will
reduce contention for server resources.
v To copy data for the purpose of installing a new server, use the EXPORT SERVER
command.

Important:
1. Because results could be unpredictable, ensure that expiration, migration,
backup, or archive processes are not running when the EXPORT NODE command
is issued.
2. The EXPORT NODE and EXPORT SERVER commands will not export data from shred
pools unless you explicitly permit it by setting the ALLOWSHREDDABLE parameter
to YES. If this value is specified, and the exported data includes data from
shred pools, but that data can no longer be shredded.
Related concepts:
“Securing sensitive client data” on page 539

Deciding when to export

When you issue an EXPORT command, the operation runs as a background process.
This process allows you to continue performing administrative tasks. In addition,
users can continue to back up, archive, migrate, restore, retrieve, or recall files
using the server.

If you choose to perform an export operation during normal working hours, be

aware that administrators can change server definitions and users may modify files
that are in server storage.

When you export to sequential media, administrators or users may modify data
shortly after it has been exported, then the information copied to tape may not be
consistent with data stored on the source server. If you want to export an exact
point-in-time copy of server control information, you can prevent administrative
and other client nodes from accessing the server.

When you export directly to another server, administrators or users may modify
data shortly after it has been exported. You can decide to merge file spaces, use
incremental export, or prevent administrative and other client nodes from
accessing the server.
Related concepts:
“Preventing administrative clients from accessing the server” on page 744
Related tasks:
“Preventing client nodes from accessing the server” on page 744
Related reference:
“Options to consider before exporting” on page 745

Chapter 22. Exporting and importing data 743

 Preventing administrative clients from accessing the server
Administrators can change administrator, policy, or client node definitions during
an export process. To prevent administrators from modifying these definitions, you
can lock out administrator access to the server and cancel any administrative
sessions before issuing an EXPORT command.

After the export process is complete, unlock administrator access.

Related tasks:
“Canceling an IBM Tivoli Storage Manager session” on page 469
“Locking and unlocking administrator IDs from the server” on page 852

Preventing client nodes from accessing the server

If client node information is exported while that client is backing up, archiving, or
migrating files, the latest file copies for the client may not be exported to tape.

About this task

To prevent users from accessing the server during export operations, cancel
existing client sessions.

Procedure

After you cancel the client sessions, you can complete one of the following steps:
1. Disable server access to prevent client nodes from accessing the server.
This option is useful when you export all client node information from the
source server and want to prevent all client nodes from accessing the server.
2. Lock out particular client nodes from server access.
This option is useful when you export a subset of client node information from
the source server and want to prevent particular client nodes from accessing
the server until the export operation is complete.

What to do next

After the export operation is complete, allow client nodes to access the server
again by:
v Enabling the server
v Unlocking client nodes

Exporting data directly to another server

You can export all server control information or a subset of that same information.

Specify one or more of the following export commands:

v EXPORT ADMIN
v EXPORT NODE
v EXPORT POLICY
v EXPORT SERVER
When you export data to a target server, you must specify the server name that
will receive the data as an import operation.

744 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Options to consider before exporting
There are several options to consider before you export, such as whether to merge
file spaces, to have an incremental export, whether to replace definitions, which
source of active client backup data to export, and the possibility of restarting your
export operation.

Merging file spaces before exporting

You can merge imported client backup, archive, and space-managed files into
existing file spaces and automatically skip duplicate files that may exist in the
target file space on the server. Optionally, you can have new file spaces created.

If you do not want to merge file spaces, see the topic on how duplicate file spaces
are managed.

Choosing to merge file spaces allows you to restart a cancelled import operation
because files that were previously imported can be skipped in the subsequent
import operation. This option is available when you issue an EXPORT SERVER or
EXPORT NODE command.

When you merge file spaces, the server performs versioning of the imported
objects based on the policy bound to the files. An import operation may leave the
target file space with more versions than policy permits. Files are versioned to
maintain the policy intent for the files, especially when incremental export (using
the FROMDATE and FROMTIME parameters) is used to maintain duplicate client file
copies on two or more servers.

The following definitions show how the server merges imported files, based on the
type of object, when you specify MERGEFILESPACES=YES.
Archive Objects
If an archive object for the imported node having the same TCP/IP
address, TCP/IP port, name, insert date, and description is found to
already exist on the target server, the imported object is skipped.
Otherwise, the archive object is imported.
Backup Objects
If a backup object for the imported node has the same TCP/IP address,
TCP/IP port, insert date, and description as the imported backup object,
the imported object is skipped. When backup objects are merged into
existing file spaces, versioning will be done according to policy just as it
occurs when backup objects are sent from the client during a backup
operation. Setting their insert dates to zero (0) will mark excessive file
versions for expiration.
Otherwise, the server performs the following tasks:
v If the imported backup object has a later (more recent) insert date than
an active version of an object on the target server with the same node,
file space, TCP/IP address, and TCP/IP port, then the imported backup
object becomes the new active copy, and the active copy on the target
server is made inactive. Tivoli Storage Manager expires this inactive
version based on the number of versions that are allowed in policy.
v If the imported backup object has an earlier (less recent) insert date than
an active copy of an object on the target server with the same node, file
space, TCP/IP address, TCP/IP port, then the imported backup object is
inserted as an inactive version.

Chapter 22. Exporting and importing data 745

 v If there are no active versions of an object with the same node, file
space, TCP/IP address, and TCP/IP port on the target server, and the
imported object has the same node, file space, TCP/IP address, and
TCP/IP port as the versions, then:
– An imported active object with a later insert date than the most recent
inactive copy will become the active version of the file.
– An imported active object with an earlier insert date than the most
recent inactive copy will be imported as an inactive version of the file
v Any imported inactive objects will be imported as other inactive
versions of the object.
Space Managed Objects
If the imported node's space-managed object has the same external object
ID, that is unique to each space managed object, already exists on the
target server then the imported object is skipped. Otherwise, the
space-managed object is imported.

The number of objects imported and skipped is displayed with the final statistics
for the import operation.
Related concepts:
“Managing duplicate file spaces” on page 766
Related tasks:
“Querying the activity log for export or import information” on page 772

Incremental export before exporting

The system administrator can limit the file data exported to objects that were
stored on the server on or after the date and time specified.

You can use the FROMDATE and FROMTIME parameters to export data based on the
date and time the file was originally stored in the server. The FROMDATE and
FROMTIME parameters only apply to client user file data; these parameters have no
effect on other exported information such as policy. If clients continue to back up
to the originating server while their data is moving to a new server, you can move
the backup data that was stored on the originating server after the export
operation was initiated. This option is available when you issue an EXPORT SERVER
or EXPORT NODE command.

You can use the TODATE and TOTIME parameters to further limit the time you specify
for your export operation.

Replace definitions before exporting

You can specify whether definitions (not file data) are replaced on the target server.
If duplicate definitions exist on the target server, they can be replaced with the
imported definitions.

Alternatively, you can have the server skip duplicate definitions. This option is
available when you issue any of the EXPORT commands.
Related concepts:
“Determining whether to replace existing definitions” on page 760

746 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Sources of active client backup data
When exporting active versions of client backup data, the server searches for active
file versions in an active-data storage pool associated with a FILE device class, if
such a pool exists.
Related concepts:
“Selection for active file versions” on page 243

Restartable export operations

You can restart a suspended server-to-server export operation if you define the
FILEDATA value as anything other than NONE when issuing an EXPORT NODE or
EXPORT SERVER command. An export might be suspended during the export
operation if a failure occurs.

The resumed export continues at a point where the suspension took place.
Therefore, data that has already been exported is not exported again and only the
data that was not sent is included in the restarted export. Issue the QUERY EXPORT
command to view all running and suspended restartable export operations, the
RESTART EXPORT command to restart an export operation, or the SUSPEND EXPORT to
suspend a running server-to-server EXPORT NODE or EXPORT SERVER process.
Suspended server-to-server export operations are not affected by a server restart.

Note: Do not issue the CANCEL PROCESS command if you want to restart the
operation at a later time. CANCEL PROCESS ends the export process and deletes all
saved status.

An export operation cannot be suspended before it completes writing definitions to

the target server. The export operation might be restarted from the point at which
the operation fails or suspends if the export operation fails after writing the file
space definitions to the target server.

If an export operation fails prior to identifying all eligible files, when the export
operation is restarted it continues to identify eligible files and may export files that
were backed up while the operation was suspended.

A restarted export operation will export only the data that was identified. During a
suspension, some files or nodes identified for export might be deleted or might
expire. To ensure that all data is exported, restart the export operation at the
earliest time and restrict operations on the selected data.

A restartable server-to-server export operation goes through (in order) the

following three phases:
1. Creating definitions on target server
2. Identifying and exporting eligible files
3. File list complete. Exporting eligible files

At any given time, a restartable export operation will be in one of the following
states:
Running - Not Suspendible
This state directly corresponds to phase 1 of a restartable export, “Creating
definitions on target server.”
Running
The operation is running as an active process and is either in phase 2:
“Identifying and exporting eligible files” or phase 3: “File list complete.
Exporting eligible files.”

Chapter 22. Exporting and importing data 747

 Running - Suspend in Progress
The operation is in the process of being suspended as a result of a SUSPEND
EXPORT command. The export operation is fully suspended when all of the
data from the export operation is saved. An export operation in this state
can be in either phase 2 on page 747: “Identifying and exporting eligible
files” or phase 3 on page 747: “File list complete. Exporting eligible files.”
Suspended
The operation is not currently running. It may still be in either phase 2 on
page 747 or phase 3 on page 747.

An export operation that fails while it is in phase 1 on page 747 cannot be

restarted, and you must re-specify the export command. The new export operation
starts and all definitions are retransmitted. Before issuing the export command,
check the state of the target server to guarantee that the definitions will be
properly imported during the new operation. If the original export command
specified MERGEFILESPACES=NO, delete any filespace definitions imported on
the target server prior to the failure to ensure that filespace definitions are correctly
imported during the new operation. If the original export command specified
REPLACEDEFS=NO and you want to export any changes in the definitions made
since the original operation failed, delete all definitions on the target that were
imported prior to the failure.

When a server-to-server export operation restarts, the data continues to be

imported onto the target server into the same file spaces that were selected in the
original export process, regardless of the value of the MERGEFILESPACES
parameter. For server-to-server export operations, data is transmitted in
transactions to ensure the source server can restart at the correct location in the
data should the process fail. The target server's TXNGROUPMAX and
MOVESIZETHRESH values are used to control the amount of data within each
export transaction.

Attention: Ensure that the target server's Tivoli Storage Manager level is newer
or the same as the source server's level. If you suspend export operations and
upgrade the source server's database, the target server may stop the export
operation if the new source server's Tivoli Storage Manager level is incompatible
with the target server's level.

Preparing to export to another server for immediate import

When you export data to another server on the network, the export results in an
immediate import on the target server. You can export data to a Tivoli Storage
Manager server of the same or different operating system as the originating server.

About this task

A server-to-server export operation performs the following steps:

1. Opens a session with the target server.
2. Authenticates with the administrator's user ID and password.
3. Starts the equivalent of an IMPORT SERVER process.
Before you export data to another server on the network, perform the following
tasks:

748 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Procedure
1. Install Tivoli Storage Manager on the target server. This includes defining disk
space for the database and recovery log, and defining initial server storage. For
more information, refer toInstallation Guide.
2. Consider setting up enterprise configuration for the target server so you can
distribute consistent backup and archive policies to the target server.
3. Issue the DEFINE SERVER command to define the name of the target server or
the originating server.
4. Ensure that the administrator that issues the export command is defined with
the same administrator name and password on the target server, and has
System authority on the target server.
Related tasks:
Chapter 21, “Managing a network of Tivoli Storage Manager servers,” on page 687

“Setting up communications among servers” on page 692

Previewing results of an export operation for immediate import

When you export data to another server, you can use the PREVIEWIMPORT option to
determine how much data will be transferred without actually moving any data.
When PREVIEWIMPORT=NO, the export operation is performed, and the data is
immediately imported to the target server. This option is available when you issue
any EXPORT command.

About this task

Issue each EXPORT command with PREVIEWIMPORT=YES to determine which objects

and how much data will be copied to the target server. Use this information to
determine how much storage pool space is required on the target server. The
server sends the messages to the activity log and to the following place for each
operation:

Procedure

To determine how much space is required to export all server data, issue the
following command:
export server filedata=all previewimport=yes

After you issue the EXPORT SERVER command, a message similar to the following
message is issued when the server starts a background process:
EXPORT SERVER started as Process 4

You can view the preview results by querying the activity log.
You can also view the results on the following applications:
v Server console
Related tasks:
“Requesting information about an export or import process” on page 770
“Canceling server processes” on page 610

Chapter 22. Exporting and importing data 749

 Directing import messages to an output file
The information generated by the validation process can help you define a storage
hierarchy that supports the storage destinations currently defined in the import
data.

About this task

You can direct import messages to an output file to capture any error messages
that are detected during the import process. Do this by starting an administrative
client session in console mode before you invoke the import command.

Procedure

To direct messages to an output file named [Link], issue the following

command:
> dsmadmc -consolemode -outfile=[Link]

Monitoring the server-to-server export process

You can view information on the server console about an import or export process
or you can view the information from an administrative client that is running in
console mode.

Be watchful of mount messages, because the server might request mounts of

volumes that are not in the library. Check-in of volumes may be required.

If you want to view the status of any server-to-server exports that can be
suspended, issue the QUERY EXPORT command. The QUERY EXPORT command lists all
running or suspended operations.

If a process completes, you can query the activity log for status information from
an administrative client running in batch or interactive mode.

You can also query the activity log for status information from the server console.

The process first builds a list of what is to be exported. The process can therefore
be running for some time before any data is transferred. The connection between
the servers might time-out. You may need to adjust the COMMTIMEOUT and
IDLETIMEOUT server options on one or both servers.

If a process completes, you can query the activity log for status information from
the server console or from an administrative client running in batch or interactive
mode. The process first builds a list of what is to be exported. The process can
therefore be running for some time before any data is transferred. The connection
between the servers might time-out. You may need to adjust the COMMTIMEOUT and
IDLETIMEOUT server options on one or both servers.

750 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Exporting administrator information to another server
When you issue the EXPORT ADMIN command, the server exports administrator
definitions.

About this task

Each administrator definition includes:

v Administrator name, password, and contact information
v Any administrative privilege classes the administrator has been granted
v Whether the administrator ID is locked from server access

You can specify a list of administrator names, or you can export all administrator
names.

You can preview the result on the server console or by querying the activity log.

Procedure

Issue the following command to export all the administrator definitions to the
target server defined as OTHERSERVER.
export admin * toserver=otherserver previewimport=yes

This lets you preview the export without actually exporting the data for immediate
import.

Exporting client node information to another server

When you issue the EXPORT NODE command, the server exports client node
definitions.

About this task

Each client node definition includes:

v User ID, password, and contact information
v Name of the policy domain to which the client is assigned
v File compression status
v Whether the user has the authority to delete backed up or archived files from
server storage
v Whether the client node ID is locked from server access

You can also specify whether to export file data. File data includes file space
definitions and authorization rules. You can request that file data be exported in
any of the following groupings of files:
v Active and inactive versions of backed up files, archive copies of files, and
space-managed files
v Active versions of backed up files, archive copies of files, and space-managed
files
v Active and inactive versions of backed up files
v Active versions of backed up files
v Archive copies of files
v Space-managed files

Chapter 22. Exporting and importing data 751

 Procedure

To export client node information and all client files for NODE1 directly to
SERVERB, issue the following example command:
export node node1 filedata=all toserver=serverb

Results

Important: When you specify a list of node names or node patterns, the server
will not report the node names or patterns that do not match any entries in the
database. Check the summary statistics in the activity log to verify that the server
exported all intended nodes.

Exporting policy information to another server

When you issue the EXPORT POLICY command, the server exports information
belonging to each specified policy domain.

About this task

The following items are exported:

v Policy domain definitions
v Policy set definitions, including the active policy set
v Management class definitions, including the default management class
v Backup copy group and archive copy group definitions
v Schedule definitions
v Associations between client nodes and schedules

Procedure

To export policy information directly to SERVERB, issue the following example

command:
export policy replacedefs=yes toserver=serverb

Exporting server data to another server

When you issue the EXPORT SERVER command, the server exports all server control
information. You can also export file data information with the EXPORT SERVER
command.

Procedure

To export server data to another server on the network and have the file spaces
merged with any existing file spaces on the target server, as well as replace
definitions on the target server and have the data, that is to be exported, to begin
with any data inserted in the originating server beginning on 10/25/2007, issue the
following command:
export server toserver=serv23 fromdate=10/25/2007 filedata=all
mergefilespaces=yes dates=relative

752 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Exporting and importing data using sequential media volumes
Before you export or import data, you can use the EXPORT or IMPORT command with
the PREVIEW parameter to verify what data will be moved and prepare sequential
media for exporting and importing data.

Using preview before exporting or importing data

You can specify PREVIEW=YES on the EXPORT and IMPORT commands to generate a
report that shows how much data will be transferred without actually moving any
data. When PREVIEW=NO, the export or import operation is performed.

Procedure
1. To determine how much data and which objects are moved, issue both EXPORT
or IMPORT commands with PREVIEW=YES.
The server sends export or import messages to the activity log and to the
following place:
v Server console
Export Reports the types of objects, number of objects, and number of bytes
that would be copied to sequential media volumes. Use this
information to determine how many sequential media volumes you
will need.
Import
Reports the number and types of objects found on the sequential media
volumes that meet your import specifications. Also reports information
about any detected problems, such as corrupted data. Use this
information to determine which data to move to the server and to
determine if you have enough storage pool space allocated on the
server.
2. To determine how much space is required to export all server data, issue the
following command:
export server filedata=all preview=yes
After you issue the EXPORT SERVER command, the server starts a background
process and issues a message similar to the following output:
EXPORT SERVER started as Process 4

You can view the preview results by querying the activity log or the following
place:
v Server console
You can request information about the background process. If necessary, you
can cancel an export or import process.
Related tasks:
“Requesting information about an export or import process” on page 770
“Canceling server processes” on page 610

Chapter 22. Exporting and importing data 753

 Planning for sequential media used to export data
To export data, you must specify a device class that supports sequential media and
identify the volumes to store the exported data.

Procedure
1. Select a device class.
You can query the source and target servers to select a device class on each
server that supports the same device type. If you cannot find a device class on
each server that supports a matching device type, define a new device class
that is available to both servers.

Tip:
a. If the mount limit for the device class is exceeded when you request an
export (that is, if all the drives are busy), the server automatically cancels
lower priority operations, such as reclamation, to make a mount point
available for the export process.
b. You can export data to a storage pool on another server by specifying a
device class whose device type is SERVER.
2. Estimate the number of removable media volumes that you must label.
To estimate the number of removable media volumes that you must label,
divide the number of bytes to be moved by the estimated capacity of a volume
in the storage pool.
You can estimate the following forms of removable media volumes:
v The number of removable media volumes that are required to store the
exported data
For example, you have LTO Generation 1 tapes with an estimated capacity of
100 GB. If the preview shows that you need to transfer 150 GB of data, then
label at least two tape volumes before you export the data.
3. Use scratch media. You can use scratch media to ensure that you have
sufficient space to store all the exported data. If you use scratch media, record
the label names and the order in which they were mounted.
Or, you can use the USEDVOLUMELIST parameter on the EXPORT command to
create a file that contains a list of volumes that were used.
4. Label the removable media volumes.
During an import process, you must specify the order in which volumes are to
be mounted.
This order must match the order in which the following media were mounted
during the export process:
v volumes
To ensure that the volumes are mounted in the correct order, label them with
information that identifies the order in which they are mounted during the
import process. For example, label them as DSM001, DSM002, DSM003.
When you export data, record the date and time for each labeled volume. Store
this information in a safe location because you need the information when you
import the data. Alternatively, if you used the USEDVOLUMELIST parameter on the
export command, save the generated output file. This file can be used on the
import command volumes parameter.
For more information about exporting data, see EXPORT SERVER in the
Administrative's Reference.
Related tasks:

754 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 “Using virtual volumes to store data on another server” on page 732
“Defining device classes” on page 177

Exporting tasks
You can export all server control information or a subset of server control
information.

Specify one or more of the following export commands:

v EXPORT ADMIN
v EXPORT NODE
v EXPORT POLICY
v EXPORT SERVER

When you export data, you must specify the device class to which export data will
be written. You must also list the volumes in the order in which they are to be
mounted when the data is imported.

You can specify the USEDVOLUMELIST parameter to indicate the name of a file where
a list of volumes used in a successful export operation will be stored. If the
specified file is created without errors, it can be used as input to the IMPORT
command on the VOLUMENAMES=FILE:filename parameter. This file will contain
comment lines with the date and time the export was done, and the command
issued to create the export.

Note: An export operation will not overwrite an existing file. If you perform an
export operation and then try the same operation again with the same volume
name, the file is skipped, and a scratch file is allocated. To use the same volume
name, delete the volume entry from the volume history file.
Related tasks:
“Planning for sequential media used to export data” on page 754

Exporting administrator information

When you issue the EXPORT ADMIN command, the server exports administrator
definitions.

About this task

Each administrator definition includes:

v Administrator name, password, and contact information
v Any administrative privilege classes the administrator has been granted
v Whether the administrator ID is locked from server access

You can specify a list of administrator names, or you can export all administrator
names.

Procedure

Issue the following command to export definitions for the DAVEHIL and PENNER
administrator IDs to the DSM001 tape volume, which the TAPECLASS device class
supports, and to not allow any scratch media to be used during this export
process:
export admin davehil,penner devclass=tapeclass
volumenames=dsm001 scratch=no

Chapter 22. Exporting and importing data 755

 Exporting client node information
When you issue the EXPORT NODE command, the server exports client node
definitions.

About this task

Each client node definition includes:

v User ID, password, and contact information
v Name of the policy domain to which the client is assigned
v File compression status
v Whether the user has the authority to delete backed up or archived files from
server storage
v Whether the client node ID is locked from server access

You can also specify whether to export file data. File data includes file space
definitions and authorization rules. You can request that file data be exported in
any of the following groupings of files:
v Active and inactive versions of backed up files, archive copies of files, and
space-managed files
v Active versions of backed up files, archive copies of files, and space-managed
files
v Active and inactive versions of backed up files
v Active versions of backed up files
v Archive copies of files
v Space-managed files

When exporting active versions of client backup data, the server searches for active
file versions in an active-data pool associated with a FILE device class, if such a
pool exists. This process minimizes the number of mounts that are required during
the export process.

If you do not specify that you want to export file data, then the server only exports
client node definitions.

For example, suppose that you want to perform the following steps:
v Export definitions for client nodes and file spaces in the ENGPOLDOM policy
domain
v Export any active backup versions of files belonging to these client nodes
v Export this information to scratch volumes in the TAPECLASS device class

Procedure

To issue this command, enter:

export node filespace=* domains=engpoldom
filedata=backupactive devclass=tapeclass

In this example, the server exports:

v Definitions of client nodes assigned to ENGPOLDOM
v File space definitions and backup authorizations for each client node in
ENGPOLDOM
v Active versions of backed up files belonging to the client nodes assigned to
ENGPOLDOM
756 IBM Tivoli Storage Manager for Windows: Administrator's Guide
Related concepts:
“Selection for active file versions” on page 243

Exporting policy information

You must issue the EXPORT POLICY command to export policy information.

About this task

When you issue the EXPORT POLICY command, the server exports the following
information belonging to each specified policy domain:
v Policy domain definitions
v Policy set definitions, including the active policy set
v Management class definitions, including the default management class
v Backup copy group and archive copy group definitions
v Schedule definitions
v Associations between client nodes and schedules

For example, suppose that you want to export policy and scheduling definitions
from the policy domain named ENGPOLDOM. You want to use tape volumes
DSM001 and DSM002, which belong to the TAPECLASS device class, but allow the
server to use scratch tape volumes if necessary.

Procedure

To issue this command, enter:

export policy engpoldom
devclass=tapeclass volumenames=dsm001,dsm002

Exporting server data

When you issue the EXPORT SERVER command, the server exports all server control
information. You can also export file data information with the EXPORT SERVER
command.

About this task

For example, you want to export server data to four defined tape cartridges, which
the TAPECLASS device class supports. You want the server to use scratch volumes
if the four volumes are not enough, and so you use the default of SCRATCH=YES.

Procedure

To issue this command, enter:

export server devclass=tapeclass
volumenames=dsm001,dsm002,dsm003,dsm004 filedata=all

During the export process, the server exports definition information before it
exports file data information. This ensures that definition information is stored on
the first tape volumes. This process allows you to mount a minimum number of
tapes during the import process, if your goal is to copy only control information to
the target server.
In the example above, the server exports:
v Administrator definitions
v Client node definitions
v Policy domain, policy set, management class, and copy group definitions

Chapter 22. Exporting and importing data 757

 v Schedule definitions and client node associations
v File space definitions
v File space authorization rules

Important: If you are using wildcard characters when retrieving or restoring

data, file space authorization rules might prohibit data from being included.
Wildcard characters might be ignored if your client access rules also use
wildcard characters when the SET ACCESS command is issued. See the
Backup-Archive Clients Installation and User's Guide for your platform.
v Backed up, archived, and space-managed files
When exporting active versions of client backup data, the server searches for active
file versions in an active-data storage pool associated with a FILE device class, if
such a pool exists.
Related concepts:
“Selection for active file versions” on page 243

Importing data from sequential media volumes

After you understand how to import server control information and file data
information, you can import any subset of data to the target server.

Procedure

Before you import data to a new target server, you must:

1. Install Tivoli Storage Manager for the target server. This step includes defining
disk space for the database and recovery log.
For information on installing Tivoli Storage Manager, see Installation Guide.
2. Define server storage for the target server.
Because each server operating system handles devices differently, server storage
definitions are not exported. Therefore, you must define initial server storage
for the target server. The target server must at least be able to use a drive that
is compatible with the export media. This task can include defining libraries,
drives, device classes, storage pools, and volumes. See the Administrator's Guide
that applies to the target server.

Results

After Tivoli Storage Manager is installed and set up on the target server, a system
administrator can import all server control information or a subset of server
control information by specifying one or more of the following import commands:
v IMPORT ADMIN
v IMPORT NODE
v IMPORT POLICY
v IMPORT SERVER

The entire process of importing all server control information and file data from
tape volumes to a new target server includes:
v Previewing information before you import data
v Importing definitions
v Tailoring server storage definitions on the target server
v Importing file data

758 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Options to consider before importing
Before you import data from sequential media you might consider whether to
merge file spaces, replace existing definitions, or use a relative date when
importing file data.

Merge file spaces:

You can merge imported client backup, archive, and space-managed files into
existing file spaces, and automatically skip duplicate files that may exist in the
target file space on the server. Optionally, you can have new file spaces created.

If you do not want to merge file spaces, look into how duplicate file spaces are
managed. Choosing to merge file spaces allows you to restart a cancelled import
operation since files that were previously imported can be skipped in the
subsequent import operation.

When you merge file spaces, the server performs versioning of the imported
objects based on the policy bound to the files. An import operation may leave the
target file space with more versions than policy permits. Files are versioned to
maintain the policy intent for the files, especially when incremental export (using
the FROMDATE and FROMTIME parameters) is used to maintain duplicate client file
copies on two or more servers.

The following definitions show how the server merges imported files, based on the
type of object, when you specify MERGEFILESPACES=YES.
Archive Objects
If an archive object for the imported node having the same TCP/IP
address, TCP/IP port, insert date, and description is found to already exist
on the target server, the imported object is skipped. Otherwise, the archive
object is imported.
Backup Objects
If a backup object for the imported node has the same TCP/IP address,
TCP/IP port, insert date, and description as the imported backup object,
the imported object is skipped. When backup objects are merged into
existing file spaces, versioning will be done according to policy just as it
occurs when backup objects are sent from the client during a backup
operation. Setting their insert dates to zero (0) will mark excessive file
versions for expiration.
Otherwise, the server performs the following tasks:
v If the imported backup object has a later (more recent) insert date than
an active version of an object on the target server with the same node,
file space, TCP/IP address, and TCP/IP port, then the imported backup
object becomes the new active copy. The active copy on the target server
is made inactive. Tivoli Storage Manager expires this inactive version
based on the number of versions that are allowed in policy.
v If the imported backup object has an earlier (less recent) insert date than
an active copy of an object on the target server with the same node, file
space, TCP/IP address, and TCP/IP port, then the imported backup
object is inserted as an inactive version.
v If there are no active versions of an object with the same node, file
space, TCP/IP address, TCP/IP port on the target server, and the
imported object has the same node, TCP/IP address, TCP/IP port as the
versions, then:

Chapter 22. Exporting and importing data 759

 – An imported active object with a later insert date than the most recent
inactive copy will become the active version of the file.
– An imported active object with an earlier insert date than the most
recent inactive copy will be imported as an inactive version of the file
v Any imported inactive objects will be imported as other inactive
versions of the object.
Space Managed Objects
If the imported node's space-managed object has an external file ID which
already exists on the target server, then the imported object is skipped.
Otherwise, the space-managed object is imported.

The number of objects imported and skipped is displayed with the final statistics
for the import operation.
Related concepts:
“Managing duplicate file spaces” on page 766
Related tasks:
“Querying the activity log for export or import information” on page 772

Determining whether to replace existing definitions:

By using the REPLACEDEFS parameter with the IMPORT command, you can specify
whether to replace existing definitions on the target server when Tivoli Storage
Manager encounters an object with the same name during the import process.

For example, if a definition exists for the ENGPOLDOM policy domain on the
target server before you import policy definitions, then you must specify
REPLACEDEFS=YES to replace the existing definition with the data from the
export tape.

Definitions that can be replaced include administrator, client node, policy, or

schedule definitions. The default is to not replace existing definitions on the target
server.

Deciding whether to use a relative date when importing file data:

When you import file data, you can keep the original creation date for backup
versions and archive copies, or you can specify that the server use an adjusted
date.

If you want to keep the original dates set for backup versions and archive copies,
use DATES=ABSOLUTE, which is the default. If you use the absolute value, any
files whose retention period has passed will be expired shortly after they are
imported to the target server.

When you specify a relative date, the dates of the file versions are adjusted to the
date of import on the target server. This is helpful when you export from a server
that is in a different time zone than the target server.

760 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Previewing information before you import data
Before you import any data to the target server, preview each IMPORT command to
determine what data you want to import to the target server. You can import all or
a subset of export data from tapes.

About this task

When you set PREVIEW=YES, tape operators must mount export tape volumes so
that the target server can calculate the statistics for the preview.

Procedure

Issue the following design to preview information for the IMPORT SERVER
command:
import server devclass=tapeclass preview=yes
volumenames=dsm001,dsm002,dsm003,dsm004

Figure 89 on page 762 shows an example of the messages sent to the activity log
and the following place:
Server console

Chapter 22. Exporting and importing data 761

 ANR0402I Session 3 started for administrator SERVER_CONSOLE (Server).
ANR1363I Import volume DSM001 opened (sequence number 1).
ANR0610I IMPORT SERVER started by SERVER_CONSOLE as process 2.
ANR0612I IMPORT SERVER: Reading EXPORT SERVER data from server SERV1
exported 05/07/1996 [Link].
ANR0639I IMPORT SERVER: Processing domain ENGPOLDOM.
ANR0640I IMPORT SERVER: Processing policy set ACTIVE in policy domain
ENGPOLDOM.
ANR0640I IMPORT SERVER: Processing policy set STANDARD in policy domain
ENGPOLDOM.
ANR0641I IMPORT SERVER: Processing management class STANDARD in domain
ENGPOLDOM, set ACTIVE.
ANR0641I IMPORT SERVER: Processing management class MCENG in domain
ENGPOLDOM, set STANDARD.
ANR0641I IMPORT SERVER: Processing management class STANDARD in domain
ENGPOLDOM, set STANDARD.
ANR0643I IMPORT SERVER: Processing archive copy group in domain
ENGPOLDOM, set ACTIVE, management class STANDARD.
ANR0643I IMPORT SERVER: Processing archive copy group in domain ENGPOLDOM,
set STANDARD, management class MCENG.
ANR0643I IMPORT SERVER: Processing archive copy group in domain ENGPOLDOM,
set STANDARD, management class STANDARD.
ANR0642I IMPORT SERVER: Processing backup copy group in domain ENGPOLDOM,
set ACTIVE, management class STANDARD.
ANR0642I IMPORT SERVER: Processing backup copy group in domain ENGPOLDOM,
set STANDARD, management class MCENG.
ANR0642I IMPORT SERVER: Processing backup copy group in domain ENGPOLDOM,
set STANDARD, management class STANDARD.
ANR0638I IMPORT SERVER: Processing administrator DAVEHIL.
ANR0638I IMPORT SERVER: Processing administrator PENNER.
ANR0635I IMPORT SERVER: Processing node TOMC.
ANR0636I IMPORT SERVER: Processing file space OS2 for node TOMC as file
space OS1.
ANR0636I IMPORT SERVER: Processing file space DRIVED for node TOMC as file
space DRIVE1.
ANR0636I IMPORT SERVER: Processing file space OS2VDISK for node TOMC as file
space OS2VDIS1.
ANR1365I Import volume DSM001 closed (end reached).
ANR1363I Import volume DSM002 opened (sequence number 2).
ANR1365I Import volume DSM002 closed (end reached).
ANR1363I Import volume DSM003 opened (sequence number 3).
ANR1365I Import volume DSM003 closed (end reached).
ANR1363I Import volume DSM004 opened (sequence number 4).
ANR1365I Import volume DSM004 closed (end reached).
ANR0617I IMPORT SERVER: Processing completed successfully.
ANR0620I IMPORT SERVER: Copied 1 domain(s).
ANR0621I IMPORT SERVER: Copied 2 policy set(s).
ANR0622I IMPORT SERVER: Copied 2 management class(es).
ANR0623I IMPORT SERVER: Copied 6 copy group(s).
ANR0625I IMPORT SERVER: Copied 2 administrator(s).
ANR0626I IMPORT SERVER: Copied 1 node definition(s).
ANR0627I IMPORT SERVER: Copied 3 file space(s), 0 archive file(s) and 462
backup file(s).
ANR0629I IMPORT SERVER: Copied 8856358 bytes of data.
ANR0611I IMPORT SERVER started by SERVER_CONSOLE as process 2 has ended.

Figure 89. Sample report created by issuing preview for an import server command

Use the value reported for the total number of bytes copied to estimate storage
pool space needed to store imported file data.
For example, Figure 89 shows that 8 856 358 bytes of data will be imported.
Ensure that you have at least 8 856 358 bytes of available space in the backup
storage pools defined to the server. You can issue the QUERY STGPOOL and QUERY
VOLUME commands to determine how much space is available in the server storage
hierarchy.
In addition, the preview report shows that 0 archive files and 462 backup files will
be imported. Because backup data is being imported, ensure that you have
sufficient space in the backup storage pools used to store this backup data.
Related tasks:

762 IBM Tivoli Storage Manager for Windows: Administrator's Guide

“Tailoring server storage definitions on the target server” on page 765
“Using preview before exporting or importing data” on page 753
Related reference:
“Monitoring export and import processes” on page 770

Importing definitions
When previewing information before importing data, you must import server
control information. This includes administrator definitions, client node definitions,
policy domain, policy set, management class, and copy group definitions, schedule
definitions, and client node associations.

About this task

However, do not import file data at this time, because some storage pools named
in the copy group definitions may not exist yet on the target server.

Before you import server control information, perform the following tasks:

Procedure
1. Read the following topics:
v “Determining whether to replace existing definitions” on page 760
v “Determining how the server imports active policy sets”
2. Start an administrative client session in console mode to capture import
messages to an output file.
3. Import the server control information from specified tape volumes.
Related tasks:
“Directing import messages to an output file” on page 764
“Importing server control information” on page 765

Determining how the server imports active policy sets:

When the server imports policy definitions, several objects are imported to the
target server.

The following items are imported:

v Policy domain definitions
v Policy set definitions, including the ACTIVE policy set
v Management class definitions
v Backup copy group definitions
v Archive copy group definitions
v Schedule definitions defined for each policy domain
v Client node associations, if the client node definition exists on the target server

If the server encounters a policy set named ACTIVE on the tape volume during the
import process, it uses a temporary policy set named $$ACTIVE$$ to import the
active policy set.

If replacedefs=yes, a copy of the active policy set named $$ACTIVE$$ is made on

the target server. The $$ACTIVE$$ policy set is updated with the definitions from
the ACTIVE policy set in the export of the source. The new active policy set at the

Chapter 22. Exporting and importing data 763

 target server will contain all the management classes from the ACTIVE policy set
from the export server and from the ACTIVE policy set that already exists on the
importing server.

After $$ACTIVE$$ is imported to the target server, the server activates this policy
set. During the activation process, the server validates the policy set by examining
the management class and copy group definitions. If any of the following
conditions occur, the server issues warning messages during validation:
v The storage destinations specified in the backup and archive copy groups do not
refer to defined storage pools.
v The default management class does not contain a backup or archive copy group.
v The current ACTIVE policy set contains management class names that are not
defined in the policy set to be activated.
v The current ACTIVE policy set contains copy group names that are not defined
in the policy set to be activated.

After each $$ACTIVE$$ policy set has been activated, the server deletes that
$$ACTIVE$$ policy set from the target server. To view information about active
policy on the target server, you can use the following commands:
v QUERY COPYGROUP
v QUERY DOMAIN
v QUERY MGMTCLASS
v QUERY POLICYSET

Results from issuing the QUERY DOMAIN command show the activated policy set as
$$ACTIVE$$. The $$ACTIVE$$ name shows you that the policy set which is
currently activated for this domain is the policy set that was active at the time the
export was performed.

Directing import messages to an output file:

The information generated by the validation process can help you define a storage
hierarchy that supports the storage destinations currently defined in the import
data.

About this task

You can direct import messages to an output file to capture any error messages
that are detected during the import process. Do this by starting an administrative
client session in console mode before you invoke the import command.

Procedure

To direct messages to an output file named [Link], issue the following

command:
> dsmadmc -consolemode -outfile=[Link]

764 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Importing server control information:

If you have completed the prerequisite steps, you might be ready to import the
server control information.

About this task

Based on the information generated during the preview operation, you know that
all definition information has been stored on the first tape volume named DSM001.
Specify that this tape volume can be read by a device belonging to the
TAPECLASS device class.

Procedure

Issue the following command:

import server filedata=none devclass=tapeclass volumenames=dsm001

You can issue the command from an administrative client session or from the
following:
server console

Tailoring server storage definitions on the target server

If you previously imported definition information, you can use the reports that are
generated by the import process to help you tailor storage for the target server.

Procedure

To tailor server storage definitions on the target server, complete the following
steps:
1. Identify any storage destinations that are specified in copy groups and
management classes that do not match defined storage pools:
v If the policy definitions you imported included an ACTIVE policy set, that
policy set is validated and activated on the target server. Error messages that
are generated during validation include whether any management classes or
copy groups refer to storage pools that do not exist on the target server. If
you directed console messages to an output file, you have a copy of these
messages in a file.
v Query management class and copy group definitions to compare the storage
destinations that are specified with the names of existing storage pools on
the target server.
To request detailed reports for all management classes, backup copy groups,
and archive copy groups in the ACTIVE policy set, enter these commands:
query mgmtclass * active * format=detailed
query copygroup * active * standard type=backup format=detailed
query copygroup * active * standard type=archive format=detailed
2. If storage destinations for management classes and copy groups in the ACTIVE
policy set refer to storage pools that are not defined, complete one of the
following tasks:
v Define storage pools that match the storage destination names for the
management classes and copy groups.
v Change the storage destinations for the management classes and copy
groups. complete the following steps:
a. Copy the ACTIVE policy set to another policy set

Chapter 22. Exporting and importing data 765

 b. Modify the storage destinations of management classes and copy groups
in that policy set, as needed
c. Activate the new policy set

Results

Depending on the amount of client file data that you expect to import, you might
want to examine the storage hierarchy to ensure that sufficient storage space is
available. Storage pools that are specified as storage destinations by management
classes and copy groups might fill up with data. For example, you might need to
define extra storage pools to which data can migrate from the initial storage
destinations.
Related tasks:
“Directing import messages to an output file” on page 764
“Defining storage pools” on page 245
Related reference:
“Defining and updating a policy set” on page 500

Importing file data information

After you have defined the appropriate storage hierarchy on the target server, you
can import file data from the tape volumes. File data includes file space definitions
and authorization rules.

You can request that file data be imported in any of the following groupings:
v Active and inactive versions of backed up files, archive copies of files, and
space-managed files
v Active versions of backed up files, archive copies of files, and space-managed
files
v Active and inactive versions of backed up files
v Active versions of backed up files
v Archive copies of files
v Space-managed files

Data being imported will not be stored in active-data pools. Use the COPY
ACTIVEDATA command to store newly imported data into an active-data pool.

Before you import file data information:

v Understand how the server handles duplicate file space names
v Decide whether to keep the original creation date for backup versions and
archive copies or to import file data using an adjusted date

Managing duplicate file spaces:

When the server imports file data information, it imports any file spaces belonging
to each specified client node. If a file space definition already exists on the target
server for the node, the server does not replace the existing file space name.

If the server encounters duplicate file space names when it imports file data
information, it creates a new file space name for the imported definition by
replacing the final character or characters with a number. A message showing the
old and new file space names is written to the system log and to the activity log. A
message showing the old and new file space names is written to the activity log
and to the following place:

766 IBM Tivoli Storage Manager for Windows: Administrator's Guide

v server console
For example, if the C_DRIVE and D_DRIVE file space names reside on the target
server for node FRED and on the tape volume for FRED, then the server imports
the C_DRIVE file space as C_DRIV1 file space and the D_DRIVE file space as
D_DRIV1 file space, both assigned to node FRED.

Deciding whether to use a relative date when importing file data:

When you import file data, you can keep the original creation date for backup
versions and archive copies, or you can specify that the server use an adjusted
date.

Because tape volumes containing exported data might not be used for some time,
the original dates defined for backup versions and archive copies may be old
enough that files are expired immediately when the data is imported to the target
server.

To prevent backup versions and archive copies from being expired immediately,
specify DATES=RELATIVE on the IMPORT NODE or IMPORT SERVER commands to
adjust for the elapsed time since the files were exported to tape.

For example, assume that data exported to tape includes an archive copy archived
five days prior to the export operation. If the tape volume resides on the shelf for
six months before the data is imported to the target server, the server resets the
archival date to five days prior to the import operation.

If you want to keep the original dates set for backup versions and archive copies,
use DATES=ABSOLUTE, which is the default. If you use the absolute value, any
files whose retention period has passed will be expired shortly after they are
imported to the target server.

Issuing an import server or import node command:

You can import file data, either by issuing the IMPORT SERVER or IMPORT NODE
command. When you issue either of these commands, you can specify which type
of files should be imported for all client nodes specified and found on the export
tapes.

About this task

You can specify any of the following values to import file data:
All Specifies that all active and inactive versions of backed up files, archive
copies of files, and space-managed files for specified client nodes are
imported to the target server
None Specifies that no files are imported to the target server; only client node
definitions are imported
Archive
Specifies that only archive copies of files are imported to the target server
Backup
Specifies that only backup copies of files, whether active or inactive, are
imported to the target server

Chapter 22. Exporting and importing data 767

 Backupactive
Specifies that only active versions of backed up files are imported to the
target server
Allactive
Specifies that only active versions of backed up files, archive copies of files,
and space-managed files are imported to the target server
Spacemanaged
Specifies that only files that have been migrated from a user’s local file
system (space-managed files) are imported

For example, suppose you want to import all backup versions of files, archive
copies of files, and space-managed files to the target server. You do not want to
replace any existing server control information during this import operation.

Procedure

Specify the four tape volumes that were identified during the preview operation.
These tape volumes can be read by any device in the TAPECLASS device class. To
issue this command, enter:
import server filedata=all replacedefs=no
devclass=tapeclass volumenames=dsm001,dsm002,dsm003,dsm004

You can limit the import to nodes that were assigned to specific policy domains on
the source server. For example, suppose you exported from the source server the
data for all nodes in all domains. To import to the target server the data only for
nodes that were in the ENGDOM on the source server, enter this command:
import node filedata=all domains=engdom devclass=tapeclass
volumenames=dsm001,dsm002,dsm003,dsm004

If the ENGDOM policy domain exists on the target server, the imported nodes are
assigned to that domain. If ENGDOM does not exist on the target server, the
imported nodes are assigned to the STANDARD policy domain.
If you do not specify a domain on the IMPORT NODE command, the imported node
is assigned to the STANDARD policy domain.

Importing subsets of information

You can use an IMPORT command to copy a subset of the information from export
tapes to the target server. For example, if a tape was created with EXPORT SERVER,
you can import only node information from the tape by using IMPORT NODE.

While the server allows you to issue any import command, data cannot be
imported to the server if it has not been exported to tape. For example, if a tape is
created with the EXPORT POLICY command, an IMPORT NODE command will not find
any data on the tape because node information is not a subset of policy
information.

See Table 64 on page 769 for the commands that you can use to import a subset of
exported information to a target server.

768 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Table 64. Importing a subset of information from tapes
If tapes were created with You can issue this import You cannot issue this import
this export command: command: command:
EXPORT SERVER IMPORT SERVER
IMPORT ADMIN Not applicable.
IMPORT NODE
IMPORT POLICY
EXPORT NODE IMPORT NODE IMPORT ADMIN
IMPORT SERVER IMPORT POLICY
EXPORT ADMIN IMPORT ADMIN IMPORT NODE
IMPORT SERVER IMPORT POLICY
EXPORT POLICY IMPORT POLICY IMPORT ADMIN
IMPORT SERVER IMPORT NODE

Recovering from errors during the import process

During import processing, the server might encounter invalid data due to
corruption during storage on tape or in the database prior to the export operation.

If invalid data is encountered during an import operation, the server uses the
default value for the new object's definition. If the object already exists, the existing
parameter is not changed.

During import and export operations, the server reports on the affected objects to
the activity log and also to the:
server console
You should query these objects when the import process is complete to see if they
reflect information that is acceptable.

Each time you run the IMPORT NODE or IMPORT SERVER command with the FILEDATA
parameter equal to a value other than NONE, Tivoli Storage Manager creates a
new file space and imports data to it. This process ensures that the current import
does not overwrite data from a previous import.

A file space definition may already exist on the target server for the node. If so, an
administrator with system privilege can issue the DELETE FILESPACE command to
remove file spaces that are corrupted or no longer needed. For more information
on the DELETE FILESPACE command, refer to the Administrator's Reference.
Related concepts:
“Managing duplicate file spaces” on page 766

Renaming a file space:

An imported file space can have the same name as a file space that already exists
on a client node. In this case, the server does not overlay the existing file space,
and the imported file space is given a new system generated file space name.

This new name may match file space names that have not been backed up and are
unknown to the server. In this case, you can use the RENAME FILESPACE command
to rename the imported file space to the naming convention used for the client
node.

Chapter 22. Exporting and importing data 769

 Monitoring export and import processes
The server lets you monitor export or import processes while they are running or
after they have completed.

You can use the following two ways to monitor export or import processes:
v You can view information about a process that is running on the server console
or from an administrative client running in console mode.
v After a process has completed, you can query the activity log for status
information from an administrative client running in batch or interactive mode.
Watch for mount messages, because the server might request mounts of volumes
that are not in the library. The process first builds a list of what is to be exported.
The process can therefore be running for some time before any data is transferred.

Check-in of volumes may be required.

Requesting information about an export or import process

After you issue an EXPORT or IMPORT command, the server starts a background
process, assigns a process ID to the operation, and displays the process ID when
the operation starts.

About this task

You can query an export or import process by specifying the process ID number.

Procedure

For example, to request information about the EXPORT SERVER operation, which
started as process 4, enter:
query process 4

If you issue a preview version of an EXPORT or IMPORT command and then query
the process, the server reports the types of objects to be copied, the number of
objects to be copied, and the number of bytes to be copied.
When you export or import data and then query the process, the server displays
the number and types of objects copied so far, and the total number of bytes that
have been transferred, along with information on any media mount requests that
may be outstanding for the process.
Related tasks:
“Requesting information about server processes” on page 791

Viewing information from the server console

When you issue an EXPORT or IMPORT command, either from the server console or
from an administrative client, information is displayed on the server console.

Figure 90 on page 771 shows an example of the information that is displayed after
issuing an EXPORT SERVER command.

770 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 ANR0610I EXPORT SERVER started by SERVER_CONSOLE as process 1.
ANR0639I EXPORT SERVER: Processing domain ENGPOLDOM.
ANR0640I EXPORT SERVER: Processing policy set ACTIVE in policy domain
ENGPOLDOM.
ANR0640I EXPORT SERVER: Processing policy set STANDARD in policy domain
ENGPOLDOM.
ANR0641I EXPORT SERVER: Processing management class STANDARD in domain
ENGPOLDOM, set ACTIVE.
ANR0641I EXPORT SERVER: Processing management class STANDARD in domain
ENGPOLDOM, set STANDARD.
ANR0643I EXPORT SERVER: Processing archive copy group in domain
ENGPOLDOM, set STANDARD, management class ACTIVE.
ANR0643I EXPORT SERVER: Processing archive copy group in domain
ENGPOLDOM, set STANDARD, management class STANDARD.
ANR0643I EXPORT SERVER: Processing backup copy group in domain
ENGPOLDOM, set STANDARD, management class ACTIVE.
ANR0643I EXPORT SERVER: Processing backup copy group in domain
ENGPOLDOM, set STANDARD, management class STANDARD.
ANR0604I EXPORT SERVER: No schedules were found in policy domain * for
exporting.
ANR0635I EXPORT SERVER: Processing node TOMC.
ANR0605I EXPORT SERVER: No schedule associations were found in
policy domain * for exporting.
ANR0637I EXPORT SERVER: Processing file space DRIVED for node TOMC.
ANR0637I EXPORT SERVER: Processing file space OS2 for node TOMC.
ANR0637I EXPORT SERVER: Processing file space OS2VDISK for node TOMC.
ANR0617I EXPORT SERVER: Processing completed successfully.
ANR0620I EXPORT SERVER: Copied 1 domain(s).
ANR0621I EXPORT SERVER: Copied 2 policy set(s).
ANR0622I EXPORT SERVER: Copied 2 management class(es).
ANR0623I EXPORT SERVER: Copied 4 copy group(s).
ANR0626I EXPORT SERVER: Copied 1 node definition(s).
ANR0627I EXPORT SERVER: Copied 3 file space(s), 16 archive file(s)
and 0 backup file(s).
ANR0629I EXPORT SERVER: Copied 3045632 bytes of data.
ANR0611I EXPORT SERVER started by SERVER_CONSOLE as process 1 has ended.

Figure 90. Sample export server output

Viewing information from an administrative client

You can use the console mode from an administrative client to monitor export or
import operations or to capture processing messages to an output file.

Procedure
1. To start an administrative session in console mode, issue the following
command:
> dsmadmc -consolemode
While the system is running in console mode, you cannot enter any
administrative commands from the client session. You can, however, start
another administrative client session for entering commands (for example,
QUERY PROCESS) if you are using a multitasking workstation, such as AIX.
2. If you want the server to write all terminal output to a file, specify the OUTFILE
option with a destination. For example, to write output to the [Link] file,
enter:
> dsmadmc -consolemode -outfile=[Link]
For information about using the CONSOLE mode option and ending an
administrative session in console mode, see the Administrator's Reference.

Chapter 22. Exporting and importing data 771

 Querying the activity log for export or import information
After an export or import process has completed, you can query the activity log for
status information and possible error messages.

About this task

To minimize processing time when querying the activity log for export or import
information, restrict the search by specifying EXPORT or IMPORT in the SEARCH
parameter of the QUERY ACTLOG command.

Procedure

To determine how much data will be moved after issuing the preview version of
the EXPORT SERVER command, query the activity log by issuing the following
command:
query actlog search=export

Results

Figure 91 on page 773 displays a sample activity log report.

772 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 Date/Time Message
-------------------- ---------------------------------------------------
07/03/2002 [Link] ANR0610I EXPORT SERVER started by ADMIN as
process 1.
07/03/2002 [Link] ANR0639I EXPORT SERVER: Processing domain ENGPOLDOM.
07/03/2002 [Link] ANR0640I EXPORT SERVER: Processing policy set
ACTIVE in policy domain ENGPOLDOM.
07/03/2002 [Link] ANR0640I EXPORT SERVER: Processing policy set
STANDARD in policy domain ENGPOLDOM.
07/03/2002 [Link] ANR0641I EXPORT SERVER: Processing management class
STANDARD in domain ENGPOLDOM, set ACTIVE.
07/03/2002 [Link] ANR0641I EXPORT SERVER: Processing management class
STANDARD in domain ENGPOLDOM, set STANDARD.
07/03/2002 [Link] ANR0643I EXPORT SERVER: Processing archive copy
group in domain ENGPOLDOM, set STANDARD, management class ACTIVE.
07/03/2002 [Link] ANR0643I EXPORT SERVER: Processing archive copy
group in domain ENGPOLDOM, set STANDARD, management class STANDARD.
07/03/2002 [Link] ANR0642I EXPORT SERVER: Processing backup copy
group in domain ENGPOLDOM, set STANDARD, management class ACTIVE.
07/03/2002 [Link] ANR0642I EXPORT SERVER: Processing backup copy
group in domain ENGPOLDOM, set STANDARD, management class STANDARD.
07/03/2002 [Link] ANR0604I EXPORT SERVER: No schedules were found in policy
domain * for exporting.
07/03/2002 [Link] ANR0635I EXPORT SERVER: Processing node TOMC.
07/03/2002 [Link] ANR0605I EXPORT SERVER: No schedule associations were
found in policy domain * for exporting.
07/03/2002 [Link] ANR0637I EXPORT SERVER: Processing file space DRIVED for
node TOMC.
07/03/2002 [Link] ANR0637I EXPORT SERVER: Processing file space OS2 for node
TOMC.
07/03/2002 [Link] ANR0637I EXPORT SERVER: Processing file space OS2VDISK for
node TOMC.
07/03/2002 [Link] ANR0617I EXPORT SERVER: Processing completed successfully.
07/03/2002 [Link] ANR0620I EXPORT SERVER: Copied 1 domain(s).
07/03/2002 [Link] ANR0621I EXPORT SERVER: Copied 2 policy set(s).
07/03/2002 [Link] ANR0622I EXPORT SERVER: Copied 2 management class(es).
07/03/2002 [Link] ANR0623I EXPORT SERVER: Copied 4 copy group(s).
07/03/2002 [Link] ANR0626I EXPORT SERVER: Copied 1 node definition(s).
07/03/2002 [Link] ANR0627I EXPORT SERVER: Copied 3 file space(s),
16 export file(s) and 0 backup file(s).
07/03/2002 [Link] ANR0629I EXPORT SERVER: Copied 3045632 bytes of data.
07/03/2002 [Link] ANR0611I EXPORT SERVER started by ADMIN as
process 1 has ended.

Figure 91. Sample activity log report on exported data

Exporting and importing data from virtual volumes

You can perform all the EXPORT and IMPORT operations to virtual volumes that
are described in the sequential media topics.

Data stored as virtual volumes appear to be sequential storage pool volumes on

the source server, but are actually stored as archive files on another server. Those
archive files can be in random or sequential access storage pools. The EXPORT and
IMPORT commands are identical to those previously shown except that the device
class specified in the commands must have a device type of SERVER.
Related tasks:
“Using virtual volumes to store data on another server” on page 732
“Exporting and importing data using sequential media volumes” on page 753

Chapter 22. Exporting and importing data 773

774 IBM Tivoli Storage Manager for Windows: Administrator's Guide
Part 5. Monitoring operations
Monitor client operations, server operations, and maintenance processes to ensure
that applications and systems are protected as expected.

About this task

For information about monitoring operations with the Operations Center, including
the monitoring of alerts, see Chapter 17, “Managing the storage environment with
the Operations Center,” on page 587.

© Copyright IBM Corp. 1993, 2013 775

776 IBM Tivoli Storage Manager for Windows: Administrator's Guide
Chapter 23. Daily monitoring tasks
As a Tivoli Storage Manager administrator, monitor operations daily to ensure that
the Tivoli Storage Manager system is functioning properly. Daily monitoring tasks
focus on examining server processes, server database, storage pools, and scheduled
operations.

About this task

You can complete the monitoring tasks by using the command-line interface (CLI)
or the Operations Center, which also includes a CLI.

The following list describes some of the items that are important to monitor daily.
Instructions for monitoring these items, and other monitoring tasks can be found
in the topics in this section. Not all of these tasks apply to all environments.
v Verify that the database file system has enough space.
v Examine the database percent utilization, available free space, and free-pages.
v Verify that there is enough disk space in the file systems that contain these log
files.
– Active log
– Archive log
– Mirror log
– Archive failover log
v Verify that the instance directory file system has enough space.
v Verify that the database backups completed successfully, and that they are
running frequently enough.
v Check the database and recovery log statistics.
v Verify that you have current backup files for device configuration and volume
history information. You can find the file names for the backups by looking in
the [Link] file for the DEVCONFIG and VOLUMEHISTORY options. Ensure that
file systems where the files are stored have sufficient space.
v Search the summary table for failed processes.
v Search the activity log for error messages.
v For storage pools that have deduplication enabled, ensure that processes are
completing successfully.
v Check the status of your storage pools to ensure that there is enough space
available.
v Check for any failed storage pool migrations.
v Check the status of sequential access storage pools.
v Check how many scratch volumes are available.
v Determine if there are any tape drives offline, or their paths that are offline.
v Determine if there are any libraries offline, or their paths that are offline.
v Verify that all of the tapes have the appropriate write-access.
v Verify the status and settings for disaster recovery manager (DRM).
v Check for failed or missed schedules.
v Check the summary table for scheduled client operations such as backup,
restore, archive, and retrieve.

© Copyright IBM Corp. 1993, 2013 777

 v Check the summary table for scheduled server operations such as migration,
reclamation, and expiration.
Related tasks:
Chapter 17, “Managing the storage environment with the Operations Center,” on
page 587

Monitoring operations using the command line

You can use the command-line interface to monitor your server processes,
database, storage pools, and scheduled operations.
Related tasks:
Chapter 17, “Managing the storage environment with the Operations Center,” on
page 587

Monitoring your server processes daily

Monitor server processes daily to ensure that your server is functioning properly,
and that no problems exist.

About this task

For detailed information about the commands mentioned here, see the
Administrator's Reference.

The examples used here are based on a 24-hour period, but your values can differ
depending on the time frame you specify.

The following steps describe the commands that you can use to monitor server
processes:

Procedure
1. Search the summary table for any server processes that failed within the
previous 24-hour period:
select activity as process, number as processnum from summary where
activity in (’EXPIRATION’,’RECLAMATION’,’MIGRATION’,’STGPOOL BACKUP’,
’FULL_DBBACKUP’,’INCR_DBBACKUP’,’REPLICATION’) and successful=’NO’
and end_time> (current_timestamp - interval ’24’ hours)

This example output indicates that backup storage pool process number 7
failed:
PROCESS: STGPOOL BACKUP
PROCESSNUM: 7

2. Search the activity log for the messages associated with the failed process
number that was indicated in the output of the command in Step 1.
select message from actlog where process=7 and date_time>(current_timestamp
- interval ’24’ hours) and severity in (’W’,’E’,’S’)

Example output:
MESSAGE: ANR1221E BACKUP STGPOOL: Process 7 terminated - insufficient space in
target storage pool FILECOPYPOOL. (SESSION: 1, PROCESS: 7)

MESSAGE: ANR1893E Process 7 for BACKUP STORAGE POOL completed with a

completion state of FAILURE. (SESSION: 1, PROCESS: 7)

778 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 3. Examine the frequency of your successful database backups to determine if
they are running often enough to clear archive log space, provide sufficient
recovery points, and allow for appropriate number of volumes to be saved for
disaster recovery purposes.

Tip: This SELECT statement is searching the last 24 hours for a full or
incremental database backup. You can adjust this statement to search the last 48
hours or longer, depending on the frequency that you have scheduled
TYPE=FULL database backups.
select count(*) as frequency from summary where end_time>(current_
timestamp - interval ’24’ hours) and successful=’YES’ and activity in
(’FULL_DBBACKUP’,’INCR_DBBACKUP’)

Example output:
FREQUENCY
------------
3

4. If storage pools are enable for deduplication, ensure that the

duplicate-identification processes are completing successfully and processing
the expected amount of data:
select activity,number,sum(examined) as filesprocessed,sum(affected)
as duplicateextents,sum(bytes) as duplicatebytes,successful from summary
where activity=’IDENTIFY’ and entity=’STORAGE_POOL_NAME’
and end_time>(current_timestamp - interval ’24’ hours) group by
activity,number,successful

Example output:
ACTIVITY: IDENTIFY
NUMBER: 5
FILESPROCESSED: 12946
DUPLICATEEXTENTS: 10504
DUPLICATEBYTES: 127364341
SUCCESSFUL: YES

Related tasks:
“Monitoring your database daily”
“Monitoring disk storage pools daily” on page 782
“Monitoring sequential access storage pools daily” on page 783
“Monitoring scheduled operations daily” on page 786
Chapter 17, “Managing the storage environment with the Operations Center,” on
page 587

Monitoring your database daily

Monitor the database daily to ensure that the database is functioning properly, and
that no problems exist.

About this task

For detailed information about the commands mentioned here, see the
Administrator's Reference.

The following steps describe the commands that you can use to monitor the
database:

Chapter 23. Daily monitoring tasks 779

 Procedure
1. Use the QUERY DBSPACE command, and then examine the file system information
reported through the query to ensure that the file system has adequate space.
Examine the total, used, and free space.
query dbspace

Example output:
Location Total Space(MB) Used Space(MB) Free Space(MB)
------------------------------ ----------- --------------- ---------------
/fvt/kolty/srv/db2/mgsA3/pat-h1 253,952.00 62,066.64 191,821.36

2. Examine the file systems where the database is located, using the appropriate
operating system commands for the following:
v Ensure that the file systems are not approaching full.
v Ensure that other applications, or unexpected users of the file system space
are not storing data in the server database directories.
v Check the operating system and device error logs for any early signs or
indications of device failures.
3. Query the database to ensure that the percent utilization is acceptable, and that
the remaining space is sufficient for the next few days or weeks of expected
activity. This includes examining the free space available, and the free-pages
values. If you find that you are approaching your space limits, take action to
ensure that you get additional space provisioned to avoid any potential
problems.
query db format=detailed

Example output:
Database Name: mgsA2
Total Size of File System (MB): 253,952
Space Used by Database(MB): 544
Free Space Available (MB): 191,821
Total Pages: 40,964
Usable Pages: 40,828
Used Pages: 33,116
Free Pages: 7,712
Buffer Pool Hit Ratio: 97.7
Total Buffer Requests: 102,279
Sort Overflows: 0
Package Cache Hit Ratio: 78.9
Last Database Reorganization: 08/24/2011 [Link]
Full Device Class Name: FILECLASS
Incrementals Since Last Full: 1
Last Complete Backup Date/Time: 08/25/2011 [Link]

4. Monitor the file systems to ensure that they are not running out of space. Verify
that there is enough disk space in the file systems that contain these log files:
v Active log
v Archive log
v Mirror log
v Archive failover log
If the archive log directory fills up it will overflow to the active log directory. If
you see archive log space file systems filling up, it might be an indication that
a database backup is not being run, or not being run often enough. It might
also be an indication that the space is shared with other applications that are
contending for the same space.

780 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 Issue this command to look at the total space used, free space, and so on.
query log format=detailed

Example output:
Total Space(MB): 1,000
Used Space(MB): 0
Free Space(MB): 995
Active Log Directory: /fvt/kolty/srv/db2/mgsA3/alog
Archive Log Directory: /fvt/kolty/srv/db2/mgsA3/archlog
Mirror Log Directory:
Archive Failover Log Directory: /fvt/kolty/srv/db2/mgsA3/flog

5. Examine the instance directory to ensure that it has enough space. If there is
insufficient space in this directory, the Tivoli Storage Manager server fails to
start.
You should also examine the instance_dir/sqllib/db2dump directory and delete
*.[Link] and *.[Link] files regularly.

V6.2 and later servers:

v The [Link] file is managed by Tivoli Storage Manager and is
automatically pruned periodically.
v The DB2DIAGSIZE option can be used to control the maximum size of the
diagnostic log files.

V6.1 servers:
v Servers that are running version 6.1 must periodically delete the [Link]
file.
6. Verify that the database backups completed successfully, and examine the
details to determine if there are any problems:
select * from summary where end_time>(current_timestamp - interval
’24’ hours) and activity in (’FULL_DBBACKUP’,’INCR_DBBACKUP’)
If there are no results to this select command, then there were no database
backups in the previous 24-hour period.
a. Issue the QUERY PROCESS command to look at current status of an active
backup:
query process

Example output:
Process Process Description Status
Number
-------- -------------------- -------------------------------------------------
5 Database Backup TYPE=FULL in progress. 62,914,560 bytes
backed up to volume /fvt/kolty/srv/Storage/143-
[Link] .

7. Check to ensure that the DEVCONFIG and VOLUMEHISTORY files configured in the
[Link] file are current and up-to-date. Ensure that the file systems where
these files are being written to are not running out of space. If there are old or
unnecessary volume history entries, consider pruning the old entries using the
DELETE VOLHISTORY command.

Important: Save the volume history file to multiple locations. Ensure that these
different locations represent different underlying disks and file systems.
Related tasks:

Chapter 23. Daily monitoring tasks 781

 “Monitoring your server processes daily” on page 778
“Monitoring disk storage pools daily”
“Monitoring sequential access storage pools daily” on page 783
“Monitoring scheduled operations daily” on page 786
Chapter 17, “Managing the storage environment with the Operations Center,” on
page 587

Monitoring disk storage pools daily

Monitor disk storage pools daily to ensure that there are no problems, and that
your storage pools are fully operational and capable of handling the required
workload.

About this task

For detailed information about the commands mentioned here, see the
Administrator's Reference.

The following steps describe the commands that you can use to monitor disk
storage pools:

Procedure
1. Check the status of storage pools, and ensure that there is enough space
available.
v Examine the percent utilization to ensure that the amount of space is
sufficient for ingestion rates.
v The high and low migration thresholds should be set to values that will
allow for proper migration cycles.
v If the storage pool is set to CACHE=YES, the percent migration should be
approaching zero.
v This indicates that items are being cleared out of the pool appropriately.
Issue the QUERY STGPOOL command to display information about one or more
storage pools.
query stgpool

Example output:
Storage Device Estimated Pct Pct High Low Next
Pool Class Capacity UtilMigr Mig Mig Storage-
Name Name Pct Pct Pool
----------- ---------- ---------- ----- ----- ---- --- -----------
ARCHIVEPOOL DISK 1,000.0 M 0.0 0.0 90 70 storage_pool
BACKUPPOOL DISK 1,000.0 M 0.0 0.0 5 1 storage_pool

2. Check the status of the disk volumes. Issue the SELECT command and specify a
particular device class name:
select volume_name, status from volumes
where devclass_name=’devclass name’

Example output:
VOLUME_NAME: /fvt/kolty/srv/Storage/ar1
STATUS: ONLINE

VOLUME_NAME: /fvt/kolty/srv/Storage/bk1
STATUS: ONLINE

782 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 3. Check for any failed migrations that could contribute to space not being
released in a timely manner:
select start_time,end_time,activity as process,number as processnum,
entity as poolname from summary where activity=’MIGRATION’ and successful=’NO’
and end_time>(current_timestamp - interval ’24’ hours)

Example output:
START_TIME: 2011-08-23 [Link].000000
END_TIME: 2011-08-23 [Link].000000
PROCESS: MIGRATION
PROCESSNUM: 7
POOLNAME: storage_pool_example

Related tasks:
“Monitoring your server processes daily” on page 778
“Monitoring your database daily” on page 779
“Monitoring sequential access storage pools daily”
“Monitoring scheduled operations daily” on page 786
Chapter 17, “Managing the storage environment with the Operations Center,” on
page 587

Monitoring sequential access storage pools daily

Monitor sequential access storage pools daily to ensure that there are no problems.

About this task

For detailed information about the commands mentioned here, see the
Administrator's Reference.

The following steps describe the commands that you can use to monitor sequential
access storage pools:

Procedure
1. Check the status of your storage pools, and ensure that there is enough space
available. Examine the percent utilization to ensure that the amount of space is
sufficient for the amount of data that is being taken in. Set the high and low
migration thresholds to values that will allow for proper migration cycles.
Issue the QUERY STGPOOL command to display information about one or more
storage pools.
query stgpool

Example output:
Storage Device Estimated Pct Pct High Low Next Storage-
Pool Name Class Name Capacity UtilMigr Mig Mig Pool
Pct Pct
----------- ---------- ---------- ----- ----- ---- --- -----------
ARCHIVEPOOL DISK 1,000.0 M 0.0 0.0 90 70 storage_pool
BACKUPPOOL DISK 1,000.0 M 0.0 0.0 5 1 storage_pool

2. Check the status of the sequential access storage pool volumes with this SELECT
command:
select volume_name,status,access,write_errors,read_errors,
error_state from volumes where stgpool_name=’STORAGE_POOL_NAME’

Chapter 23. Daily monitoring tasks 783

 The select statement can be modified to limit the results based on error state,
read-write errors, or current-access state. Example output:
VOLUME_NAME: /fvt/kolty/srv/Storage/[Link]
STATUS: FULL
ACCESS: READONLY
WRITE_ERRORS: 0
READ_ERRORS: 0
ERROR_STATE: NO

VOLUME_NAME: /fvt/kolty/srv/Storage/[Link]
STATUS: FULL
ACCESS: READWRITE
WRITE_ERRORS: 0
READ_ERRORS: 0
ERROR_STATE: NO

3. Verify that all of the tapes have the appropriate write-access by issuing this
command:
select volume_name,access from volumes
where stgpool_name=’TAPEPOOL’ and access!=’READWRITE’

For example, this output indicates that the following volumes are not available
for use:
VOLUME_NAME: A00011L4
ACCESS: DESTROYED

VOLUME_NAME: KP0033L3
ACCESS: UNAVAILABLE

4. Use the QUERY DIRSPACE command to display information about free space in
the directories that are associated with a device class with a device type of
FILE.
query dirspace

Example ouput:
Device Class Directory Estimated Estimated
Name Capacity Available
------------ ------------------------------------ --------- ---------
FILECLASS /fvt/kolty/srv/Storage 253,952 M 185,616 M

Tip: Ensure that the amount of available space is higher than the total capacity
of all storage pools assigned to the device class or classes using that directory.
5. Determine how many scratch volumes are available in tape libraries with this
SELECT command:
select library_name,count(*) "Scratch volumes" from libvolumes
where status=’Scratch’ group by library_name

Example output:
LIBRARY_NAME Scratch volumes
------------------------- ----------------
TS3310 6

6. Determine how many scratch volumes can be potentially allocated out of the
storage pools using those tape libraries.
select stgpool_name,(maxscratch-numscratchused)
as "Num Scratch Allocatable" from stgpools
where devclass=’DEVICE_CLASS_NAME’

784 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 Example output:
STGPOOL_NAME Num Scratch Allocatable
-------------------------------- ------------------------
COPYPOOL 4
TAPEPOOL 2

Tip: Ensure that the number of allocatable scratch volumes is equal to the
number of available scratch library volumes in the assigned tape library.
7. Issue these SELECT commands to determine if there are any tape drives or paths
that are offline:
a. Check to ensure that the drives are online:
select drive_name,online from drives
where online<>’YES’

Example output:
DRIVE_NAME ONLINE
-------------------------------- -----------------------------------------
DRIVEA NO

b. Check to ensure that the paths to the drives are also online. A drive can be
online, while the path is offline.
select library_name,destination_name,online
from paths where online<>’YES’ and destination_type=’DRIVE’

Example output:
LIBRARY_NAME: TS3310
DESTINATION_NAME: DRIVEA
ONLINE: NO

8. Check to see if there are any library paths that are offline with this SELECT
command:
select destination_name,device,online from paths
where online<>’YES’ and destination_type=’LIBRARY’

Example output:
DESTINATION_NAME: TS3310
DEVICE: /dev/smc0
ONLINE: NO

9. If you are using the DRM, check the status and settings.
a. Check to see which copy storage pool volumes are onsite:
select stgpool_name,volume_name,upd_date,voltype from drmedia
where state in (’MOUNTABLE’,’NOTMOUNTABLE’)

Example output:
STGPOOL_NAME: COPYPOOL
VOLUME_NAME: CR0000L5
UPD_DATE: 2011-04-17 [Link].000000
VOLTYPE: Copy

b. Check to see if any volumes are eligible for return to onsite:

select stgpool_name,volume_name,upd_date,voltype from drmedia
where state=’VAULTRETRIEVE’

Chapter 23. Daily monitoring tasks 785

 Example output:
STGPOOL_NAME: COPYPOOL
VOLUME_NAME: CR0014L5
UPD_DATE: 2011-08-22 [Link].000000
VOLTYPE: Copy

c. Ensure that the DRM settings are valid:

select * from drmstatus

Example output:
PLANPREFIX:
INSTRPREFIX:
PLANVPOSTFIX: @
NONMOUNTNAME: NOTMOUNTABLE
COURIERNAME: COURIER
VAULTNAME: VAULT
DBBEXPIREDAYS: 60
CHECKLABEL: Yes
FILEPROCESS: No
CMDFILENAME:
RPFEXPIREDAYS: 60

Related tasks:
“Monitoring your server processes daily” on page 778
“Monitoring your database daily” on page 779
“Monitoring disk storage pools daily” on page 782
“Monitoring scheduled operations daily”
Chapter 17, “Managing the storage environment with the Operations Center,” on
page 587

Monitoring scheduled operations daily

Monitor scheduled operations daily to ensure that client and server scheduled
operations are completing successfully, and that no problems exist.

About this task

For detailed information about the commands mentioned here, see the
Administrator's Reference.

The following steps describe the commands that you can use to monitor scheduled
operations:

Procedure
1. The most valuable command that you can use to check the status of your
scheduled operations is the QUERY EVENT command. Issue this command and
look for any missed or failed scheduled operations that might indicate a
problem:
query event * * type=client
query event * type=admin

Tip: The EXCEPTIONSONLY parameter can be used to control what type of

information you want displayed on scheduled or completed events.
2. Use this SELECT command to specifically check for missed schedules:
select time(scheduled_start)"Start Time",node_name"Client"
from events where status=’Missed’

786 IBM Tivoli Storage Manager for Windows: Administrator's Guide

3. Use this SELECT command to specifically check for any failed schedules:
select time(scheduled_start)"Start Time",node_name"Client"
from events where status=’Failed’
4. Check the activity log for any unexpected messages that might indicate a
problem. Client messages are reported to the server with a prefix of ANE.
Review the messages to examine the client view of operations such as
successes, failures, and skipped files.
query actlog search=ANE????E begindate=today
query actlog search=ANE????W begindate=today
query actlog search=ANE????S begindate=today
5. Check the activity log for server messages, which use the ANR prefix:
query actlog search=ANR????E begindate=today
query actlog search=ANR????W begindate=today
query actlog search=ANR????S begindate=today
query actlog search=ANR9999D begindate=today
6. Check the summary table for various scheduled client operations such as
BACKUP, RESTORE, ARCHIVE, and RETRIEVE.
select * from summary where schedule_name is not null and end_time>
(current_timestamp - interval ’24’ hours) and activity in
(’BACKUP’,’ARCHIVE’,’RETRIEVE’,’RESTORE’)
The select command can be modified to limit the results based on schedule,
entity, success, failure, and so on. Analyze the results and compare to the
previous results to determine if there are any anomalies or other indicators of
potential issues.
7. Check for scheduled server processes such as MIGRATION, RECLAMATION,
and EXPIRATION. This SELECT command reports successes and failures for the
operations:
select * from summary where schedule_name is not null and end_time>
(current_timestamp - interval ’24’ hours) and activity in
(’EXPIRATION’,’MIGRATION’,’RECLAMATION’,’STGPOOL BACKUP’,
’FULL_DBBACKUP’,’INCR_DBBACKUP’)
Related tasks:
“Monitoring your server processes daily” on page 778
“Monitoring your database daily” on page 779
“Monitoring disk storage pools daily” on page 782
“Monitoring sequential access storage pools daily” on page 783
Chapter 17, “Managing the storage environment with the Operations Center,” on
page 587

Chapter 23. Daily monitoring tasks 787

788 IBM Tivoli Storage Manager for Windows: Administrator's Guide
Chapter 24. Basic monitoring methods
You can monitor operations using a variety of methods, such as using queries to
display information about settings and status, gain insight into database
operations, or display messages in the activity log.

Using IBM Tivoli Storage Manager queries to display information

IBM Tivoli Storage Manager provides QUERY commands to display formatted
information about definitions, settings, processes, and status.

About this task

For some commands, you can display the information in either a standard or
detailed format. The standard format presents less information than the detailed
format, and is useful in displaying an overview of many objects. For displaying
more information about a particular object, use the detailed format when
supported by a given command.

For information about creating customized queries of the database, see “Using SQL
to query the IBM Tivoli Storage Manager database” on page 794.

Requesting information about IBM Tivoli Storage Manager

definitions
During Tivoli Storage Manager system setup, an administrator can define many
objects, for example, storage management policies, storage pools, and device
classes. Tivoli Storage Manager provides queries that display information about
these objects.

About this task

Most of these definition queries let you request standard format or detailed format.
Standard format limits the information and usually displays it as one line per
object. Use the standard format when you want to query many objects, for
example, all registered client nodes. Detailed format displays the default and
specific definition parameters. Use the detailed format when you want to see all
the information about a limited number of objects.

Here is an example of the standard output for the QUERY NODE command:
Node Name Platform Policy Days Days Locked?
Domain Since Since
Name Last Password
Access Set
---------- -------- --------- ------ -------- -------
CLIENT1 AIX STANDARD 6 6 No
GEORGE Linux86 STANDARD 1 1 No
JANET HPUX STANDARD 1 1 No
JOE2 Mac STANDARD <1 <1 No
TOMC WinNT STANDARD 1 1 No

Here is an example of the detailed output for the QUERY NODE command:

© Copyright IBM Corp. 1993, 2013 789

 Node Name: JOE
Platform: WinNT
Client OS Level: 5.00
Client Version: Version 5, Release 1, Level 5.0
Policy Domain Name: STANDARD
Last Access Date/Time: 05/19/2002 [Link]
Days Since Last Access: 6
Password Set Date/Time: 05/19/2002 [Link]
Days Since Password Set: 6
Invalid Sign-on Count: 0
Locked?: No
Contact:
Compression: Client’s Choice
Archive Delete Allowed?: Yes
Backup Delete Allowed?: No
Registration Date/Time: 03/19/2002 [Link]
Registering Administrator: SERVER_CONSOLE
Last Communication Method Used: Tcp/Ip
Bytes Received Last Session: 108,731
Bytes Sent Last Session: 698
Duration of Last Session (sec): 0.00
Pct. Idle Wait Last Session: 0.00
Pct. Comm. Wait Last Session: 0.00
Pct. Media Wait Last Session: 0.00
Optionset:
URL: [Link]
Node Type: Client
Password Expiration Period: 60
Keep Mount Point?: No
Maximum Mount Points Allowed: 1
Auto Filespace Rename: No
Validate Protocol: No
TCP/IP Name: JOE
TCP/IP Address: [Link]
Globally Unique ID: 11.9c.54.e0.8a.b5.11.d6.b3.c3.[Link].c1.5b
Transaction Group Max: 0
Session Initiation: ClientOrServer
HLADDRESS:
LLADDRESS:

Requesting information about client sessions

When administrators or users access Tivoli Storage Manager, an administrative or
client node session is established with the server. The server assigns each client
session a unique session number.

About this task

You can use the QUERY SESSION command to request information about client
sessions. Figure 92 shows a sample client session report.

Sess Comm. Sess Wait Bytes Bytes Sess Platform Client Name
Number Method State Time Sent Recvd Type
------ ------ ------ ------ ------- ------- ----- -------- --------------------
3 Tcp/Ip IdleW 9 S 7.8 K 706 Admin WinNT TOMC
5 Tcp/Ip IdleW 0 S 1.2 K 222 Admin AIX GUEST
6 Tcp/Ip Run 0 S 117 130 Admin Mac2 MARIE

Figure 92. Information about client sessions

Check the wait time to determine the length of time (seconds, minutes, hours) the
server has been in the current state. The session state reports status of the session
and can be one of the following:
Start Connecting with a client session.

790 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 Run Running a client request.
End Ending a client session.
RecvW
Waiting to receive an expected message from the client while a database
transaction is in progress. A session in this state is subject to the
COMMTIMEOUT limit.
SendW
Waiting for acknowledgment that the client has received a message sent by
the server.
MediaW
Waiting for removable media to become available.
IdleW Waiting for communication from the client, and a database transaction is
not in progress. A session in this state is subject to the IDLETIMEOUT
limit.
For example, Tivoli Storage Manager cancels the client session if the
IDLETIMEOUT option is set to 30 minutes, and a user does not initiate
any operations within those 30 minutes. The client session is automatically
reconnected to the server when it starts to send data again.

Requesting information about server processes

When a command runs in the foreground (synchronous command execution), you
are unable to issue any other commands until the process completes. When a
command runs in the background (asynchronous command execution), you can
issue other commands while the process completes.

About this task

Most commands run in the foreground, but others generate background processes.
In some cases, you can specify that a process run in the foreground. Tivoli Storage
Manager issues messages that provide information about the start and end of
processes. In addition, you can request information about active background
processes. If you know the process ID number, you can use the number to limit the
search. However, if you do not know the process ID, you can display information
about all background processes by issuing the QUERY PROCESS command.

Figure 93 shows a server background process report after a DELETE FILESPACE

command was issued. The report displays a process ID number, a description, and
a completion status for each background process.

Process Process Description Status

Number
-------- ------------------------ ---------------------------------------------
2 DELETE FILESPACE Deleting filespace DRIVE_D for node CLIENT1:
172 files deleted.

Figure 93. Information about background processes

Chapter 24. Basic monitoring methods 791

 Requesting information about server settings
Any administrator can request general server information, most of which is
defined by SET commands, by issuing the QUERY STATUS command.

About this task

The displayed text includes a variety of information, such as:

v The server name and TCP/IP settings
v Server password and authentication settings
v Client node settings
v Activity log settings and status
v License audits and compliance status
v Client/server session limits
v Central scheduler settings and status
v Recovery log and backup trigger modes
v Refresh settings and status
v Table of contents retention period
v Machine globally unique ID at last startup
v Archive retention protection status
v Data encryption strength

This list is not all-inclusive. For a detailed explanation of the QUERY STATUS
command, see the Administrator's Reference.

Querying server options

Use the QUERY OPTION command to display information about one or more
server options.

About this task

Task Required Privilege Class
Query server options Any administrator

You can issue the QUERY OPTION command with no operands to display general
information about all defined server options. You also can issue it with a specific
option name or pattern-matching expression to display information on one or more
server options. You can set options by editing the server options file.

Options can also be set through the IBM Tivoli Storage Manager Console.

See the QUERY OPTION command in the Administrator's Reference for more
information.

792 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Querying the system
The QUERY SYSTEM command combines multiple queries of your Tivoli Storage
Manager system into a single command. This command can be used to collect
statistics and provide information for problem analysis by IBM service.

About this task

When you enter the QUERY SYSTEM command, the server issues the following
queries:
QUERY ASSOCIATION
Displays all client nodes that are associated with one or more client
schedules
QUERY COPYGROUP
Displays all backup and archive copy groups (standard format)
QUERY DB
Displays information about the database (detailed format)
QUERY DBSPACE
Displays display information about the directories used by the database.
QUERY DEVCLASS
Displays all device classes (detailed format)
QUERY DOMAIN
Displays all policy domains (standard format)
QUERY LOG
Displays information about the recovery log (detailed format)
QUERY MGMTCLASS
Displays all management classes (standard format)
QUERY OPTION
Displays all server options
QUERY PROCESS
Displays information about all active background processes
QUERY SCHEDULE
Displays client schedules (standard format)
QUERY SESSION
Displays information about all administrative and client node sessions in
standard format
QUERY STATUS
Displays general server parameters, such as those defined by SET
commands
QUERY STGPOOL
Displays information about all storage pools (detailed format)
QUERY VOLUME
Displays information about all storage pool volumes (standard format)
QUERY VOLHISTORY
Displays sequential volume history information that has been collected by
the server.
SELECT
Displays the results of two SQL queries:

Chapter 24. Basic monitoring methods 793

 select platform_name,count(*) from nodes group by platform_name
select stgpool_name,devclass_name,count(*) from volumes
group by stgpool_name,devclass_name

The first command displays the number of client nodes by platform.

The second command displays the name and associated device class of all
storage pools having one or more volumes assigned to them.

Using SQL to query the IBM Tivoli Storage Manager database

You can use a standard SQL SELECT statement to get information from the
database.

About this task

IBM Tivoli Storage Manager Versions 6.1 and later use the DB2 open database
connectivity (ODBC) driver to query the database and display the results.

DB2 provides its own ODBC driver which can also be used to access the Tivoli
Storage Manager server DB2 database. For more information on the DB2 native
ODBC driver, refer to DB2 documentation at: [Link]
db2luw/v10r5. Search on Introduction to DB2 CLI and ODBC

Using SELECT commands

SELECT commands allow you to create and format customized queries of the IBM
Tivoli Storage Manager database.

About this task

For SELECT statement syntax and guidelines, refer to DB2 documentation:

[Link] Search on the term
Select-statement.

You can issue the SELECT command from the command line of an administrative
client. You cannot issue this command from the server console.

Learning what information is available: system catalog tables

System catalog tables provide information about information that is available in
the database.

About this task

To help you find what information is available in the database, Tivoli Storage
Manager provides three system catalog tables:
[Link]
Contains information about all tables that can be queried with the SELECT
command.
[Link]
Describes the columns in each table.
[Link]
Defines the valid values for each enumerated type and the order of the
values for each type.

794 IBM Tivoli Storage Manager for Windows: Administrator's Guide

You can issue the SELECT command to query these tables and determine the
location of the information that you want. For example, to get a list of all tables
available for querying in the database TSMDB1 enter the following command:
select tabname from [Link] where tabschema=’TSMDB1’ and type=’V’

The results are:

TABNAME: ACTLOG
TABNAME: AF_VOL_SEGMENTS
TABNAME: ARCHDESC_NAMEVIEW
TABNAME: ARCHIVES
TABNAME: ARCHIVE_NAMEVIEW
TABNAME: AR_COPYGROUPS
TABNAME: ASSOCIATIONS
TABNAME: AS_VOLUME_ASSIGNMENT
TABNAME: BACKUPS
TABNAME: BACKUPSETS
TABNAME: BACKUP_NAMEVIEW
TABNAME: BU_COPYGROUPS
TABNAME: CLIENT_ADMINISTRATORS
TABNAME: CONTENTS
TABNAME: DB
TABNAME: DEVCLASSES
TABNAME: DF_VOL_CONTENTS
TABNAME: DRIVES
TABNAME: DRMSTATUS
TABNAME: EVENTS
TABNAME: FILESPACEVIEW
TABNAME: GROUPMEMBER
TABNAME: LIBRARIES

You can also issue the SELECT command to query columns. For example, to get a
list of columns for querying in the database TSMDB1 and the table name ACTLOG,
enter the following command:
select colname from [Link] where tabschema=’TSMDB1’and tabname=’ACTLOG’

The results are:

COLNAME: DATE_TIME
COLNAME: DOMAINNAME
COLNAME: MESSAGE
COLNAME: MSGNO
COLNAME: NODENAME
COLNAME: ORIGINATOR
COLNAME: OWNERNAME
COLNAME: PROCESS
COLNAME: SCHEDNAME
COLNAME: SERVERNAME
COLNAME: SESSID
COLNAME: SESSION
COLNAME: SEVERITY

Chapter 24. Basic monitoring methods 795

 Customizing queries using the SELECT command
With the SELECT command, you can customize a wide variety of queries.

About this task

This section shows two examples.

For many more examples of the command, see the Administrator's Reference.

Example 1: Find the number of nodes by type of operating system by issuing the
following command:
select platform_name,count(*) as "Number of Nodes" from nodes
group by platform_name

This command gives results like the following:

PLATFORM_NAME Number of Nodes

------------- ---------------
OS/2 45
AIX 90
Windows 35

Example 2: For all active client sessions, determine how long they have been
connected and their effective throughput in bytes per second:
select session_id as "Session", client_name as "Client", state as "State",
current_timestamp-start_time as "Elapsed Time",
(cast(bytes_sent as decimal(18,0)) /
cast(second(current_timestamp-start_time) as decimal(18,0)))
as "Bytes sent/second",
(cast(bytes_received as decimal(18,0)) /
cast(second(current_timestamp-start_time) as decimal(18,0)))
as "Bytes received/second"
from sessions

This command gives results like the following:

Session: 24
Client: ALBERT
State: Run
Elapsed Time: 4445.000000
Bytes sent/second: 564321.9302768451
Bytes received/second: 0.0026748857944

Session: 26
Client: MILTON
State: Run
Elapsed Time: 373.000000
Bytes sent/second: 1638.5284210992221
Bytes received/second: 675821.6888561849

Tip: When issuing SELECT * FROM DB statements, the output column

PHYSICAL_VOLUMES is included for legacy compatibility purposes only. The
number of physical volumes listed represents the DBSPACE locations defined to
the server.

For example:

796 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 DATABASE_NAME: mgsA62
TOT_FILE_SYSTEM_MB: 511872
USED_DB_SPACE_MB: 448
FREE_SPACE_MB: 452802
PAGE_SIZE: 16384
TOTAL_PAGES: 32772
USABLE_PAGES: 32636
USED_PAGES: 24952
FREE_PAGES: 768
BUFF_HIT_RATIO: 99.7
TOTAL_BUFF_REQ: 385557
SORT_OVERFLOW: 0
LOCK_ESCALATION: 0
PKG_HIT_RATIO: 99.8
LAST_REORG:
FULL_DEV_CLASS:
NUM_BACKUP_INCR: 0
LAST_BACKUP_DATE:
PHYSICAL_VOLUMES: 1

Using SELECT commands in Tivoli Storage Manager scripts

A Tivoli Storage Manager script is one or more commands that are stored as an
object in the database. You can define a script that contains one or more SELECT
commands.

About this task

A script can be run from an administrative client or the server console. You can
also include it in an administrative command schedule to run automatically. See
“Tivoli Storage Manager server scripts” on page 626 for details.

Tivoli Storage Manager is shipped with a file that contains a number of sample
scripts. The file, [Link], is in the server directory. To create and store the
scripts as objects in your server's database, issue the DSMSERV RUNFILE
command during installation:
> dsmserv runfile [Link]

You can also run the file as a macro from an administrative command line client:
macro [Link]

The sample scripts file contains Tivoli Storage Manager commands. These
commands first delete any scripts with the same names as those to be defined,
then define the scripts. The majority of the samples create SELECT commands, but
others do such things as back up storage pools. You can also copy and change the
sample scripts file to create your own scripts.

Here are a few examples from the sample scripts file:

def script q_inactive_days ’/* --------------------------------------------*/’
upd script q_inactive_days ’/* Script Name: Q_INACTIVE */’
upd script q_inactive_days ’/* Description: Display nodes that have not */’
upd script q_inactive_days ’/* accessed Tivoli Storage Manager for a */’
upd script q_inactive_days ’/* specified number of days */’
upd script q_inactive_days ’/* Parameter 1: days */’
upd script q_inactive_days ’/* Example: run q_inactive_days 5 */’
upd script q_inactive_days ’/* --------------------------------------------*/’
upd script q_inactive_days "select node_name,lastacc_time from nodes where -"
upd script q_inactive_days " cast((current_timestamp-lastacc_time)days as -"
upd script q_inactive_days " decimal) >= $1 "

Chapter 24. Basic monitoring methods 797

 /* Display messages in the activity log of severity X or Y */

def script q_msg_sev desc=’Show msgs in the activity log of severity X or Y’

upd script q_msg_sev ’/* ------------------------------------------------*/’
upd script q_msg_sev ’/* Script Name: Q_MSG_SEV */’
upd script q_msg_sev ’/* Description: Display messages in the */’
upd script q_msg_sev ’/* activity log that have either */’
upd script q_msg_sev ’/* of two specified severities. */’
upd script q_msg_sev ’/* Parameter 1: severity 1 */’
upd script q_msg_sev ’/* Parameter 2: severity 2 */’
upd script q_msg_sev ’/* where severity is I, W, E, S, or D */’
upd script q_msg_sev ’/* Example: run q_msg_sev S E */’
upd script q_msg_sev ’/* ------------------------------------------------*/’
upd script q_msg_sev "select date_time,msgno,message from actlog -"
upd script q_msg_sev " where severity=upper(’$1’) or severity=upper(’$2’)"

Querying the SQL activity summary table

You can query the SQL activity summary table to view statistics about client
operations and server processes.

About this task

Some of the client operations recorded to the table are BACKUP, RESTORE,
ARCHIVE and RETRIEVE. Server processes include MIGRATION,
RECLAMATION and EXPIRATION.

To list column names and their descriptions from the activity summary table, enter
the following command:
select colname,remarks from columns where tabname=’summary’

Here are a few example queries of the activity summary table.

v To display all events starting at 00:00 a.m. of the current day until the present
time, enter:
select * from summary

The result might look like this:

START_TIME: 2008-10-10 [Link].000000
END_TIME: 2008-10-10 [Link].000000
ACTIVITY: BACKUP
NUMBER: 10
ENTITY: NODE1
COMMMETH: Tcp/Ip
ADDRESS: [Link]
SCHEDULE_NAME:
EXAMINED: 3
AFFECTED: 3
FAILED: 0
BYTES: 36631067
IDLE: 0
MEDIAW: 0
PROCESSES: 2
SUCCESSFUL: YES
VOLUME_NAME:
DRIVE_NAME:
LIBRARY_NAME:
LAST_USE:
COMM_WAIT: 2
NUM_OFFSITE_VOLS:

ANS8002I Highest return code was 0.

798 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 v To display all events starting at or after 00:00 a.m. on October 10, 2008 until the
present time, enter:
select * from summary where start_time>=’2008-10-10 [Link]’

You can determine how long to keep information in the summary table. For
example, to keep the information for 5 days, enter the following command:
set summaryretention 5

To keep no information in the table, specify a value of 0.

Tivoli Storage Manager does not create records in the SQL activity summary table
for manual backups or for successful scheduled backups of 0 bytes. Records are
created in the summary table for successful scheduled backups only if data is
backed up.

Creating output for use by another application

You can redirect the output of SELECT commands for use in another program (for
example, a spreadsheet or database program). The use of command output
redirection and one of the delimited output format options lets you create queries
whose output can be further processed in other applications.

About this task

For example, based on the output of a SELECT command, a spreadsheet program

could produce graphs of average file sizes and file counts summarized by type of
client platform. When using another program for data formatting, the output to be
used should be written in a format that is easily processed. Two standard formats
for tabular data files are comma-separated values (CSV) and tab-separated values (TSV).
Most modern applications that can import tabular data can read one or both of
these formats.

Use the administrative client command line options -COMMADELIMITED or

-TABDELIMITED to select one of these formats for tabular query output. All
tabular output created during the administrative session will be formatted into
either comma-separated or tab-separated values.

For details about using command line options and redirecting command output,
see the Administrator's Reference.

Using the Tivoli Storage Manager activity log

The activity log contains messages that are normally sent to the server console
during server operation. The only exceptions are responses to commands entered
at the console, such as responses to QUERY commands.

About this task

Task Required Privilege Class
Request information from the activity log Any administrator
Set the activity log retention period System
Set the activity log size limit System

Examples of messages sent to the activity log include:

Chapter 24. Basic monitoring methods 799

 v When client sessions start or end
v When migration starts and ends
v When backup versions expire
v What data is exported to tape
v When expiration processing is performed
v What export or import processing is performed

You can also query the activity log for client session information. For example,
issue the following command to search the activity log for any messages that were
issued in relation to session 4:
query actlog search="(SESSION:4)"

This command displays the output of the query:

EXAMPLE 2 (Client SESSION that performs a backup)

09/23/2003 [Link] ANR0406I Session 4 started for node FRED (WinNT) (Tcp/Ip
colind(2463)). (SESSION: 4)
09/23/2003 [Link] ANR8493I FILE volume C:\CODE\522\[Link] mounted in
drive OUTFILE4 (FILE) in library OUTFILE. (SESSION: 4)
09/23/2003 [Link] ANR8340I FILE volume C:\CODE\522\[Link] mounted.
(SESSION: 4)
09/23/2003 [Link] ANR8468I FILE volume C:\CODE\522\[Link] dismounted
from drive OUTFILE4 (FILE) in library OUTFILE. (SESSION:4)
09/23/2003 [Link] ANR0403I Session 4 ended for node FRED (WinNT).
(SESSION: 4)

Any error messages sent to the server console are also stored in the activity log.

Use the following sections to adjust the size of the activity log, set an activity log
retention period, and request information about the activity log.

Requesting information from the activity log

You can request information stored in the activity log.

About this task

To minimize processing time when querying the activity log, you can:
v Specify a time period in which messages have been generated. The default for
the QUERY ACTLOG command shows all activities that have occurred in the
previous hour.
v Specify the message number of a specific message or set of messages.
v Specify a string expression to search for specific text in messages.
v Specify the QUERY ACTLOG command from the command line for large
queries instead of using the graphical user interface.
v Specify whether the originator is the server or client. If it is the client, you can
specify the node, owner, schedule, domain, or session number. If you are doing
client event logging to the activity log and are only interested in server events,
then specifying the server as the originator will greatly reduce the size of the
results.

For example, to review messages generated on May 30 between 8 a.m. and 5 p.m.,
enter:
query actlog begindate=05/30/2002 enddate=05/30/2002
begintime=08:00 endtime=17:00

800 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 To request information about messages related to the expiration of files from the
server storage inventory, enter:
query actlog msgno=0813

Refer to Messages for message numbers.

You can also request information only about messages logged by one or all clients.
For example, to search the activity log for messages from the client for node JEE:
query actlog originator=client node=jee

Setting a retention period for the activity log

You can use the SET ACTLOGRETENTION command to specify how long activity
log information is kept in the database.

About this task

Activity log management is retention-based when the optional parameter

MGMTSTYLE is set to its default value, DATE. The server automatically deletes
messages from the activity log once the number of days that are specified pass. At
installation, activity log management is retention-based, and the retention period is
set to 30 days. To change the retention period to 10 days, for example, enter:
set actlogretention 10

To disable activity log retention, set the SET ACTLOGRETENTION command to

zero. To display the current retention period and size of the activity log, query the
server status.

Note: With retention-based management, you lose some control over the amount
of space that the activity log occupies. For more information on size-based activity
log management, see “Setting a size limit for the activity log.”

Setting a size limit for the activity log

You can use size-based activity log management as an alternative to
retention-based management. This allows greater control over the amount of space
that the activity log occupies.

About this task

The server will periodically remove the oldest activity log records until the activity
log size no longer exceeds the configured maximum size allowed. To manage the
activity log by size, the parameter MGMTSTYLE must be set to the value SIZE. To
change the maximum size of the activity log to 12 MB, for example, enter:
set actlogretention 12 mgmtstyle=size

To disable activity log retention, set the SET ACTLOGRETENTION command to

zero. To display the current and maximum size of the activity log, query the server
status.

Note: With size-based management, you lose some control over the length of time
that activity log messages are kept. For more information on retention-based
activity log management, see “Setting a retention period for the activity log.”

Chapter 24. Basic monitoring methods 801

802 IBM Tivoli Storage Manager for Windows: Administrator's Guide
Chapter 25. Monitoring Tivoli Storage Manager accounting
records
Tivoli Storage Manager accounting records show the server resources that are used
during a session. This information lets you track resources that are used by a client
node session.

About this task

Task Required Privilege Class
Set accounting records on or off System

At installation, accounting defaults to OFF. You can turn accounting on by using

the SET ACCOUNTING command. When accounting is set to ON, the server
creates a session resource usage accounting record whenever a client node session
ends.

Accounting records are stored in a file named [Link], which is located in

the server directory.

The accounting file contains text records that can be viewed directly or can be read
into a spreadsheet program. The file remains opened while the server is running
and accounting is set to ON. The file continues to grow until you delete it or prune
old records from it. To close the file for pruning, either temporarily set accounting
off or stop the server.

There are 31 fields, which are delimited by commas (,). Each record ends with a
new-line character. Each record contains the following information:

Field Contents
1 Product version
2 Product sublevel
3 Product name, ‘ADSM',
4 Date of accounting (mm/dd/yyyy)
5 Time of accounting (hh:mm:ss)
6 Node name of Tivoli Storage Manager client
7 Client owner name (UNIX)
8 Client Platform
9 Authentication method used
10 Communication method used for the session
11 Normal server termination indicator (Normal=X'01', Abnormal=X'00')
12 Number of archive store transactions requested during the session
13 Amount of archived files, in kilobytes, sent by the client to the server
14 Number of archive retrieve transactions requested during the session
15 Amount of space, in kilobytes, retrieved by archived objects
16 Number of backup store transactions requested during the session
17 Amount of backup files, in kilobytes, sent by the client to the server
18 Number of backup retrieve transactions requested during the session
19 Amount of space, in kilobytes, retrieved by backed up objects
20 Amount of data, in kilobytes, communicated between the client node and the
server during the session

© Copyright IBM Corp. 1993, 2013 803

 Field Contents
21 Duration of the session, in seconds
22 Amount of idle wait time during the session, in seconds
23 Amount of communications wait time during the session, in seconds
24 Amount of media wait time during the session, in seconds
25 Client session type. A value of 1 or 4 indicates a general client session. A value
of 5 indicates a client session that is running a schedule. Values other than 1,
4, or 5 are reserved for the Tivoli Storage Manager server's internal use, and
you can ignore them.
26 Number of space-managed store transactions requested during the session
27 Amount of space-managed data, in kilobytes, sent by the client to the server
28 Number of space-managed retrieve transactions requested during the session
29 Amount of space, in kilobytes, retrieved by space-managed objects
30 Product release
31 Product level

The following shows a sample record:

3,8,ADSM,08/03/2000,[Link],node1,,AIX,1,Tcp/Ip,0,254,1713,0,0,47,1476,0,0,3316,
960,27,5,1,4,0,0,0,0,7,2

804 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Chapter 26. Reporting and monitoring with Tivoli Monitoring
for Tivoli Storage Manager
Tivoli Monitoring for Tivoli Storage Manager is not delivered with Tivoli Storage
Manager Version 7.1.

You can use Tivoli Monitoring for Tivoli Storage Manager Version 6.3.4 with any
Tivoli Storage Manager Version 5.5 or later server. For more information about
installing and configuring Tivoli Monitoring for Tivoli Storage Manager, see
Technote 1649883 at: [Link]
[Link]?uid=swg21649883.

© Copyright IBM Corp. 1993, 2013 805

806 IBM Tivoli Storage Manager for Windows: Administrator's Guide
Chapter 27. Logging IBM Tivoli Storage Manager events to
receivers
The server and client messages provide a record of Tivoli Storage Manager activity
that you can use to monitor the server. You can log server messages and most
client messages as events to one or more repositories called receivers.

About this task

You can log the events to any combination of the following receivers:
Tivoli Storage Manager server console and activity log
See “Logging events to the IBM Tivoli Storage Manager server console and
activity log” on page 809.
File and user exits
See “Logging events to a file exit and a user exit” on page 810.
Tivoli event console
See “Logging events to the Tivoli Enterprise Console” on page 811.
Event server receiver (Enterprise Event Logging)
Routes the events to an event server. See “Enterprise event logging:
logging events to another server” on page 820.
Simple Network Management Protocol (SNMP)
See “Logging events to an SNMP manager” on page 815.
The Windows Event Log
See “Logging events to the Windows event log” on page 820.

In addition, you can filter the types of events to be enabled for logging. For
example, you might enable only severe messages to the event server receiver and
one or more specific messages, by number, to another receiver. Figure 94 on page
808 shows a possible configuration in which both server and client messages are
filtered by the event rules and logged to a set of specified receivers.

© Copyright IBM Corp. 1993, 2013 807

 Tivoli Storage
Manager Server Activity Log

Server Console

Client
File
Messages Event
Rules
User Exit

Tivoli Event
Console

Event Server

Server
Messages

Figure 94. Event logging overview

Task Required Privilege Class

Enable or disable events System

Begin or end event logging

You can control event logging through the following actions:

1. Enable or disable logging for one or more event types and for one or more
receivers. See “Enabling and disabling events.”
2. Begin or end logging to one or more receivers. See “Beginning and ending
event logging” on page 809.

Enabling and disabling events

You can enable and disable events using the ENABLE EVENTS and DISABLE EVENTS
commands.

About this task

When you enable or disable events, you can specify the following:
v A message number or an event severity (ALL, INFO, WARNING, ERROR, or
SEVERE).
v Events for one or more client nodes (NODENAME) or for one or more servers
(SERVERNAME).

To enable or disable events, issue the ENABLE EVENTS and DISABLE EVENTS
commands. For example,
v To enable event logging to a user exit for all error and severe server messages,
enter:
enable events userexit error,severe
v To enable event logging to a user exit for severe client messages for all client
nodes, enter:
enable events userexit severe nodename=*
v To disable event logging to a user exit for error server messages, enter:

808 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 disable events userexit error
If you specify a receiver that is not supported on any platform, or if you specify an
invalid event or name, Tivoli Storage Manager issues an error message. However,
any valid receivers, events, or names that you specified are still enabled. Certain
events, such as messages that are issued during server startup and shutdown,
automatically go to the console. They do not go to other receivers, even if they are
enabled.

Note: Server messages in the SEVERE category and message ANR9999 can provide
valuable diagnostic information if there is a serious problem. For this reason, you
should not disable these messages. Use the SET CONTEXTMESSAGING ON command to
get additional information that could help determine the cause of ANR9999D
messages. The IBM Tivoli Storage Manager polls the server components for
information that includes process name, thread name, session ID, transaction data,
locks that are held, and database tables that are in use.

Beginning and ending event logging

The BEGIN EVENTLOGGING and END EVENTLOGGING commands can be used
to log events when event logging is not automatically started during server
startup.

About this task

At server startup, event logging begins automatically to the server console and
activity log and for any receivers that are started based on entries in the server
options file. A receiver for which event logging has begun is an active receiver.

To begin logging events to receivers for which event logging is not started
automatically, issue the BEGIN EVENTLOGGING command. You can also use this
command after you have disabled event logging to one or more receivers. To end
event logging for an active receiver issue the END EVENTLOGGING command.

For example,
v To begin logging events to the event server, enter:
begin eventlogging eventserver
v To end logging events to the event server, enter:
end eventlogging eventserver

Logging events to the IBM Tivoli Storage Manager server console and
activity log
Logging events to the server console and activity log begins automatically at server
startup.

About this task

Enabling client events to the activity log will increase the database utilization. You
can set a retention period or size limit for the log records by using the SET
ACTLOGRETENTION command (see “Setting a retention period for the activity
log” on page 801 and “Setting a size limit for the activity log” on page 801). At
server installation, activity log management is retention-based, and this value is set
to one day. If you increase the retention period or the size limit, utilization is

Chapter 27. Logging IBM Tivoli Storage Manager events to receivers 809
 further increased. For more information about the activity log, see “Using the
Tivoli Storage Manager activity log” on page 799.

You can disable server and client events to the server console and client events to
the activity log. However, you cannot disable server events to the activity log.
Also, certain messages, such as those issued during server startup and shutdown
and responses to administrative commands, will still be displayed at the console
even if disabled.

To enable all error and severe client events to the console and activity log, you can
issue the ENABLE EVENTS command. See the Administrator's Reference for more
information.

Logging events to a file exit and a user exit

A file exit is a file that receives all the information related to its enabled events.
You can log events to a file exit and a user exit.

About this task

Be aware that this file can rapidly grow in size depending on the events enabled
for it. There are two versions of the file exit: binary and text. The binary file exit
stores each logged event as a record, while the text file exit stores each logged
event as a fixed-sized, readable line. For more information about the text file exit,
see “Readable text file exit (FILETEXTEXIT) format” on page 826.

A user exit is an external interface in the form of an executable, user-written

program. Tivoli Storage Manager supports DLL user exits.

See “Adding a file exit or user exit option.”

Adding a file exit or user exit option

File and user exits receive event data in the same data block structure. Setting up
logging for these receivers is similar.

Procedure
1. Add an option for the exit to the server options file:
v For a file exit: Add either the FILEEXIT option (for a binary file exit) or
FILETEXTEXIT (for a text file exit) option.
a. Specify whether event logging to the file exit receiver begins
automatically at server startup. The parameters are YES and NO. If you
do not specify YES, you must begin event logging manually by issuing
the BEGIN EVENTLOGGING command.
b. Specify the file where each logged event is to be stored.
c. Specify how files will be stored if the file being stored already exists.
REPLACE will overwrite the existing file, APPEND will append data to
the existing file, and PRESERVE will not overwrite the existing file.
For example,
fileexit yes \tsm\server\data replace

filetextexit yes \tsm\server\data replace

v For a user exit: Add the USEREXIT option.
– Specify whether event logging to the user exit receiver begins
automatically at server startup. The parameters for this option are YES
810 IBM Tivoli Storage Manager for Windows: Administrator's Guide
 and NO. If you do not specify YES, you must begin event logging
manually by issuing the BEGIN EVENTLOGGING command.
– Specify a DLL name that contains the user-exit function.
– Specify the name of the user-exit function in the DLL.
– Specify the name of the user-exit function in the service program.
For example,
userexit yes [Link] dllmodulename
2. Enable events for the receiver. You must specify the name of the user exit in the
USEREXIT server option and the name of the file in the FILEEXIT server
option. Here are two examples:
enable events file error
enable events userexit error,severe
You can also enable events to one or more client nodes or servers by specify
the NODENAME OR SERVERNAME parameter. See “Enabling and disabling
events” on page 808 for more information.
3. If you did not specify YES in the server option, begin event logging. For
example, to begin event logging for a user-defined exit, enter:
begin eventlogging userexit

See “Beginning and ending event logging” on page 809 for more information.

Logging events to the Tivoli Enterprise Console

Tivoli Storage Manager includes the Tivoli receiver, a Tivoli Enterprise Console
adapter for sending events to the Tivoli Enterprise Console. You can specify the
events to be logged based on their source.

About this task

Application clients, Data Protection for IBM ESS for DB2, and Data Protection for
IBM ESS for Oracle must have enhanced Tivoli Enterprise Console support enabled
in order to route the events to the Tivoli Enterprise Console. Because of the
number of messages, you should not enable all messages from a node to be logged
to the Tivoli Enterprise Console.

The valid event names are:

Event Name Source

TSM_SERVER_EVENT Tivoli Storage Manager server
TSM_CLIENT_EVENT Tivoli Storage Manager clients
TSM_APPL_EVENT Tivoli Storage Manager application program
interface
TSM_TDP_DOMINO_EVENT Data Protection for Lotus Domino
TSM_TDP_EXCHANGE_EVENT Data Protection for Microsoft Exchange
Server
TSM_TDP_INFORMIX_EVENT Data Protection for Informix®
TSM_TDP_ORACLE_EVENT Data Protection for Oracle
TSM_TDP_SQL_EVENT Data Protection for Microsoft SQL Server
TSM_TDP_SAP_R3_EVENT Data Protection for [Link] Technology
and Data Protection for IBM ESS for
[Link] Technology

Chapter 27. Logging IBM Tivoli Storage Manager events to receivers 811
 Event Name Source
TSM_TDP_ESS_DB2_EVENT Data Protection for IBM ESS for DB2
TSM_TDP_ESS_ORACLE_EVENT Data Protection for IBM ESS for Oracle

Controlling the format of events

The server options UNIQUETECEVENTS and UNIQUETDPTECEVENTS are
available to control the format of events sent from the Tivoli Storage Manager
server to the Tivoli Enterprise Console.

About this task

Enabling either of these options not only changes the event class format, but also
generates a unique event class for individual Tivoli Storage Manager messages for
the client, the server, application clients, Data Protection for IBM ESS for DB2, Data
Protection for IBM ESS for Oracle, and Data Protection for IBM ESS for R/3.

Option Name Function

UNIQUETECEVENTS Changes the event class format and
generates a unique event class for client,
server, and some Data Protection messages
UNIQUETDPTECEVENTS Changes the event class format and
generates a unique event class for all client,
server, and all Data Protection messages

Setting UNIQUETDPTECEVENTS to YES will dynamically set

UNIQUETECEVENTS to YES. However, the Tivoli Storage Manager will not update
the server options file to reflect this change.

Depending on your particular environment, enabling one or both of these options

can facilitate rule-evaluation performance on the Tivoli Enterprise Console server.
Enabling one or both of these options can also incur a performance penalty during
event reception on the Tivoli Enterprise Console server. Test the options in your
own environment. It is possible that you might not benefit from enabling either of
these two options.

If the UNIQUETECEVENTS option is enabled, unique events have the following

event class format:
TSM_SERVER_ANR####
TSM_CLIENT_ANE####
TSM_APPL_ANE####
TSM_TDP_DOMINO_ACD####
TSM_TDP_EXCHANGE_ACN####
TSM_TDP_ORACLE_ANS####
TSM_TDP_INFORMIX_ANS####
TSM_TDP_SQL_ACO####

where #### represents the message number.

If the UNIQUETDPTECEVENTS option is enabled, the messages logged to the

Tivoli Storage Manager server for Data Protection for IBM ESS for DB2, Data
Protection for IBM ESS for Oracle, and Data Protection for R/3 will have unique
events with the following formats:
TSM_TDP_ESS_DB2_EEP####TSM_TDP_ESS_ORACLE_EEO####
TSM_TDP_SAP_R3_BKI#### (includes messages for Data Protection for IBM ESS for R/3)

812 IBM Tivoli Storage Manager for Windows: Administrator's Guide

where #### represents the message number. For exact details of the event class
format, look at the appropriate baroc file.

Application clients can issue unique events in the following ranges. All events
follow the IBM 3.4 naming convention, which uses a three-character prefix
followed by four digits.

Application client Event range

Data Protection for Microsoft Exchange ACN3500 to ACN3649
Server
Data Protection for Lotus Domino ACD5200 to ACD5299
Data Protection for Microsoft SQL Server ACO3000 to ACO3999
Data Protection for Oracle ANS0500 to ANS0599
Data Protection for Informix ANS0600 to ANS0699

If UNIQUETDPTECEVENTS is enabled, Data Protection for IBM ESS for DB2, Data
Protection for IBM ESS for Oracle, and Data Protection for R/3 can issue unique
events in the following ranges:

Application client Event range

Data Protection for IBM ESS for DB2 EEP0000 to EEP9999
Data Protection for IBM ESS for Oracle EEO0000 to EEO9999
Data Protection for R/3 and Data Protection BKI0000 to BKI9999
for IBM ESS for R/3

As noted, enabling UNIQUETDPTECEVENTS also enables UNIQUETECEVENTS.

This means that all recognized Data Protection messages will be sent as unique
events from the Tivoli Storage Manager server.

Based upon the setting of the option or options on the Tivoli Storage Manager
server, the Tivoli Enterprise Console administrator must create a rule base using
one of the following baroc files:

UNIQUETDPTECEVENTS UNIQUETECEVENTS Baroc File

Setting Setting
NO NO [Link]
NO YES [Link]
YES Defaults to YES because [Link]
UNIQUETDPTECEVENTS is
set to YES.

Each successive baroc file accepts the events of the previous baroc file. For
example, [Link] accepts all events in [Link], and [Link]
accepts all events contained in [Link].

Chapter 27. Logging IBM Tivoli Storage Manager events to receivers 813
 Encoding events to UTF-8
Tivoli Storage Manager supports the following option to encode a Tivoli Enterprise
Console event into UTF-8 before sending it to the Tivoli Enterprise Console server.
Some Tivoli Enterprise Console patches (for example, Patch 0004 for Tivoli
Enterprise Console Version 3.6 Modification 2) require UTF-8 encoded events for
some locales.

About this task

Option Name Function
TECUTF8EVENT Encodes a Tivoli Enterprise Console event
into UTF-8

To determine whether this option is enabled, issue the QUERY OPTION command.

Setting up a Tivoli Enterprise Console as a receiver

You can set up a Tivoli Enterprise Console as a receiver for event logging.

About this task

To set up Tivoli as a receiver for event logging, complete the following procedure:

Procedure
1. Define the Tivoli Storage Manager event classes to the Tivoli Enterprise Console
with the baroc file for your operating system:
[Link]
This file is distributed with the server.

Note: Please refer to Tivoli Enterprise Console documentation for instructions

on removing an existing baroc file, if needed, and installing a new baroc file.
Before the events are displayed on a Tivoli Enterprise Console, you must
import the baroc file into an existing rule base or create a new rule base and
activate it. To do this, complete the following steps:
a. From the Tivoli desktop, click on the Rule Base icon to display the pop-up
menu.
b. Select Import, then specify the location of the baroc file.
c. Select the Compile pop-up menu.
d. Select the Load pop-up menu and Load, but activate only when server
restarts from the resulting dialog.
e. Shut down the event server and restart it.
To create a new rule base, complete the following steps:
a. Click on the Event Server icon from the Tivoli desktop. The Event Server
Rules Bases window will open.
b. Select Rule Base from the Create menu.
c. Optionally, copy the contents of an existing rule base into the new rule base
by selecting the Copy pop-up menu from the rule base to be copied.
d. Click on the RuleBase icon to display the pop-up menu.
e. Select Import and specify the location of the baroc file.
f. Select the Compile pop-up menu.

814 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 g. Select the Load pop-up menu and Load, but activate only when server
restarts from the resulting dialog.
h. Shut down the event server and restart it.
2. To define an event source and an event group:
a. From the Tivoli desktop, select Source from the EventServer pop-up menu.
Define a new source whose name is Tivoli Storage Manager from the
resulting dialog.
b. From the Tivoli desktop, select Event Groups from the EventServer pop-up
menu. From the resulting dialog, define a new event group for Tivoli
Storage Manager and a filter that includes event classes
IBMTSMSERVER_EVENT and IBMTSMCLIENT_EVENT.
c. Select the Assign Event Group pop-up menu item from the Event Console
icon and assign the new event group to the event console.
d. Double-click on the Event Console icon to start the configured event
console.
3. Enable events for logging to the Tivoli receiver. See “Enabling and disabling
events” on page 808 for more information.
4. In the server options file, specify the location of the host on which the Tivoli
server is running. For example, to specify a Tivoli server at the IP address
[Link]:1555, enter the following:
techost [Link]
tecport 1555
5. Begin event logging for the Tivoli receiver. You do this in one of two ways:
v To begin event logging automatically at server start up, specify the following
server option:
tecbegineventlogging yes
Or
v Enter the following command:
begin eventlogging tivoli

See “Beginning and ending event logging” on page 809 for more information.

Logging events to an SNMP manager

IBM Tivoli Storage Manager supports the simple network management protocol
(SNMP) together with event logging.

About this task

You can do the following:

v Set up an SNMP heartbeat monitor to regularly check that the Tivoli Storage
Manager server is running.
v Send messages known as traps to an SNMP manager, such as NetView® or Tivoli
Enterprise Console.
v Run Tivoli Storage Manager scripts and retrieve output and return codes. See
“Tivoli Storage Manager server scripts” on page 626 for details.

Tivoli Storage Manager also implements an SNMP subagent that can be configured
to report exception conditions and provide support for a management information
base (MIB). The management information base (MIB), which is shipped with Tivoli
Storage Manager, defines the variables that will run server scripts and return the
server scripts' results. You must register SNMPADMIN, the administrative client

Chapter 27. Logging IBM Tivoli Storage Manager events to receivers 815
 the server runs these scripts under. Although a password is not required for the
subagent to communicate with the server and run scripts, a password should be
defined for SNMPADMIN to prevent access to the server from unauthorized users.
An SNMP password (community name) is required, however, to access the SNMP
agent, which forwards the request to the subagent.

Note: Because the SNMP environment has weak security, you should consider not
granting SNMPADMIN any administrative authority. This restricts SNMPADMIN
to issuing only Tivoli Storage Manager queries.

SNMP SET requests are accepted for the name and input variables associated with
the script names stored in the MIB by the SNMP subagent. This allows a script to
be processed by running a GET request for the ibmAdsm1ReturnValue and
ibmAdsm2ReturnValue variables. A GETNEXT request will not cause the script to
run. Instead, the results of the previous script processed will be retrieved. When an
entire table row is retrieved, the GETNEXT request is used. When an individual
variable is retrieved, the GET request is used.

Here is a typical Tivoli Storage Manager configuration with SNMP:

Procedure
1. Systems A, B, C: A Tivoli Storage Manager server communicates with a local
subagent.
2. System D: A DPI-enabled SNMP agent is installed. This is required for
communication between the Tivoli Storage Manager SNMP subagent,
dsmsnmp, and the SNMP Manager you are using. A DPI-enabled SNMP agent
is available as part of the AIX operating system.
3. System E: An SNMP manager, such as NetView, is installed.
4. The subagents on systems A, B, and C communicate with the agent on system
D.
5. The agent on system D forwards SNMP traps to NetView on system E.

Results

To run an arbitrary command from an SNMP management application, for

example, NetView, follow these steps:
1. Choose the name and parameters for a Tivoli Storage Manager script.
2. Use the application to communicate with the SNMP agent. This agent changes
the Tivoli Storage Manager MIB variable for one of the two script names that
the Tivoli Storage Manager subagent maintains. The SNMP agent also sets the
parameter variables for one of the two scripts.
3. Use the application to retrieve the variable ibmAdsmReturnValue1.x or
ibmAdsmReturnValue2.x, where x is the index of the server that is registered
with the subagent.

To set the variables associated with the script, the nodes on which the subagent
and the agent are run must have read-write authority to the MIB variables. This is
done through the SNMP configuration process on the system that the SNMP agent
runs on.

An SNMP agent is needed for communication between an SNMP manager and its
managed systems. The SNMP agent is realized through the snmpd daemon. The
Distributed Protocol Interface (DPI) Version 2 is an extension of this SNMP agent.

816 IBM Tivoli Storage Manager for Windows: Administrator's Guide

SNMP managers can use the MIB that is shipped with Tivoli Storage Manager to
manage the server. Therefore, an SNMP agent supporting DPI Version 2 must be
used to communicate with the Tivoli Storage Manager subagent. This SNMP agent
is not included with Tivoli Storage Manager. A supported DPI agent ships with
AIX. The Tivoli Storage Manager subagent is included with Tivoli Storage Manager
and, before server startup, must be started as a separate process communicating
with the DPI-enabled SNMP agent.

The SNMP manager system can reside on the same system as the Tivoli Storage
Manager server, but typically would be on another system connected through
SNMP. The SNMP management tool can be any application, such as NetView or
Tivoli Enterprise Console, which can manage information through SNMP MIB
monitoring and traps. The Tivoli Storage Manager server system runs the processes
needed to send Tivoli Storage Manager event information to an SNMP
management system. The processes are:
v SNMP agent (snmpd)
v Tivoli Storage Manager SNMP subagent (dsmsnmp)
v Tivoli Storage Manager server (dsmserv)

Cross-system support for communication between the subagent and agent is

supported, and in some cases required. Figure 95 illustrates a typical Tivoli Storage
Manager implementation:

SNMP Manager
Windows Linux
System
Tivoli Storage Tivoli Storage
AIX Manager server Manager server
SNMP Manager

SNMP Protocol SNMP Protocol

SNMP Subagent SNMP Subagent

SNMP Agent SNMP DPI SNMP DPI

SNMP DPI

SNMP DPI SNMP DPI SNMP DPI

SNMP Subagent SNMP Subagent SNMP Subagent

Tivoli Storage Tivoli Storage Tivoli Storage

Manager server Manager server Manager server

Solaris HP-UX

Figure 95. Tivoli Storage Manager SNMP Implementation

Figure 96 on page 818 shows how the communication for SNMP works in a Tivoli
Storage Manager system:
v The SNMP manager and agent communicate with each other through the SNMP
protocol. The SNMP manager passes all requests for variables to the agent.
v The agent then passes the request to the subagent and sends the answer back to
the manager. The agent responds to the manager's requests and informs the
manager about events by sending traps.
v The agent communicates with both the manager and subagent. It sends queries
to the subagent and receives traps that inform the SNMP manager about events
taking place on the application monitored through the subagent. The SNMP
agent and subagent communicate through the Distributed Protocol Interface
(DPI). Communication takes place over a stream connection, which typically is a
TCP connection but could be another stream-connected transport mechanism.

Chapter 27. Logging IBM Tivoli Storage Manager events to receivers 817
 v The subagent answers MIB queries of the agent and informs the agent about
events by sending traps. The subagent can also create and delete objects or
subtrees in the agent's MIB. This allows the subagent to define to the agent all
the information needed to monitor the managed application.

SNMP Manager
get/set
respond
SNMP Protocol SNMP Protocol
trap

SNMP Agent
query
reply
SNMP DPI register SNMP DPI
trap

SNMP Subagent

Figure 96. Manager-Agent-Subagent communication

Note:
1. You can start dsmsnmp and the server in any order. However, starting dsmsnmp
first is more efficient in that it avoids retries.
2. The MIB file name is [Link]. The file name is located in the directory in
which the server is installed.
3. Merge the contents of the [Link] file into the /etc/[Link] file.

Configuring IBM Tivoli Storage Manager SNMP

You can configure SNMP by completing the following procedure.

About this task

The IBM Tivoli Storage Manager SNMP set up procedure is illustrated by Figure 97
on page 819:

818 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 SNMP Manager
AIX
SNMP Protocol SNMP Protocol

snmpd

SNMP DPI

Windows
SNMP DPI

dsmsnmp
[Link]
Tivoli Storage
Manager server

Figure 97. IBM Tivoli Storage Manager SNMP Set Up

To set up Tivoli Storage Manager monitoring through SNMP, do the following:

Procedure
1. Modify the server options file to specify the SNMP communication method.
Figure 98 displays an example of a SNMP communication method setting in the
server options file.
You must specify the COMMMETHOD and SNMPSUBAGENT options. The
SNMPSUBAGENT option must specify a host that is an AIX system with a
DPI-enabled SNMP agent, such as the SystemView agent.

commmethod snmp
snmpsubagent hostname jimbo communityname public timeout 600
snmpsubagentport 1521
snmpheartbeatinterval 5
snmpmessagecategory severity

Figure 98. Example of SNMP communication method options

For details about server options, see the server options section in
Administrator's Reference.
2. Install, configure, and start the SNMP agent as described in the documentation
for that agent. The SNMP agent must support the DPI Version 2.0 standard.
Tivoli Storage Manager supports the SNMP agent that is built into the AIX
operating system.
Before starting the agent, ensure that the dpid2 and snmpd subsystems have
been started.

Important: The dpid2 subsystem is started automatically through the snmpd

subsystem in SNMP, V3. The dpid2 subsystem must be manually started
independently of the snmpd subsystem with SNMP, V1.
3. Start the Tivoli Storage Manager SNMP subagent by running the dsmsnmp
executable.
4. Start the Tivoli Storage Manager server to begin communication through the
configured TCP/IP port with the subagent.
5. Begin event logging for the SNMP receiver, and enable events to be reported to
SNMP. For example, issue the following commands:

Chapter 27. Logging IBM Tivoli Storage Manager events to receivers 819
 begin eventlogging snmp
enable event snmp all
6. Define the Tivoli Storage Manager SNMP MIB values for the SNMP manager to
help format and display the Tivoli Storage Manager SNMP MIB variables and
messages. The [Link] file ships with the Tivoli Storage Manager server
and must be loaded by the SNMP manager. This file is in the installation
directory of the server. For example, when you run NetView for Windows as an
SNMP manager, the [Link] file is copied to the \netview_path\SNMP_MIB
directory and then loaded through the following command:
[C:\] loadmib -load [Link]

Logging events to the Windows event log

The Windows event log lets you display enabled events in the Windows
Application Log in the Windows Event Viewer.

About this task

The information displayed includes:

v The date and time that the Windows Event Log received the event
v The category of the event
v The event number
v The user security identifier
v The name of the computer that logged the event

To enable severe and error events for logging on the Event Log, you can issue the
ENABLE EVENTS command. For example:
enable events nteventlog severe,error

Enterprise event logging: logging events to another server

One or more servers can send server events and events from their own clients to
another server for logging.

About this task

The sending server receives the enabled events and routes them to a designated
event server. This is done by a receiver that IBM Tivoli Storage Manager provides.
At the event server, an administrator can enable one or more receivers for the
events being routed from other [Link] 99 on page 821 shows the
relationship of a sending Tivoli Storage Manager server and a Tivoli Storage
Manager event server.

820 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 Tivoli Storage Tivoli Storage Manager
Manager Server Event Server
Activity Log

Server Console
Client
Messages Event Event File
Rules EVENTS Rules

User Exit

Tivoli Event
Console

Server
Messages

Figure 99. Server-to-server event logging

The following scenario is a simple example of how enterprise event logging can
work.

The administrator at each sending server does the following:

1. Defines the server that will be the event server. For details about
communication set up, see “Setting up communications for enterprise
configuration and enterprise event logging” on page 692.
define server server_b password=cholla hladdress=[Link] lladdress=1505
2. Identifies the server just defined as the event server:
define eventserver server_b
3. Enables the logging of severe, error, and warning server messages from the
sending server and severe and error messages from all clients to the event
server receiver by issuing the following commands:
enable events eventserver severe,error,warning
enable events eventserver severe,error nodename=*
4. Begins event logging by issuing the following command:
begin eventlogging eventserver
The administrator at the event server does the following:
5. Enables the logging of severe and error messages to a file named events that
are sent to it from the sending servers. The administrator defines the file with
the following option in the server options file:
fileexit yes events append

Then the administrator enables the events by issuing the ENABLE EVENTS
command for each sending server. For example, for SERVER_A the
administrator would enter:
enable events file severe,error servername=server_a

Note: By default, logging of events from another server is enabled to the event
server activity log. However, unlike events originating from a local server,
events originating from another server can be disabled for the activity log at an
event server.

Chapter 27. Logging IBM Tivoli Storage Manager events to receivers 821
 One or more servers can send events to an event server. An administrator at the
event server enables the logging of specific events from specific servers. In the
previous example, SERVER_A routes severe, error, and warning messages to
SERVER_B. SERVER_B, however, logs only the severe and error messages. If a
third server sends events to SERVER_B, logging is enabled only if an ENABLE
EVENTS command includes the third server. Furthermore, the SERVER_B
determines the receiver to which the events are logged.

Attention: It is important that you do not set up server-to-server event logging in

a loop. In such a situation, an event would continue logging indefinitely, tying up
network and memory resources. Tivoli Storage Manager will detect such a
situation and issue a message. Here are a few configurations to avoid:
v SERVER_A logs to SERVER_B, and SERVER_B logs to SERVER_A.
v SERVER_A logs to SERVER_B; SERVER_B logs to SERVER_C; SERVER_C logs to
SERVER_A.

Querying event logging

You can use the QUERY ENABLED command to display a list of server or client
events that are enabled or disabled by a specified receiver.

About this task

Because the lists of enabled and disabled events could be very long, Tivoli Storage
Manager displays the shorter of the two lists.

For example, assume that 1000 events for client node HSTANFORD were enabled
for logging to the user exit and that later two events were disabled. To query the
enabled events for HSTANFORD, you can enter:
query enabled userexit nodename=hstanford

The output would specify the number of enabled events and the message names of
disabled events:
998 events are enabled for node HSTANFORD for the USEREXIT receiver.
The following events are DISABLED for the node HSTANFORD for the USEREXIT
receiver:
ANE4000, ANE49999

The QUERY EVENTRULES command displays the history of events that are
enabled or disabled by a specific receiver for the server or for a client node.
query enabled userexit nodename=hstanford

User exit and file exit receivers

The data structure of the user exit receivers applies to the file exit receivers. To use
one of these exits with Tivoli Storage Manager, you must specify the corresponding
server option (FILEEXIT, FILETEXTEXIT, or USEREXIT) in the server options file.

The samples for the C, H, and make files are shipped with the server code in the
\win32app\ibm\adsm directory.

You can also use Tivoli Storage Manager commands to control event logging. For
details, see Chapter 27, “Logging IBM Tivoli Storage Manager events to receivers,”
on page 807 and Administrator's Reference.

822 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 Sample user-exit declarations
USEREXITSAMPLE.H contains declarations for a user-exit program.

Figure 100. Sample user exit declarations

/***********************************************************************
* Name: USEREXITSAMPLE.H
* Description: Declarations for a user exit
* Environment: WINDOWS NT
***********************************************************************/

#ifndef _H_USEREXITSAMPLE
#define _H_USEREXITSAMPLE

#include <stdio.h>
#include <sys/types.h>

/***** Do not modify below this line *****/

#define BASE_YEAR 1900

typedef short int16;

typedef int int32;

#ifndef uchar
typedef unsigned char uchar;
#endif
/* DateTime Structure Definitions - TSM representation of a timestamp */

typedef struct
{
uchar year; /* Years since BASE_YEAR (0-255) */
uchar mon; /* Month (1 - 12) */
uchar day; /* Day (1 - 31) */
uchar hour; /* Hour (0 - 23) */
uchar min; /* Minutes (0 - 59) */
uchar sec; /* Seconds (0 - 59) */
} DateTime;
/******************************************
* Some field size definitions (in bytes) *
******************************************/

#define MAX_SERVERNAME_LENGTH 64
#define MAX_NODE_LENGTH 64
#define MAX_COMMNAME_LENGTH 16
#define MAX_OWNER_LENGTH 64
#define MAX_HL_ADDRESS 64
#define MAX_LL_ADDRESS 32
#define MAX_SCHED_LENGTH 30
#define MAX_DOMAIN_LENGTH 30
#define MAX_MSGTEXT_LENGTH 1600

/**********************************************
* Event Types (in [Link]) *
**********************************************/

#define TSM_SERVER_EVENT 0x03 /* Server Events */

#define TSM_CLIENT_EVENT 0x05 /* Client Events */

/***************************************************
* Application Types (in [Link]) *
***************************************************/

#define TSM_APPL_BACKARCH 1 /* Backup or Archive client */

Chapter 27. Logging IBM Tivoli Storage Manager events to receivers 823
 #define TSM_APPL_HSM 2 /* Space manage client */
#define TSM_APPL_API 3 /* API client */
#define TSM_APPL_SERVER 4 /* Server (ie. server to server ) */

/*****************************************************
* Event Severity Codes (in [Link]) *
*****************************************************/

#define TSM_SEV_INFO 0x02 /* Informational message. */

#define TSM_SEV_WARNING 0x03 /* Warning message. */
#define TSM_SEV_ERROR 0x04 /* Error message. */
#define TSM_SEV_SEVERE 0x05 /* Severe error message. */
#define TSM_SEV_DIAGNOSTIC 0x06 /* Diagnostic message. */
#define TSM_SEV_TEXT 0x07 /* Text message. */

/************************************************************
* Data Structure of Event that is passed to the User-Exit. *
* The same structure is used for a file receiver *
************************************************************/

typedef struct evRdata

{
int32 eventNum; /* the event number. */
int16 sevCode; /* event severity. */
int16 applType; /* application type (hsm, api, etc) */
int32 sessId; /* session number */
int32 version; /* Version of this structure (1) */
int32 eventType; /* event type */
*(TSM_CLIENT_EVENT, TSM_SERVER_EVENT)*/
DateTime timeStamp; /* timestamp for event data. */
uchar serverName[MAX_SERVERNAME_LENGTH+1]; /* server name */
uchar nodeName[MAX_NODE_LENGTH+1]; /* Node name for session */
uchar commMethod[MAX_COMMNAME_LENGTH+1]; /* communication method */
uchar ownerName[MAX_OWNER_LENGTH+1]; /* owner */
uchar hlAddress[MAX_HL_ADDRESS+1]; /* high-level address */
uchar llAddress[MAX_LL_ADDRESS+1]; /* low-level address */
uchar schedName[MAX_SCHED_LENGTH+1]; /*sched name if applicable */
uchar domainName[MAX_DOMAIN_LENGTH+1];/*omain name for node */
uchar event[MAX_MSGTEXT_LENGTH]; /* event text */
} elEventRecvData;

/************************************
* Size of the Event data structure *
************************************/

#define ELEVENTRECVDATA_SIZE sizeof(elEventRecvData)

/*************************************
* User Exit EventNumber for Exiting *
*************************************/

#define USEREXIT_END_EVENTNUM 1822 /* Only user-exit receiver to exit*/

#define END_ALL_RECEIVER_EVENTNUM 1823 /* All receivers told to exit */

/**************************************
*** Do not modify above this line. ***
**************************************/

/********************** Additional Declarations **************************/

#endi

824 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 Sample user exit program
USEREXITSAMPLE.C is a sample user-exit program invoked by the Tivoli Storage
Manager server.

Figure 101. Sample user exit program

/***********************************************************************
* Name: USEREXITSAMPLE.C
* Description: Example user-exit program that is invoked by
* the TSM V3 Server
* Environment: *********************************************
* ** This is a platform-specific source file **
* ** versioned for: "WINDOWS NT" **
* *********************************************
***********************************************************************/

#include <stdio.h>
#include <stdlib.h>
#include <sys/types.h>
#include <io.h>
#include <windows.h>
#include "USEREXITSAMPLE.H"

/**************************************
*** Do not modify below this line. ***
**************************************/

#define DllExport __declspec(dllexport)

/****************
*** DLL MAIN ***
****************/

BOOL WINAPI
DllMain(HMODULE hMod, DWORD fdwReason, LPVOID lpvReserved)
{
return(TRUE);
} // End of WINAPI

/******************************************************************
* Procedure: adsmV3UserExit
* If the user-exit is specified on the server, a valid and
* appropriate event will cause an elEventRecvData structure
* (see USEREXITSAMPLE.H) to be passed to a procedure named
* adsmV3UserExit that returns a void.
*
* This procedure can be named differently:
* ----------------------------------------
* The procedure name must match the function name specified in
* the server options file (4th arg). The DLL name generated from
* this module must also match in the server options file
* (3rd arg).
* INPUT : A (void *) to the elEventRecvData structure
* RETURNS: Nothing
******************************************************************/

DllExport void adsmV3UserExit( void *anEvent )

{
/* Typecast the event data passed */
elEventRecvData *eventData = (elEventRecvData *)anEvent;

/**************************************
*** Do not modify above this line. ***
**************************************/

Chapter 27. Logging IBM Tivoli Storage Manager events to receivers 825
 fprintf(stderr,"NT UserExit: Got event

if( ( eventData->eventNum == USEREXIT_END_EVENTNUM ) ||

( eventData->eventNum == END_ALL_RECEIVER_EVENTNUM ) )
{
/* Server says to end this user-exit. Perform any cleanup, *
* but do NOT exit() !!! */
return;
}

/* Field Access: eventData->.... */

/* Your code here ... */

/* Be aware that certain function calls are process-wide and can cause
* synchronization of all threads running under the TSM Server process!
* Among these is the system() function call. Use of this call can
* cause the server process to hang and otherwise affect performance.
* Also avoid any functions that are not thread-safe. Consult your
* system’s programming reference material for more information.
*/

return; /* For picky compilers */

} /* End of adsmV3UserExit() */

Readable text file exit (FILETEXTEXIT) format

If you specify the readable text file exit (FILETEXTEXIT), each logged event is
written to a fixed-size, readable line.

The following table presents the format of the output. Fields are separated by
blank spaces.
Table 65. Readable text file exit (FILETEXTEXIT) format
Column Description
0001-0006 Event number (with leading zeros)
0008-0010 Severity code number
0012-0013 Application type number
0015-0023 Session ID number
0025-0027 Event structure version number
0029-0031 Event type number
0033-0046 Date/Time (YYYYMMDDDHHmmSS)
0048-0111 Server name (right padded with spaces)
1
0113-0176 Node name
1
0178-0193 Communications method name
1
0195-0258 Owner name
1
0260-0323 High-level internet address (n.n.n.n)
1
0325-0356 Port number from high-level internet address
1
0358-0387 Schedule name
1
0389-0418 Domain name
0420-2019 Event text
2020-2499 Unused spaces

826 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Table 65. Readable text file exit (FILETEXTEXIT) format (continued)
Column Description
2500 New line character
1
Columns 113 - 418 contain data only for events that originate in a client or in another
Tivoli Storage Manager server. Otherwise, columns 113 - 418 contain blanks.

Chapter 27. Logging IBM Tivoli Storage Manager events to receivers 827
828 IBM Tivoli Storage Manager for Windows: Administrator's Guide
Part 6. Protecting the server
Disasters, by their very nature, cannot be predicted, in either their intensity, timing,
or long-term effects. The ability to recover from a disaster, if one occurs, is
essential. To protect your system infrastucture and data and to recover from a
disaster, use the tools and procedures that Tivoli Storage Manager provides.

© Copyright IBM Corp. 1993, 2013 829

830 IBM Tivoli Storage Manager for Windows: Administrator's Guide
Chapter 28. Managing Tivoli Storage Manager security
Administrators have specific activities to manage IBM Tivoli Storage Manager
security.

The security of your data is the most important aspect of managing data. You can
control access to the server and client nodes, encrypt data transmission, and
protect administrator and node passwords through authentication processes. The
two methods of authentication are LOCAL and LDAP. The LOCAL password
authentication takes place on the Tivoli Storage Manager server, and those
passwords are not case-sensitive.

LDAP password authentication takes place on the LDAP directory server, and the
passwords are case-sensitive. When using LDAP authentication, the password is
sent to the server by the client. By default, Secure Sockets Layer (SSL) is required
when LDAP authentication is used, to avoid exposing the password. SSL is used
when authenticating the server to the client and secures all communication
between the client and server. You can choose not to use SSL with LDAP
authentication if other security measures are in place to protect the password. One
example of an alternative security measure is a virtual private network (VPN)
connection.
Related concepts:
“Managing Tivoli Storage Manager administrator IDs” on page 844
“Managing passwords and logon procedures” on page 852
“Securing the server console” on page 843
“Securing sensitive client data” on page 539
Related reference:
“Managing access to the server and clients” on page 844
“Administrative authority and privilege classes” on page 845

Securing communications
You can add more protection for your data and passwords by using Secure Sockets
Layer (SSL).

SSL is the standard technology for creating encrypted sessions between servers and
clients. SSL provides a secure channel for servers and clients to communicate over
open communication paths. With SSL, the identity of the server is verified through
the use of digital certificates.

To ensure better system performance, use SSL only for sessions when it is needed.
Consider adding additional processor resources on the Tivoli Storage Manager
server to manage the increased requirements.

If you authenticate passwords with an LDAP directory server, Transport Layer

Security (TLS) protects passwords between the Tivoli Storage Manager server and
the LDAP server. TLS, a form of SSL, is required for all LDAP password
communications. See “Configuring SSL or TLS for LDAP directory servers” on
page 839.

© Copyright IBM Corp. 1993, 2013 831

 Secure Sockets Layer and Transport Layer Security
communication
You can use the Secure Sockets Layer (SSL) or the Transport Layer Security (TLS)
protocol to provide transport layer security for a secure connection between Tivoli
Storage Manager servers, clients, and storage agents. If you send data between the
server, client, and storage agent, use SSL or TLS to encrypt the data.

Tip: Any Tivoli Storage Manager documentation that indicates "SSL" or to "select
SSL" applies to TLS.

SSL is provided by the Global Security Kit (GSKit) that is installed with the Tivoli
Storage Manager server that the server, client, and storage agent use. The
Operations Center and Reporting agent do not use GSKit.

Each Tivoli Storage Manager server, client, or storage agent that enables SSL must
use a trusted self-signed certificate or obtain a unique certificate that is signed by a
certificate authority (CA). You can use your own certificates or purchase certificates
from a CA. Either certificate can be installed and added to the key database on the
Tivoli Storage Manager server, client, or storage agent. The certificate is verified by
the SSL client or server that requests or initiates the SSL communication.

SSL is set up independently on the Tivoli Storage Manager server, client, and
storage agent.

The Tivoli Storage Manager server, client, or storage agent can serve as SSL clients
during communication. An SSL client is the component that initiates
communication and verifies the certificate for an SSL server. For example, if a
Tivoli Storage Manager client initiates the SSL communication with a Tivoli Storage
Manager server, the Tivoli Storage Manager client is the SSL client and the server
is the SSL server.

For more information about the components that can be an SSL client or SSL
server, see Table 66.
Table 66. SSL clients and servers in a Tivoli Storage Manager environment
SSL client SSL server Description
Client Server The Tivoli Storage Manager client initiates a communication
request with the Tivoli Storage Manager server. The client
verifies the certificate. The server provides the certificate.
Server (such Server (such The Tivoli Storage Manager source server initiates a
as a source as a target communication request with the Tivoli Storage Manager
server) server) target server. The source server acts as an SSL client and
verifies the certificate that the target server provides.

This type of communication is common during replication

processing.

832 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 Table 66. SSL clients and servers in a Tivoli Storage Manager environment (continued)
SSL client SSL server Description
Client Server The client uses the Tivoli Storage Manager storage agent to
through a verify the certificate that the Tivoli Storage Manager server
storage agent provides.

When the storage agent communicates with the server by

using the SSL communication protocol, the storage agent acts
as an SSL client and verifies the certificate that the server
provides.

The storage agent can be the SSL client and the SSL provider
at the same time.
Server LDAP server The Tivoli Storage Manager server initiates a communication
request with the LDAP server. The Tivoli Storage Manager
server acts as the SSL client and verifies the certificate that
the LDAP server provides.
Operations Server The Operations Center initiates a communication request
Center with the Tivoli Storage Manager server. The Operations
Center acts as the SSL client and verifies the certificate that
the Tivoli Storage Manager server provides.
Reporting Server The Reporting agent initiates a communication request with
the Tivoli Storage Manager server. The Reporting feature acts
as the SSL client and verifies the certificate that the Tivoli
Storage Manager server provides.

Setting up SSL
You can set up SSL or TLS on the Tivoli Storage Manager server, backup-archive
client, and storage agent to ensure that your data is encrypted during
communication. You can use an SSL certificate to verify an SSL communication
request between the server, client, and storage agent.

Procedure

To configure Tivoli Storage Manager servers and clients for SSL or TLS, complete
the following steps:
1. Specify the TCP/IP port on which the server waits for client communications
that are enabled for SSL or TLS. You can use the SSLTCPADMINPORT option or
SSLTCPPORT option, or both, to specify TLS port numbers. The options are
stored in the [Link] file.
2. Create the key database file if it does not exist. Complete the following steps to
create the key database file for the server, client, and storage agent:
v For Tivoli Storage Manager V6.3.3 and later servers, the [Link] file and
other files that are related to SSL or TLS are created when the server is first
started. Tivoli Storage Manager automatically creates the server key database
file, [Link]. If a password exists for the server database, it is reused for the
[Link] key database. After you create the database, the key database access
password is generated and stored.
v To create the key database file for the storage agent, issue the DSMSTA
SETSTORAGESERVER command and specify the SSL=YES and
STAKEYDBPW=password parameters.
v To create the key database file, [Link], for the client, issue the following
command in the bin directory on the client:

Chapter 28. Managing security 833

 gsk8capicmd_64 -keydb -create -populate
-db [Link] -pw password -stash
3. Use one of the following certificates for SSL or TLS communication:
Self-signed certificate
You must import a .arm file for the server, backup-archive client, and
storage agent according to the default label that is used for the server
self-signed certificate. The following table shows you which file to
import:
Table 67. Determining the .arm file to use
Import this file for Import this file for
Default label in the Import this file for server-server storage agent-server
key database clients communication communication
“TSM Server [Link] [Link] [Link]
SelfSigned Key”
“TSM Server [Link] [Link] [Link]
SelfSigned SHA Key”

Important: To use TLS 1.2, the default label must be “TSM Server
SelfSigned SHA key”. You must specify the SSLTLS12 YES server option
in the server options file and the storage agent options file, if necessary.
CA-signed certificate
You must obtain a unique certificate that is signed by a CA or use a
trusted self-signed certificate for each server that enables SSL or TLS.
Backup-archive clients use the [Link] or [Link] files to import
the self-signed certificates, which the server automatically generates.
4. Manually transfer the appropriate Tivoli Storage Manager server .arm file to the
client computers. If you transfer the [Link] file, you must first change the
default certificate in the [Link] file to the “TSM Server SelfSigned SHA Key”
label. To change the default certificate, issue the following command from the
server instance directory:
gsk8capicmd_64 -cert -setdefault -db [Link]
-stashed -label "TSM Server SelfSigned SHA Key"
5. Using a backup-archive client user ID, specify the ssl yes and tcpport options
in the client options file:
v [Link]
The server is normally set up for SSL or TLS connections on a different port. If
you use an SSL or TLS connection, two ports are open on the server. One port
accepts regular non-SSL or non-TLS client connections and the other port
accepts SSL or TLS connections only.
6. If you want to use a certificate that is issued by a certificate authority (CA),
you do not need to complete steps 4 and 5. Install the CA root certificate on all
clients. A set of default root certificates are preinstalled if you specified the
-populate parameter in the command when you created the key database file.
Related reference:
“Adding a certificate to the key database” on page 836

834 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Specifying communication ports
The Tivoli Storage Manager server can be configured to listen on four TCP/IP
ports: two for regular protocols and two for the Transport Layer Security (TLS)
protocols.

If you authenticate passwords with an LDAP directory server, the connection

between the LDAP directory server and the Tivoli Storage Manager server must be
protected. The connection between a Tivoli Storage Manager server and an LDAP
directory server defaults to port 389. You do not have to use this port number and
can define the port by setting the LDAPURL option.

For IPv4 or IPv6, the COMMMETHOD server option must specify either TCPIP or
V6TCPIP. The server options for TLS communications are SSLTCPPORT and
SSLTCPADMINPORT. The server can listen on separate ports for the following
communications:
v Backup-archive clients that use the regular protocol
v Administrator IDs that use the regular protocol
v Backup-archive clients that use the TLS protocol
v Administrator IDs that use the TLS protocol
Use the TCPADMINPORT and SSLTCPADMINPORT options to separate administrative
client traffic from regular client traffic on TCPPORT and SSLTCPPORT options. If the
TCPADMINPORT and SSLTCPADMINPORT options are not used, administrative traffic
flows on client ports as well.

You can use the following components with TLS:

v Command-line client
v Administrative command-line client
v Backup-archive client graphical user interface (GUI)
v Client API

If the ADMINONCLIENTPORT option is set to NO, TLS administrative client sessions

require that you specify the SSLTCPADMINPORT option with a port number other than
the one specified by the SSLTCPPORT option. The SSLTCPPORT and SSLTCPADMINPORT
options do not affect the TCPPORT or TCPADMINPORT options and their interaction
with the ADMINONCLIENTPORT option.

The backup-archive client user decides which protocol to use and which port to
specify in the [Link] file for the SSLTCPADMINPORT option. If the
backup-archive client requires TLS authentication but the server is not in TLS
mode, the session fails.
Related concepts:
“Managing passwords and logon procedures” on page 852
Related tasks:
“Configuring Tivoli Directory Server for TLS on the iKeyman GUI” on page 840
“Configuring Tivoli Directory Server for TLS on the command line” on page 841
Related reference:
“Configuring Windows Active Directory for TLS/SSL” on page 842

Chapter 28. Managing security 835

 Adding a certificate to the key database
To use Transport Layer Security (TLS), the certificate must be installed on the
server, and for some root certificates, they must be installed on the client. Each
server that enables TLS must obtain a unique certificate signed by a certificate
authority (CA) or use a trusted self-signed certificate.

You can use your own certificates or purchase certificates from a CA. Either can be
installed and added to the key database. If you include the -stashpw parameter on
a GSKit gsk8capicmd_64 command, the password that you define is saved for later
use. The server and the storage agent create the key databases by using the stash
facility.

The key database is created when you start the Tivoli Storage Manager server. If
the certificate is signed by a trusted CA, obtain the certificate, install it in the key
database, and restart the server. Because the certificate is provided by a trusted
authority, the certificate is accepted by Tivoli Storage Manager and communication
between server and client can start.

For example, to add a certificate to the key database for the client, issue the
following command:
gsk8capicmd_64 -cert -add -label "TSM061" -format ascii
-file [Link] -db [Link] -pw password

The [Link] file is generated by the V6.3 server for distribution to the V6.3 or
later backup-archive clients, and for use in server-server communication and
storage agent-server communication. The [Link] file might also be generated by
the V6.3 server, but is not designed for passwords that authenticate with an LDAP
server.

If a certificate expires, the certificate is rejected when you attempt SSL

communication. You must obtain a new certificate and add it to the key database
for the server and the storage agent. The self-signed certificates for the server are
created with an expiration time of 10 years. After 10 years, the certificates must be
re-created and redistributed. You can re-create a self-signed certificate by deleting
the [Link] and [Link] files and by deleting the self-signed certificates from
the key database. When you restart the server or the storage agent, new certificates
and .arm files are created.

The first time that you enter a password you can use the -stashpw password
parameter. The password is then saved in the *.sth file and retrieved when you
use the -stashed parameter in later commands. You can use the -stashpw password
parameter with the CREATE and POPULATE parameters.

To show the certificates that are available to use for SSL or TLS communication,
issue the following command from the client directory:
gsk8capicmd_64 -cert -list -db [Link] -stashed
Related concepts:
“Managing passwords and logon procedures” on page 852

836 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Adding a CA-signed SSL certificate:

You can use a Transport Layer Security (TLS) certificate if the client trusts the
certificate authority (CA). Trust is established when you add a signed certificate to
the server key database and use a root certificate for the CA in the client key
database.

About this task

The Global Security Kit (GSKit) is included in the Tivoli Storage Manager server
installation. The backup-archive client and server communicate with SSL through
services provided by GSKit.

Procedure

Complete the following steps to add a certificate to the key database using GSKit:
1. Obtain a signed, server key database certificate from your CA.
2. To receive the signed certificate and make it the default for communicating
with clients, issue the following command:
gsk8capicmd_64 -cert -receive -db [Link]
-pw password -stash -file cert_signed.arm -default_cert yes

The server key database file name is [Link].

Important: If your client operating system is 32-bit, replace gsk8capicmd_64

with gsk8capicmd in all GSKit commands.
3. Restart the server.
4. Transfer the root certificate ([Link]) to the client directory.
5. To add the root certificate to the client key database, issue the gsk8capicmd_64
-cert -add command. For example:
gsk8capicmd_64 -cert -add -db [Link]
-pw password -label "my CA"
-file [Link] -format ascii

Tip: For this example, the client key database name is [Link].
6. To verify that the client can successfully connect, issue the dsmc query session
command.

Troubleshooting the certificate key database

Backup copies of the [Link] file ensure that Transport Layer Security (TLS) starts
when you restore the Tivoli Storage Manager server. If you have a backup copy,
you can restore the file and restart the server.

Procedure

If you do not have a backup copy of the [Link] file, perform the following steps:
1. Issue the DELETE KEYRING server command to delete the entry for it that is
located in the Tivoli Storage Manager database.
2. Delete all remaining cert.* files.
3. Shut down the server.
4. Start the server. The server automatically creates a new [Link] file and a
corresponding entry in the Tivoli Storage Manager database. If you do not issue
the DELETE KEYRING command, the server attempts, on startup, to create the key
database with the previous password.

Chapter 28. Managing security 837

 5. Redistribute the new [Link] file to all backup-archive clients that are using
TLS. Reinstall any third-party certificates on the backup-archive client. If you
are using an LDAP directory server to authenticate passwords, add the root
certificate that was used to sign the LDAP server’s certificate. If the root
certificate is already a default trusted certificate, you do not have to add it
again.

Configuring a storage agent and server to use SSL

You can set up a storage agent and the Tivoli Storage Manager server to use the
SSL communication method. SSL is set up independently on both the storage agent
and the Tivoli Storage Manager server.

Procedure

To set up the storage agent to use SSL communication with the Tivoli Storage
Manager server and client, complete the following steps:
1. On the storage agent, issue the DSMSTA SETSTORAGESERVER command to
initialize the storage agent and add communication information to the device
configuration file and the storage agent options file [Link]:

Hint: The following command is entered on one line, but is displayed here on
multiple lines to make it easier to read.
dsmsta setstorageserver myname=sta
mypa=sta_password
myhla=ip_address
servername=server_name
serverpa=server_password
hla=ip_address
lla=ssl_port
STAKEYDBPW=password
ssl=yes

Requirement:
v When you set the SSL=YES and STAKEYDBPW=password parameters, a key
database file is set up in the storage agent options file, [Link]. All
passwords are obfuscated in [Link].
v To enable SSL communication, ensure that the Tivoli Storage Manager LLA
parameter specifies the server SSLTCPADMIN port. If the SSLTCPADMIN is not
specified, use the SSLTCPPORT port instead. Set the SSL parameter to YES.
2. Import the Tivoli Storage Manager server certificate, [Link], to the key
database file for the storage agent. Ensure that the required SSL certificates are
in the key database file that belongs to each storage agent that uses SSL
communication. To import the SSL certificate, switch to the storage agent
directory and issue the following command:
gskcapicmd_64 -cert -add -label server_example_name
-db [Link] -stashed -file [Link] -format ascii
3. Specify the SSLTCPPORT and the SSLTCPADMINPORT options in the [Link]
options file.
4. Create the key database certificate and default certificates by starting the
storage agent.

Tip: To provide the new password to the storage agent, you must change the
key database password and then issue the DSMSTA SETSTORAGESERVER
command.

838 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 a. Open a command window and change the password by issuing the
following command:
gsk8capicmd_64 -keydb -changepw -db [Link] -pw oldpw -newpw newpw
b. Rerun the DSMSTA SETSTORAGESERVER command and specify the
STAKEYDBPW=newpassword parameter.
5. On the Tivoli Storage Manager server, issue the following command:
define server sta
hla=ip_address
lla=ssl_port
serverpa=password
ssl=yes
6. Stop the storage agent.
7. Stop the Tivoli Storage Manager server.
8. Import the [Link] certificate from the storage agent to the key database
file for the Tivoli Storage Manager server. Ensure that the required SSL
certificates are in the key database file that belongs to each server that uses
SSL communication before you restart the server. To import the SSL certificate
from the storage agent, issue the following command:
gskcapicmd_64 -cert -add -label server_example_name
-db [Link] -stashed -file [Link] -format ascii
9. Stop and restart the Tivoli Storage Manager server.
10. Restart the storage agent.

Results

When the Tivoli Storage Manager server and storage agent initiate communication,
SSL certificate information is displayed to indicate that SSL is in use.
Related reference:
“Adding a certificate to the key database” on page 836

Configuring SSL or TLS for LDAP directory servers

Before you can authenticate passwords by using an LDAP directory server, you
must configure Secure Sockets Layer (SSL) or Transport Layer Security (TLS) to use
the directory server. You must use a trusted certificate that is verified during SSL
or TLS communication.

TLS is provided by the Global Security Kit (GSKit) that is installed with the Tivoli
Storage Manager server.

If you use an LDAP directory server to authenticate passwords, you can use the
SSL or TLS protocol to secure connections between the Tivoli Storage Manager
server and the LDAP server. When you configure the Tivoli Storage Manager
server to work with an LDAP directory server, you must consider the directory
server to use before you configure SSL or TLS.

The LDAP directory server must supply a trusted certificate to the Tivoli Storage
Manager server. If the Tivoli Storage Manager server determines that the certificate
is valid, an SSL or TLS connection is established. If not, the connection fails. The
root certificate that is used with the LDAP Directory server certificate must be
added to the key database file for the Tivoli Storage Manager server. If the root
certificate is not added, the LDAP Directory server certificate cannot be used.

The directory servers that are available are IBM Tivoli Directory Server V6.2 or 6.3
or Windows Active Directory 2003 or 2008. You can configure Tivoli Directory

Chapter 28. Managing security 839

 Server with the graphical user interface or with the command line interface.

Configuring Tivoli Directory Server for TLS on the iKeyman GUI:

Configuring IBM Tivoli Directory Server is one of the preliminary tasks you must
do before you can authenticate passwords with an LDAP directory server. The
Tivoli Directory Server can use a self-signed certificate to secure the
communication between server and backup-archive client, and the LDAP directory
server.

About this task

You can use the iKeyman graphical user interface (GUI) to set up Tivoli Directory
Server. If the Tivoli Storage Manager server already has a trusted certificate from
your LDAP server, you do not have to complete the steps that are documented
here. If the LDAP directory server already has a signed certificate, you do not have
to complete these steps.

Procedure

To configure Tivoli Directory Server for Transport Layer Security (TLS) by using
the iKeyman GUI, complete the following steps:
1. Install and configure Java Runtime Environment 1.4.1 or later before you
install Tivoli Directory Server.
2. Ensure that you have GSKit version 8 installed. See Instructions to install,
uninstall, and upgrade GSKit 8 ([Link]
[Link]?uid=swg21577384).
3. Set the JAVA_HOME environmental variable to point to the Java subdirectory
of the Tivoli Directory Server installation directory. The installation directory is
/opt/IBM/ldap/Vx.x, where x.x is the version number.
4. Ensure that the local_policy.jar and US_export_policy.jar files are in the
$JAVA_HOME/jre/lib/security directory.
5. Verify whether the Content Management system (CMS) provider is registered.
Open the $JAVA_HOME/jre/lib/security/[Link] file to see whether the
entry exists. If the entry does not exist, open the [Link] file and add
the following entry:
[Link].X=[Link]

where X is the next number in the order. For example:

[Link].1=[Link]
[Link].2=[Link]
[Link].3=[Link]
6. Create the key database file in the Tivoli Directory Server installation directory
as Tivoli Directory Server instance user. Complete the following steps:
a. Open a command prompt and enter ikeyman to start the Java utility.
b. Click Key Database File > New for a new key database file or click Open
if the key database file exists.
c. Specify the key database type (CMS) and the key database file name, with
its location. For example, server_file.kdb. Click OK.
d. Supply the password for the key database file when you are prompted.
e. Select the Stash a password to a file check box and click OK.
7. Click Create > New Self-Signed Certificate.
8. Extract the certificate.

840 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 a. On the server where Tivoli Directory Server is, open the server_file.kdb
file.
b. Click Personal Certificates > Extract Certificate.
c. Enter the file name and location.
d. Save the certificate to a text-based (ASCII) file.
9. Click Server administration > Manage security properties > Key database.
Specify the key database path and file name. If you defined a password stash
file, it is assumed to have the same file specification, with an extension of
.sth. See step 6e on page 840.
10. Specify the key password. If you are not using a stashed password, specify
the password for the key database file.
11. Specify the key label, which indicates the part of the key database that is to be
used.
12. Copy the certificate file ([Link]) to the Tivoli Storage Manager server
key database.

Configuring Tivoli Directory Server for TLS on the command line:

You must configure IBM Tivoli Directory Server before you can authenticate
passwords with an LDAP directory server. The Tivoli Directory Server can use a
self-signed certificate to secure the communication between server and
backup-archive client, and the LDAP directory server.

About this task

If the Tivoli Storage Manager server already has a trusted certificate from your
LDAP server, you do not have to complete the steps that are documented here. If
the LDAP directory server already has a signed certificate, you do not have to
complete these steps.

Procedure

To configure Tivoli Directory Server for Transport Layer Security (TLS), complete
the following steps:
1. Using the Tivoli Directory Server instance user name, create the key database
by issuing the following command:
gsk8capicmd_64 -keydb -create -db "directory/[Link]"
-pw “pa$$=w0rd” -stashpw -populate
2. Create a self-signed certificate or get one from a certificate authority (CA). To
create a self-signed certificate, issue the following command:
gsk8capicmd_64 -cert -create -db "directory/[Link]" -stashed -label
"LDAP_directory_server" -dn "cn=[Link]"
-san_dnsname [Link] -size 2048
-sigalg SHA256WithRSA -expire 3650
3. Extract the certificate to a file by issuing the following command:
gsk8capicmd_64 -cert -extract -db "directory/[Link]" -stashed -label
"LDAP_directory_server" -target [Link] -format ascii
4. Copy the certificate file ([Link]) to the Tivoli Storage Manager server.
5. To add the certificate to the Tivoli Storage Manager server key database, issue
the following command from the Tivoli Storage Manager server. You must issue
the command from the instance user ID from the instance directory.
gsk8capicmd_64 -cert -add -db "[Link]" -stashed -label
"LDAP_directory_server" -format ascii -file [Link]

Chapter 28. Managing security 841

 6. Configure the key database file to work with Tivoli Directory Server. To set the
key database for TLS, issue the following command:
idsldapmodify -D <adminDN> -w <adminPW> -i <filename>

where filename contains the following entry:

dn: cn=SSL,cn=Configuration
changetype: modify
replace: ibm-slapdSSLKeyDatabase
ibm-slapdSSLKeyDatabase: <databasename>
-
replace: ibm-slapdSSLKeyDatabasePW
ibm-slapdSSLKeyDatabasePW: <password>
-
replace: ibm-slapdSslKeyRingFilePW
ibm-slapdSslKeyRingFilePW: <password>
7. Enable TLS on the LDAP directory server so that the Tivoli Storage Manager
server can run the “start TLS extended” operation. This operation secures
LDAP communication. Do not enable client authentication with TLS, because
the Tivoli Storage Manager server does not support certificate authentication as
an LDAP client.

Tip: The Tivoli Storage Manager server authenticates with the “LDAP simple
password authentication” method.
8. Restart Tivoli Directory Server and Tivoli Storage Manager server.

Configuring Windows Active Directory for TLS/SSL:

You must configure Windows Active Directory before the Tivoli Storage Manager
server can authenticate passwords.

To set up the Windows Active Directory server, complete the following steps:
1. Turn off automatic root certificate updates to Windows Update if your
Windows Active Directory server does not have access to the internet.
2. Synchronize the system times of the Tivoli Storage Manager server and the
Windows Active Directory system. You can use a Network Time Protocol (NTP)
server. For more information about synchronizing the system times, see your
operating system documentation. You can also see the Microsoft website for
information about synchronizing Active Directory (http://
[Link]/en-us/library/cc786897).
3. Set up Transport Layer Security (TLS) for LDAP server connections. Go to the
Microsoft website ([Link] and search for LDAP and SSL.
a. Obtain a signed certificate. Active Directory requires that a signed certificate
be in the Windows certificate store to enable TLS. You can obtain a signed
certificate from the following sources:
v A third-party certificate authority (CA)
v Install the Certificate Services role on a system that is joined to the Active
Directory domain and configure an enterprise root CA

Important: Do not install the Certificate Services role on the Active

Directory server. Some Active Directory Domain configurations are not
suited to accept an installed Certificate Services role. For example, when
there are multiple Active Directory servers in a domain.
b. Add the signed certificate to the Windows certificate store on the Windows
Active Directory server.

842 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 4. Determine whether the Tivoli Storage Manager server trusts the CA. As the
instance owner from the instance directory, issue the following command:
gsk8capicmd_64 -cert -list -db [Link] -stashed
gsk8capicmd_64 -cert -details -label "My CA" -db [Link] -stashed
5. If the CA is not trusted, export the CA certificate from the certificate server. You
can select either Distinguished Encoding Rules (DER) binary or Base64-encoded
ASCII.

Tip: To determine whether the file is DER binary or ASCII, open the certificate
in a text editor. If you can read the characters, then the file is ASCII.
Ensure that you have the root certificate and that the subject on the certificate
matches the CA name. The “Issued by” and “Issued to/subject” for the root
certificate must be the same. Export the CA certificate by using one of the
following methods:
v Export the CA certificate from the “Certificates (Local Computer)” Microsoft
Management Console (MMC) snap-in.
v Copy the certificate from C:\Windows\system32\certsrv\CertEnroll\*.crt
into the server key database. The file is in DER binary format.
v Download the CA certificate file from the Certificate Services web interface
[Link] server hostname>/certsrv/, if it is enabled through
the Certificate Enrollment Web Services.
6. Copy the certificate to the Tivoli Storage Manager server.
7. Add the certificate to the Tivoli Storage Manager server key database. Issue the
following command from the Tivoli Storage Manager server as the instance
user from the instance directory:
gsk8capicmd_64 -cert -add -db "[Link]" -stashed
-label "My CA" -format ascii -file [Link]

Tip: The Tivoli Storage Manager server authenticates with the “LDAP simple
password authentication” method.
Related tasks:
Setting up TLS

Securing the server console

At installation, the server console is defined with a special user ID, which is
named SERVER_CONSOLE. This name is reserved and cannot be used by another
administrator.

An administrator with system privilege can revoke or grant new privileges to the
SERVER_CONSOLE user ID. However, an administrator cannot update, lock,
rename, or remove the SERVER_CONSOLE user ID. The SERVER_CONSOLE user
ID does not have a password.

Therefore, you cannot use the user ID from an administrative client unless you set
authentication to off.

Chapter 28. Managing security 843

Managing access to the server and clients
An administrator can control access to the server and clients through a number of
methods.

Table 68 describes the typical tasks for managing access to the server and clients.
Table 68. Managing access
Task Details
Allow a new administrator to access the 1. “Registering administrator IDs” on page
server 847
2. “Granting authority to administrators” on
page 848
Modify authority for registered “Managing Tivoli Storage Manager
administrators administrator IDs”
Give a user authority to access a client “Managing client access authority levels” on
remotely page 447
Give an administrator authority to create a “Generating client backup sets on the server”
backup set for a client node on page 544
Prevent administrators from accessing the “Locking and unlocking administrator IDs
server from the server” on page 852
Prevent new sessions with the server, but “Disabling or enabling access to the server”
allow current sessions to complete on page 471
Prevent clients from accessing the server “Locking and unlocking client nodes” on
page 440
Change whether passwords are required to “Disabling the default password
access IBM Tivoli Storage Manager authentication” on page 865
Change requirements for passwords v “Modifying the default password
expiration period for passwords that are
managed by the Tivoli Storage Manager
server” on page 860
v “Setting a limit for invalid password
attempts” on page 863
v “Setting a minimum length for a
password” on page 864
Prevent clients from initiating sessions within “Server-initiated sessions” on page 433
a firewall
Tip: For information on connecting with IBM
Tivoli Storage Manager across a firewall, refer
to the Installation Guide.

Managing Tivoli Storage Manager administrator IDs

The administrator can register other administrator IDs, grant levels of authority to
them, rename or remove them, and lock and unlock them from the server.

A privilege class is a level of authority granted to an administrator. The privilege

class determines the administrative tasks that the administrator can complete.

Task Required Privilege Class

Registering an administrator System
Granting administrative authority System

844 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Task Required Privilege Class
Modifying administrative authority System
Updating information about other System
administrators
Updating information about yourself Any administrator
Displaying information about administrators Any administrator
Renaming an administrator user ID System
Removing administrators System
Locking or unlocking administrators from System
the server

Related reference:
“Administrative authority and privilege classes”

Administrative authority and privilege classes

After administrators are registered, they can perform a limited set of tasks. By
default, administrators can request command-line help and issue queries.

To perform other tasks, administrators must be granted authority by being

assigned one or more administrative privilege classes. Privilege classes determine
the authority level for an administrator. Figure 102 on page 846 illustrates the
privilege classes. An administrator with system privilege class can perform any
task with the server. Administrators with policy, storage, operator, or node
privileges can perform subsets of tasks.

Important: Two server options give you additional control over the ability of
administrators to perform tasks.
v QUERYAUTH allows you to select the privilege class that an administrator must
have to issue QUERY and SELECT commands. By default, no privilege class is
required. You can change the requirement to one of the privilege classes,
including system.
v REQSYSAUTHOUTFILE allows you to specify that system authority is required for
commands that cause the server to write to an external file (for example,
BACKUP DB). By default, system authority is required for such commands.
See the Administrator's Reference for details on server options.

Chapter 28. Managing security 845

 System

Policy Storage Operator Node

Restricted Restricted

Unrestricted Unrestricted

Figure 102. Administrative privilege classes

Table 69 summarizes the privilege classes, and gives examples of how to set
privilege classes.
Table 69. Authority and privilege classes
Privilege Class Capabilities
System Perform any administrative task with the
grant authority rocko classes=system server.
v System-wide responsibilities
v Manage the enterprise
v Manage IBM Tivoli Storage Manager
security
Unrestricted Policy Manage the backup and archive services for
grant authority smith classes=policy nodes assigned to any policy domain.
v Manage nodes
v Manage policy
v Manage schedules
Restricted Policy Same capabilities as unrestricted policy
grant authority jones domains=engpoldom except authority is limited to specific policy
domains.
Unrestricted Storage Manage server storage, but not definition or
grant authority coyote classes=storage deletion of storage pools.
v Manage the database and recovery log
v Manage IBM Tivoli Storage Manager
devices
v Manage IBM Tivoli Storage Manager
storage
Restricted Storage Manage server storage, but limited to specific
grant authority holland stgpools=tape* storage pools.
v Manage IBM Tivoli Storage Manager
devices
v Manage IBM Tivoli Storage Manager
storage

846 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Table 69. Authority and privilege classes (continued)
Privilege Class Capabilities
Operator Control the immediate operation of the
grant authority bill classes=operator server and the availability of storage media.
v Manage the IBM Tivoli Storage Manager
server
v Manage client sessions
v Manage tape operations
Node Access a Web backup-archive client to
grant authority help1 perform backup and restore operations.
classes=node node=labclient

Related concepts:
“Overview of remote access to web backup-archive clients” on page 445
“Managing Tivoli Storage Manager administrator IDs” on page 844

Registering administrator IDs

An administrator can register other administrator IDs and set their authentication
method by using the REGISTER ADMIN command.

About this task

If you want to register an administrator ID and have that ID authenticate with an

LDAP directory server, ensure that all LDAP configuration steps are completed.
When you use AUTHENTICATION=LDAP in a REGISTER ADMIN command, the LDAPURL
setting must be specified in the [Link] file and you must define the LDAPUSER
and LDAPPASSWORD options. If you do not specify an authentication method, the
default (LOCAL) is used.

Examples: Registering an administrator

Example 1: Register an administrator with a user ID of DAVEHIL, the password
birds, and a password expiration period of 120 days.
The authentication method defaults to LOCAL.
register admin davehil birds passexp=120 contact=’backup team’
Example 2: Register an administrator with a user ID of PAULJONES, the
password Ne^Pa$$w0rd, and the authentication method of LDAP.
register admin pauljones Ne^Pa$$w0rd authentication=ldap

The expiration period, length of password, and other characteristics of

LDAP-authenticated passwords can be configured on the LDAP server. But
most IBM Tivoli Storage Manager password-related commands also work
for passwords that authenticate to an LDAP server. For more information
about password policies, see “Setting the policy for an LDAP-authenticated
password” on page 856.

Chapter 28. Managing security 847

 Querying the server for information about administrators
Any administrator can query the server to show administrator information. You
can restrict the query to all administrators authorized with a specific privilege
class.

Procedure

To query the system for a detailed report on administrator ID DAVEHIL, issue the
following example QUERY ADMIN command:
query admin davehil format=detailed

The results are shown in the following detailed report:

Administrator Name: DAVEHIL

Last Access Date/Time: 2002.09.04 17.10.52
Days Since Last Access: <1
Password Set Date/Time: 2002.09.04 17.10.52
Days Since Password Set: 26
Invalid Sign-on Count: 0
Locked?: No
Contact:
System Privilege: Yes
Policy Privilege: **Included with system privilege**
Storage Privilege: **Included with system privilege**
Operator Privilege: **Included with system privilege**
Client Access Privilege: **Included with system privilege**
Client Owner Privilege: **Included with system privilege**
Registration Date/Time: 05/09/2002 [Link]
Registering Administrator: SERVER_CONSOLE
Managing profile:
Password Expiration Period: 90 Day (s)
Authentication: Local

Figure 103. A detailed administrator report

Example: Query administrator IDs that authenticate with an LDAP

directory server
query admin authentication=ldap

Only administrator IDs that authenticate to the LDAP directory server are listed in
the report.

Granting authority to administrators

You can grant authority by issuing the GRANT AUTHORITY command.

Procedure

To grant restricted policy privilege to administrator JONES for the domain

ENGPOLDOM, issue the following example command:
grant authority jones domains=engpoldom

848 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Extending authority for administrators
You can grant and extend authority by issuing the GRANT AUTHORITY command. If
an ID already has some level of authority, granting additional authority adds to
any existing privilege classes; it does not override those classes.

About this task

For example, JONES has restricted policy privilege for policy domain
ENGPOLDOM.

Procedure
1. To extend JONES’ authority to policy domain MKTPOLDOM and add operator
privilege, issue the following example command:
grant authority jones domains=mktpoldom classes=operator
2. As an additional example, assume that three tape storage pools exist:
TAPEPOOL1, TAPEPOOL2, and TAPEPOOL3. To grant restricted storage
privilege for these storage pools to administrator HOLLAND, you can issue the
following command:
grant authority holland stgpools=tape*
3. HOLLAND is restricted to managing storage pools with names that begin with
TAPE, if the storage pools existed when the authority was granted. HOLLAND
is not authorized to manage any storage pools that are defined after authority
has been granted. To add a new storage pool, TAPEPOOL4, to HOLLAND’s
authority, issue the following command:
grant authority holland stgpools=tapepool4

Reducing authority for administrators

You can revoke part of an administrator’s authority by issuing the REVOKE
AUTHORITY command.

About this task

For example, rather than revoking all of the privilege classes for administrator
JONES, you want to revoke only the operator authority and the policy authority to
policy domain MKTPOLDOM.

Procedure

Issue the following command to revoke only the operator authority and the policy
authority to policy domain MKTPOLDOM:
revoke authority jones classes=operator domains=mktpoldom

JONES still has policy privilege to the ENGPOLDOM policy domain.

Chapter 28. Managing security 849

 Reducing privilege classes
You can reduce an administrator’s authority simply by revoking one or more
privilege classes and granting one or more other classes.

Procedure

For example, administrator HOGAN has system authority. To reduce authority for
HOGAN to the operator privilege class, perform the following steps:
1. Revoke the system privilege class by issuing the following command:
revoke authority hogan classes=system
2. Grant operator privilege class by issuing the following command:
grant authority hogan classes=operator

Revoking authority for administrators

You can revoke an administrator's authority by issuing the REVOKE AUTHORITY
command. To revoke all administrative privilege classes, do not specify any
privilege classes, policy domains, or storage pools.

For example, to revoke both the storage and operator privilege classes from
administrator JONES, issue the following command:
revoke authority jones

Updating information about other administrators

An administrator can reset another administrator's password by issuing the UPDATE
ADMIN command. The administrator must have system privileges.

About this task

Tip: If you authenticate a password with an LDAP directory server, the letters and
characters that comprise the password are case-sensitive.

Procedure

For example, administrator DAVEHIL changes their password to ganymede by

issuing the following command:
update admin davehil ganymede

Note: The SERVER_CONSOLE administrator's ID and contact information cannot

be updated.

Renaming an administrator ID
You can rename an administrator ID if it needs to be identified by a new ID. You
can also assign an existing administrator ID to another person by issuing the
RENAME command. You cannot rename an administrator ID to one that exists on the
system.

About this task

For example, if administrator HOLLAND leaves your organization, you can assign
administrative privilege classes to another user by completing the following steps:

Procedure
1. Assign HOLLAND's user ID to WAYNESMITH by issuing the RENAME ADMIN
command:

850 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 rename admin holland waynesmith
By renaming the administrator's ID, you remove HOLLAND as a registered
administrator from the server. In addition, you register WAYNESMITH as an
administrator with the password, contact information, and administrative
privilege classes previously assigned to HOLLAND.
2. Change the password to prevent the previous administrator from accessing the
server by entering:
update admin waynesmith new_password contact="development"

Remember: If you rename an administrator ID that authenticates with an

LDAP directory server, administrators on other servers that share namespace
are not renamed. You must issue a RENAME command for each administrator ID.
When you rename an administrator ID with SYNCLDAPDELETE=YES, the original
name is removed from the LDAP directory server and replaced with the new
name. Administrators, with the original name, on other servers can no longer
authenticate with the LDAP directory server. Those administrators must
register with the LDAP directory server or get renamed to the new name.

Results

Note: The administrator SERVER_CONSOLE cannot be renamed.

Related concepts:
“Securing the server console” on page 843

Removing administrator IDs

You can remove administrator IDs from the server so that they no longer have
access to administrative functions. You must have system privileges to remove an
administrator.

Procedure

To remove registered administrator ID SMITH, issue the following example

command:
remove admin smith

Important:
1. You cannot remove the last system administrator from the system.
2. You cannot remove the administrator SERVER_CONSOLE.

Results

Remember: You can remove an administrator ID that authenticates to the IBM

Tivoli Storage Manager server, and remove the corresponding administrator
namespace on the LDAP directory server. If other Tivoli Storage Manager servers
are not using entries on the LDAP directory server, you do not have to keep
corresponding LDAP entries on the LDAP server.

Example: Remove administrator ID pauljones from the Tivoli Storage

Manager server and the LDAP directory server

Administrator ID pauljones authenticates with an LDAP directory server and the

person who owns that ID is leaving the company. You must find and remove all
occurrences of the administrator ID, pauljones, on the server.
remove admin pauljones syncldapdelete=yes

Chapter 28. Managing security 851

 In this example, the SYNCLDAPDELETE parameter deletes administrator pauljones
from the LDAP server. If administrator pauljones is on other Tivoli Storage
Manager servers, you need to run this same command on each one. Do not delete
entries on the LDAP directory server if they are still needed by another Tivoli
Storage Manager server. Issue SYNCLDAPDELETE=NO if other Tivoli Storage Manager
servers use the administrator namespace.
Related concepts:
“Securing the server console” on page 843

Locking and unlocking administrator IDs from the server

You can issue the LOCK ADMIN command to lock out other administrator IDs and
temporarily prevent them from accessing IBM Tivoli Storage Manager.

About this task

You can also lock or unlock administrator IDs according to the form of
authentication that they use. When you specify AUTHENTICATION=LOCAL in the
command, all administrator IDs that authenticate with the Tivoli Storage Manager
server are affected. When you specify AUTHENTICATION=LDAP in the command, all
administrator IDs that authenticate with an LDAP directory server are affected.

For our example, administrator ID MARYSMITH, takes a leave of absence from

your business.

Procedure
1. Lock MARYSMITH out by entering the following example command:
lock admin marysmith
2. When MARYSMITH returns, any system administrator can unlock the
administrator ID by issuing the following command:
unlock admin marysmith
MARYSMITH can now access the server to complete administrative tasks. You
cannot lock or unlock the SERVER_CONSOLE ID from the server.

Example: Lock all administrator IDs that authenticate locally

lock admin * authentication=local
Related concepts:
“Securing the server console” on page 843

Managing passwords and logon procedures

IBM Tivoli Storage Manager requires the server to identify authorized
administrator IDs and nodes by using a password. You can authenticate
administrator and node passwords with a Lightweight Directory Access Protocol
(LDAP) directory server.

Restriction: Backup-archive clients must be at V6.4 or later to authenticate

passwords with an LDAP directory server. Storage agents authenticating node IDs
with an LDAP directory server must use a secure connection, such as Transport
Layer Security (TLS) or a virtual private network.

852 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 Tivoli Storage
Manager server LDAP server

Tivoli Storage
Manager backup-
archive clients DB2

5. Import the trusted certificate 1. Set up the LDAP directory

from the LDAP directory server. server and configure TLS.

6. Set the LDAPURL option. 2. Create the Base DN.

7. Define the LDAP user ID and
password for node and 3. Grant a specific user ID
password administration. access to the Base DN.

8. Register or update nodes and 4. Copy the trusted certificate to the

administrators. Tivoli Storage Manager server.

Figure 104. Configuring the server to authenticate passwords with an LDAP directory server

The first step in authenticating passwords with an LDAP directory server is to

complete the configuration tasks on the Tivoli Storage Manager server and the
LDAP directory server. The following table shows you which steps are
accomplished on the two servers:
Table 70. The steps that are required to authenticate passwords with an LDAP directory
server, and where the steps are configured
Steps to authenticate passwords with an
LDAP directory server Where to complete the steps
1. Set up an LDAP directory server LDAP server
2. Create the Base DN (distinguished name) LDAP server
3. Grant access to the Base DN to a specific LDAP server
user ID.
4. Copy the trusted certificate from the LDAP server
LDAP directory server to the Tivoli Storage
Manager server
5. Import the trusted certificate from the Tivoli Storage Manager server
LDAP directory server to the Tivoli Storage
Manager server. If you already have a
certificate on the LDAP directory server, you
do not have to generate a new certificate.
You can use the existing certificate to secure
communication between the LDAP directory
server and the Tivoli Storage Manager
server.
6. From the Tivoli Storage Manager server, Tivoli Storage Manager server
update the [Link] file with the LDAPURL
option. See the Administrator's Reference for
more information about the LDAPURL option.

Chapter 28. Managing security 853

 Table 70. The steps that are required to authenticate passwords with an LDAP directory
server, and where the steps are configured (continued)
Steps to authenticate passwords with an
LDAP directory server Where to complete the steps
7. Define the user ID that administers node Tivoli Storage Manager server
and administrator passwords with the LDAP
directory server
8. Define the password for the user ID that Tivoli Storage Manager server
administers node and administrator
passwords. See the Administrator's Reference
for more information about the SET
LDAPPASSWORD command.
9. Update or register node and administrator Tivoli Storage Manager server
IDs to authenticate with an LDAP directory
server. See the Administrator's Reference for
more information about commands.

The LDAP directory server interprets letters differently from the Tivoli Storage
Manager server. The LDAP directory server distinguishes the case that is used,
either uppercase or lowercase. For example, the LDAP directory server can
distinguish between secretword and SeCretwOrd. The Tivoli Storage Manager server
interprets all letters for LOCAL passwords as uppercase.

The following terms are used when describing the LDAP directory server
environment:
Distinguished name (DN)
A unique name in an LDAP directory. The DN consists of the following
information. The information must be ordered in this way.
v The relative distinguished name (RDN)
v The organizational unit (ou)
v The organization (o)
v The country (c)
For example:
uid=jackspratt,ou=marketing,o=[Link],c=us
uid=cbukowski,ou=manufacturing,o=[Link],c=us
uid=abbynormal,ou=sales,o=[Link],c=us

In this example, the value of the RDN on the first line is of an

administrator whose user ID is jackspratt. The organizational unit
(marketing), organization ([Link]), and country (us) comprise the DN.
Bind To validate that a certificate is trusted between an LDAP server and
another server.
Bind DN
The distinguished name that is used to authenticate with the LDAP server.
(This is also the DN of the user ID that is defined in the Tivoli Storage
Manager SET LDAPUSER command.) For example, if our SET LDAPUSER
command is:
set ldapuser "uid=jackspratt,ou=media,cn=security"

then uid=jackspratt,ou=media,cn=security is the bind DN for the LDAP

directory server.

854 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 Bind DN password
The password that is associated with the bind DN.

Configuring a directory server for password authentication

Before you can authenticate passwords with the LDAP server, you must configure
the LDAP server to communicate with the Tivoli Storage Manager server.

Before you begin

You must know the user ID that was specified in the SET LDAPUSER command. For
information about the Tivoli Directory access control lists, go to the Tivoli
Directory server information center ([Link]
tivihelp/v2r1/topic/[Link]/[Link]).

Note: Windows Active Directory users who change passwords when the “Enforce
password history” policy is enabled can authenticate with the previous password
for one hour. For more information, see the Microsoft site (http://
[Link]/?id=906305).

Procedure

Complete the following steps to set up the LDAP directory server so that it can
authenticate passwords:
1. Ensure that you have a directory server installed on the LDAP server. Use one
of the following directory servers:
v IBM Tivoli Directory Server V6.2 or 6.3
v Windows Active Directory version 2003 or 2008

Requirement: If you use Tivoli Directory Server V6.2, you must update Global
Security Kit (GSKit) to V7.0.4.33 or later. For more information, see SSL errors
after upgrading to ITDS 6.3 client ([Link]
[Link]?uid=swg21469388).
2. Create the base distinguished name (Base DN) on the LDAP directory server
for the Tivoli Storage Manager namespace. The Base DN is the part of the
LDAP directory structure from which Tivoli Storage Manager operates,
specified in the LDAPURL option. For example, ou=armonk,cn=tsmdata can be a
Base DN. See your LDAP documentation for how to create a Base DN.
3. Edit the access controls on the LDAP directory server and grant access to the
Base DN to the user ID, which is specified in the SET LDAPUSER command. This
ID cannot be a part of the Base DN. You can grant access to the Base DN to
more than one user ID. However, the security of the LDAP server can be easily
compromised if you have too many user IDs with full permission over the Base
DN.
4. Set up the directory server. See “Configuring SSL or TLS for LDAP directory
servers” on page 839.

What to do next

To verify that the LDAP directory server is properly set up, complete the following
steps on the Tivoli Storage Manager server:
1. Test the forward- and reverse-DNS lookup of the LDAP directory server.
2. Test the network connection with the LDAP directory server.

Chapter 28. Managing security 855

 3. Use an LDAP utility test to connect to the LDAP server and search without
Secure Sockets Layer (SSL)/Transport Layer Security (TLS).
4. Use an LDAP utility test to connect to the LDAP server and search with
SSL/TLS.
Related tasks:
“Configuring Tivoli Directory Server for TLS on the iKeyman GUI” on page 840
“Configuring Tivoli Directory Server for TLS on the command line” on page 841
Related reference:
“Configuring Windows Active Directory for TLS/SSL” on page 842

Setting the policy for an LDAP-authenticated password

Passwords that you authenticate with a Lightweight Directory Access Protocol
(LDAP) server can be managed by the Tivoli Storage Manager server and the
LDAP directory server.

You establish policies for passwords that will be authenticated by each server.

Restriction: You can issue Tivoli Storage Manager server commands to manage
your password policies. If you set a password policy on both the LDAP server and
Tivoli Storage Manager server, the settings might conflict. The result might be that
you are not able to access a node or log on with an administrator ID. For
information on the maximum invalid attempts policy, see the table in “Setting a
limit for invalid password attempts” on page 863.

In addition to setting a policy for case sensitivity, you can configure the
LDAP-authenticated password policy to set the following options:
Password history
The password history is the number of times that you must define a new
password before you can reuse a password.
Minimum age
The minimum age is the length of time before you can change the
password.
Maximum age
The maximum age is the length of time before you must change the
password.
A combination of characters
You can determine the number of special characters, numbers, and
alphabetical characters for your passwords. For example, some products set
up a password policy to enforce the following rules:
v The password cannot contain the user account name or parts of the user
full name that exceed three consecutive characters
v The password must be at least eight characters in length
v The password must contain characters from two of the following four
categories:
– English uppercase characters (A through Z)
– English lowercase characters (a through z)
– Base 10 digits (0 through 9)
– Non-alphabetic characters (for example, !, $, #, %)
Use any of the following characters when you create a password:

856 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 a b c d e f g h i j k l m n o p q r s t u v w x y z
A B C D E F G H I J K L M N O P Q R S T U V W X Y Z
0 1 2 3 4 5 6 7 8 9
~ ! @ # $ % ^ & * _ - + = ` | ( ) { } [ ] : ; < > , . ? /

Important: If you issue a REGISTER NODE or REGISTER ADMIN command,

place special characters within quotation marks. For example:
register admin JohnDoe "pa$$w0rd<new>" authentication=ldap

The LDAP server that you use determines the complexity that you can
have for passwords outside of Tivoli Storage Manager.

Configuring the Tivoli Storage Manager server to authenticate

passwords with an LDAP directory server
You can authenticate passwords with an LDAP directory server only after
configuring the Tivoli Storage Manager server.

Procedure

Complete the following steps on the Tivoli Storage Manager server to authenticate
passwords with an LDAP directory server:
1. Import the key database file from the LDAP directory server. You can use any
method to copy the file from the LDAP directory server to the Tivoli Storage
Manager server.
2. Open the [Link] file and specify the LDAP directory server with the
LDAPURL option. Specify the LDAP directory server URL and the base
distinguished name (DN) on the LDAPURL option. For example:
LDAPURL ldap://[Link]/cn=project_x

The default port is 389. If you want to use a different port number, specify it as
part of the LDAPURL option. For example, to specify a port of 222:
LDAPURL ldap://[Link]/cn=project_x
3. Restart the Tivoli Storage Manager server.
4. Issue the SET LDAPUSER command to define the ID of the user who can
administer Tivoli Storage Manager operations on the LDAP directory server.
This user ID must have full administrative authority over the Base DN and be
able to add, delete, and modify all Base DN entries. For example:
set ldapuser "cn=apastolico,ou=manufacturing,o=dhs,c=us"
See the Administrator’s Reference for more information about the SET LDAPUSER
command.
5. Issue the SET LDAPPASSWORD command to define the password for the user ID
that is defined in the LDAPUSER option. For example:
set ldappassword "boX=T^p$"

Chapter 28. Managing security 857

 Registering nodes and administrator IDs to authenticate
passwords with an LDAP directory server
After you configure the LDAP directory server and the Tivoli Storage Manager
server, you can register nodes and administrator IDs to manage the Tivoli Storage
Manager server.

About this task

If the user ID and password are verified to be correct, communication lines are
opened and the node or administrator ID can run Tivoli Storage Manager
applications.

Procedure

Complete the following step to register either a node or an administrator ID to

authenticate with an LDAP directory server:
1. Issue the REGISTER NODE or the REGISTER ADMIN command from the
command-line interface and specify an authentication method. The
authentication methods that are available are LDAP or LOCAL.
LDAP The administrator ID or node authenticates their password with an
LDAP directory server.
LOCAL
The administrator ID or node authenticates their password with the
Tivoli Storage Manager server.

For example:
register admin admin1 “c0m=p1e#Pa$$w0rd?s” authentication=ldap
register node node1 “n0de^Passw0rd%s” authentication=ldap

After you issue the commands, the passwords for administrator ID admin1 and
the node ID node1 can be authenticated with an LDAP directory server.

Tip: A node and its password or an administrator ID and its password each
occupy one inetOrgPerson object on the LDAP directory server. For information
about inetOrgPerson objects, see Definition of the inetOrgPerson LDAP Object
Class ([Link]
To know which authentication method is in use, issue the QUERY NODE
FORMAT=DETAILED or QUERY ADMIN FORMAT=DETAILED command.
2. Optional: To register all new nodes and administrator IDs with a default
authentication method, issue the SET DEFAULTAUTHENTICATION command. Any
REGISTER NODE or REGISTER ADMIN commands that are issued after you issue the
SET DEFAULTAUTHENTICATION command create nodes and administrators with the
default authentication method. You can set the authentication methods to
LDAP or LOCAL.
For information about the SET DEFAULTAUTHENTICATION command, see the
Administrator's Reference.

858 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Updating nodes and administrator IDs to authenticate
passwords with a directory server
Node and administrator ID passwords can be authenticated with an LDAP
directory server or with Tivoli Storage Manager. You can use either authentication
method.

About this task

When you authenticate nodes and administrator IDs with an LDAP directory
server, you ensure more protection for your passwords. Communication lines
between the LDAP directory server and Tivoli Storage Manager are protected with
Transport Layer Security (TLS).

You can change a password authentication method after you configure the LDAP
directory server and the Tivoli Storage Manager server. However, you cannot
update the authentication method for your own user ID unless you have system
authority. If necessary, another administrator must change the authentication
method.

Procedure

To change a node or administrator password authentication method, complete the

following steps:
1. From the Tivoli Storage Manager server, issue the UPDATE NODE or UPDATE ADMIN
command. You do not have to include a password in the command. If you do
not include a password, you are asked for a new password at the next logon.
You must then enter a password with a combination of uppercase, lowercase,
and numbers or valid characters. For example:
update admin harry authentication=ldap

The following example UPDATE NODE command has a password that is made up
of characters that are supported by the Tivoli Storage Manager server:
update node node1 n0de^87^n0de authentication=ldap

Tip: A shared LDAP server might have a password that is on the LDAP
directory server. In that case, the user is not prompted to enter a new
password.
2. Optional: Issue the QUERY NODE FORMAT=DETAILED or the QUERY ADMIN
FORMAT=DETAILED command to view the results. If you must change the
authentication method for several nodes or administrator IDs, you can use a
wildcard character (*). For example,
update node * authentication=ldap

In the preceding example, the authentication method for all nodes is changed
to “LDAP pending.”

What to do next

All nodes and administrator IDs require new passwords after you run the
UPDATE command. Before the node and administrative IDs are given a password,
they are in the LDAP pending state. The node and administrator IDs are updated
to use LDAP authentication, but you must first give them a password.

Chapter 28. Managing security 859

 Determining which nodes and administrator IDs are
configured to authenticate with an LDAP server
You might have a mixture of nodes or administrator IDs that can and cannot
authenticate with an LDAP directory server. After you update nodes and
administrator IDs to authenticate with an LDAP directory server, you can see
which nodes or administrator IDs are using them.

Tivoli Storage Manager backup-archive clients must be at version 6.4 or later to

authenticate passwords with an LDAP directory server. If all of the backup-archive
clients are not upgraded to version 6.4 or later, you cannot authenticate all nodes
with an LDAP directory server. To determine which nodes or administrator IDs are
configured to authenticate with an LDAP directory server, you can use the QUERY
NODE or QUERY ADMIN command.

Example: Determine which nodes are configured to authenticate

with an LDAP server

Find the nodes that are authenticated with the LDAP directory server:
query node authentication=ldap

Example: Determine which administrator IDs are configured to

authenticate with the Tivoli Storage Manager server

Find the administrator IDs that do not authenticate their passwords with an LDAP
directory server:
query admin authentication=local

Example: Determine the password authentication method for one

node

You can query individual nodes or administrator IDs to determine whether they
authenticate with an LDAP directory server. To determine the password
authentication method for node tivnode_12 issue the following command:
query node tivnode_12 format=detailed

Modifying the default password expiration period for

passwords that are managed by the Tivoli Storage Manager
server
By default, the Tivoli Storage Manager server sets a password expiration of 90
days. The expiration period begins when an administrator ID or client node is first
registered to the server. If a user’s password is not changed within this period, the
user must change the password the next time they access the server.

About this task

Issue the SET PASSEXP command to set the password expiration period for selected
administrator IDs or client nodes. You must specify the administrator ID or node
name with the ADMIN or NODE parameter in the SET PASSEXP command. If you set
the expiration period only for selected users, the expiration period can be 0 - 9999
days. A value of 0 means that user's password never expires.

Restriction: The SET PASSEXP command does not affect administrator IDs and
nodes if their passwords are authenticated with an LDAP directory server.

860 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 The password expiration period for passwords that are authenticated with an
LDAP directory server must be defined on the LDAP server.

Procedure

Issue the following command to set the expiration period of client node
node_tsm12 to 120 days:

set passexp 120 node=node_tsm12

After you explicitly set password expiration for an administrator ID, it is not
modified if you later set a password expiration for all users. You can use the RESET
PASSEXP command to reset the password expiration period to the common
expiration period. Use the QUERY STATUS command to display the common
password expiration period.
Related concepts:
“Setting the policy for an LDAP-authenticated password” on page 856

Scenarios for authenticating passwords

The following example scenarios are for authenticating passwords with an LDAP
directory server.

Scenario 1: Register one node to authenticate its password with

an LDAP directory server

The Tivoli Storage Manager server administrator has a new node that must
authenticate its password with an LDAP directory server. The first action is to
create the “cn=tsmdata” entry and Base DN on the LDAP directory server. The
server administrator can then set up the LDAPURL option that is based on the Base
DN. Here is an example entry for the LDAPURL option:
[Link]
LDAPURL ldaps://[Link]/cn=tsmdata

After you set the LDAPURL option, restart the server. Complete the following steps
to configure the server:
1. Issue the query option ldapurl command to validate that you entered all of
the values correctly.
2. Issue the set ldapuser uid=tsmserver,ou=Users,cn=aixdata command to
configure the LDAPUSER.
3. Issue the SET LDAPPASSWORD adsm4Data command to define the password.
4. For this scenario, the node that must be added is NODE1. Issue the following
command:
register node c0mplexPassw0rd NODE1 authentication=ldap

command.

Scenario 2: Update a node to authenticate with an LDAP

directory server

A single node (UPDNODE1) that currently authenticates with the Tivoli Storage
Manager server is now required to authenticate with an LDAP directory server. For
UPDNODE1, use the AUTHENTICATION parameter in the UPDATE NODE command. For
example:
update node updnode1 newC0mplexPW$ authentication=ldap

Chapter 28. Managing security 861

 If you do not want to specify a password now, you can run the command without
it. At the next logon, either you enter a new password or the password is
generated automatically. If the passwordaccess=generate option is set, a new
password is generated for you.

Scenario 3: Update all nodes to authenticate their password with

an LDAP directory server

If you want to update all your nodes to authenticate with an LDAP directory
server, you can use a wildcard. Issue the following command to have all the nodes
authenticate with an LDAP directory server:
update node * authentication=ldap

Scenario 4: Determine the authentication method of a node

If you have nodes that authenticate with the Tivoli Storage Manager server and
nodes that authenticate with an LDAP directory server, you can determine where
nodes are authenticating. Issue the following command to determine which nodes
authenticate with an LDAP directory server:
query node authentication=ldap

Issue the following command to determine which nodes authenticate with the
Tivoli Storage Manager server:
query node authentication=local

Scenario 5: Locking nodes that authenticate with the Tivoli

Storage Manager server

You can issue a LOCK NODE command to lock all nodes that authenticate with the
Tivoli Storage Manager server. These nodes might be rarely used, and you might
not know by which password authentication method they are supposed to be
managed. When you lock the nodes, the node owners must consult with you. At
that point, you can find out whether they want to use the LDAP directory server
or stay with the Tivoli Storage Manager server. You can issue the LOCK NODE or
UNLOCK NODE commands with a wildcard to lock or unlock all nodes in that group.
To lock all nodes that authenticate with the Tivoli Storage Manager server, issue
the following command:
lock node * authentication=local

Scenario 6: Setting up a default authentication method

After you configure everything, you can design it so that every new node and
administrator authenticate with an LDAP directory server. After you issue the SET
DEFAULTAUTH command, you do not have to designate the authentication method
for any REGISTER NODE or REGISTER ADMIN commands. Issue the following
command to set the default authentication method to LDAP:
set defaultauth=ldap

Any REGISTER NODE or REGISTER ADMIN command that is issued after this SET
DEFAULTAUTH command inherits the authentication method (LDAP). If you want to
register a node that authenticates with the Tivoli Storage Manager server, include
AUTHENTICATION=LOCAL in the REGISTER NODE command.

862 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Setting a limit for invalid password attempts
By default, Tivoli Storage Manager does not check the number of times a user
attempts to logon with an invalid password. You can set a limit on consecutive
invalid password attempts for all client nodes. When the limit is exceeded, the
server locks the node.

About this task

If you authenticate passwords with an LDAP directory server, determine on which

server the invalid attempts are to be configured. Invalid password attempts can be
configured on an LDAP directory server, outside of the Tivoli Storage Manager
server. But the consequence of setting the number of invalid attempts on the LDAP
directory server might pose some problems. For example, when the REGISTER NODE
command is issued, the default behavior is to name the node administrator the
same name as the node. The LDAP server does not recognize the difference
between the node “NODE_Q” and the administrator “NODE_Q”. The node and
the administrator can authenticate to the LDAP server if they have the same
password. If the node and administrator have different passwords, the
authentication fails for either the node or administrator. If the node or the
administrator fail to logon consistently, their IDs are locked. You can avoid this
situation by issuing the REGISTER NODE command with USERID=userid or
USERID=NONE.

Consider the maximum invalid attempts policy:

Table 71. The consequences of setting the maximum invalid attempts on both the Tivoli
Storage Manager server and the LDAP server for a node and admin that share a name
Authentication Tivoli Storage Manager
method password setting LDAP directory server setting
(A maximum of 3 invalid (A maximum of 3 invalid
attempts) attempts)
With an LDAP The password provided during The password provided during
directory server authentication is attempted for authentication is attempted for
(LDAP) both the node and administrator both the node and administrator
that share a name. The policy does that share a name. If the node and
not affect the node or administrator have different
administrators that are passwords, the authentication on
authenticating to the Tivoli Storage the LDAP server fails for either the
Manager server, even when their node or administrator. If the
passwords are different. The Tivoli correct node password is provided
Storage Manager server recognizes on 3 consecutive attempts, the
that the node and administrator administrator authentication fails
share a name and authenticates on those same 3 attempts. The
both simultaneously. If the administrator ID gets locked out
passwords are not the same and after the third attempt.
the correct node password is
provided on 3 attempts, the
administrator's incorrect password
is ignored. No invalid login
attempts are recorded.
With the Tivoli Has 3 attempts to get the Not applicable.
Storage Manager password right
server (LOCAL)

On the Tivoli Storage Manager server, issue the SET INVALIDPWLIMIT command to
limit the invalid password attempts for the Tivoli Storage Manager namespace.

Chapter 28. Managing security 863

 Procedure

To set a system-wide limit of three consecutive invalid password attempts, issue

the following example command:
set invalidpwlimit 3

The default value at installation is 0. A value of 0 means that invalid password

attempts are not checked. You can set the value from 0 to 9999 attempts.

If you initially set a limit of 4 and then lower the limit, some clients might fail
verification during the next logon attempt.

After a client node is locked, only an administrator with storage authority can
unlock the node.

An administrator can also force a client to change their password on the next
logon by specifying the FORCEPWRESET=YES parameter on the UPDATE NODE or UPDATE
ADMIN command. For more information, see the Administrator's Reference.
Related tasks:
“Locking and unlocking client nodes” on page 440
“Locking and unlocking administrator IDs from the server” on page 852

Setting a minimum length for a password

By default, IBM Tivoli Storage Manager does not check the length of a password.
The administrator can specify a minimum password length that is required for
Tivoli Storage Manager passwords.

About this task

This feature affects all node and administrator passwords, whether the password
authenticates with the Tivoli Storage Manager server or the LDAP directory server.

You can configure the parameters of your LDAP-authenticated passwords on the

LDAP directory server. The settings that you configure might conflict with the
Tivoli Storage Manager server settings.

Procedure

To set the minimum password length to eight characters, issue the following
example command:
set minpwlength 8

The default value at installation is 0. A value of 0 means that the password length
is not checked. You can set the length value from 0 to 64.

864 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Disabling the default password authentication
By default, the server automatically sets password authentication to “on.” With
password authentication set to on, all users must enter a password when accessing
the server.

About this task

You can only disable password authentication for passwords that authenticate with
the Tivoli Storage Manager server (LOCAL).

Procedure

To allow administrators and client nodes to access the Tivoli Storage Manager
server without entering a password, issue the following command:
set authentication off

Remember: Setting password authentication to “off” reduces data security.

Enabling unified logon with backup-archive clients

The IBM Tivoli Storage Manager unified logon feature allows the Tivoli Storage
Manager server to use the Windows user account database when authenticating a
backup-archive client logon.

With this feature, a user can log on to a Windows computer and access the
backup-archive client without entering another password. When unified logon is
enabled, the server continues to use its normal authentication methods for
protocols other than Named Pipes.

The procedure described here assumes that the Tivoli Storage Manager server and
all the Tivoli Storage Manager client computers are in the same Windows domain.
A Windows domain is a way of allowing the Windows Domain Controller to
manage the user accounts for all members of the domain. The Tivoli Storage
Manager unified logon procedure takes advantage of the convenience of allowing
the Windows domain to manage user accounts.

Tip: The Tivoli Storage Manager server can run successfully on the Windows
server or workstation operating system and does not have to reside on the
Windows Domain Controller computer.

To enable unified logon, you must have the following system requirements:
v The backup-archive client must be installed on a supported Windows operating
system.
v The Tivoli Storage Manager server must enable the Named Pipes protocol.
v The backup-archive client must use the Named Pipes communications method.

The unified logon feature applies only to node and administrator passwords that
authenticate to the Tivoli Storage Manager server. Passwords that are stored and
authenticate to an LDAP directory server are not affected by this feature.
Authentication must be LOCAL (Tivoli Storage Manager server) for unified logon to
apply.

Chapter 28. Managing security 865

 Enabling unified logon
Unified logon allows the IBM Tivoli Storage Manager server to use the Windows
user account database. With this feature, a user can log on to a Windows computer
and access the backup-archive client without entering another password.

About this task

Unified logon affects node and administrator passwords that authenticate to the
Tivoli Storage Manager server. Passwords that authenticate to an LDAP directory
server are not affected.

Procedure

Perform the following steps to enable unified logon:

1. From the Windows Domain Controller, perform the following steps to
authorize Tivoli Storage Manager users by creating a Windows global group
and adding users:
a. Open the User Manager for Domains from Administrative Tools.
b. Click User > New Global Group.
c. For Group Name enter, for example, ADSMSERVER.
d. Add users to the ADSMSERVER group by selecting Add.
e. Click OK to add the group.
f. Reboot the server workstation to enable the groups.
2. To stop the server, issue the following command from the Tivoli Storage
Manager command prompt:
halt
3. Perform the following steps to configure Tivoli Storage Manager server service
to use the administrator account:
a. Right-click on the server you want to work with, and select Properties. On
Windows: Start > Control Panel > Administrative Tools > Services.
b. From the Startup Type section, select Automatic.
c. From the Log On As section, select This Account and change LocalSystem
to DOMAIN\Administrator.
d. Enter the password of the Windows Domain administrator account.
e. Click OK to continue.
4. The preceding procedure is the same as entering the following lines in the
server options file ([Link]):
adsmgroupname tsmserver
commmethod namedpipe
namedpipename \\.\pipe\server1
securepipes yes
a. Click the Named Pipes tab and perform the following steps:
5. Perform the following steps to restart the Tivoli Storage Manager server:
a. From the left pane, expand the Tivoli Storage Manager server.
b. Expand Reports.
c. Click Service Information.
d. From the right pane, select the server and click Start. The status for the
server changes to Running.
6. Ensure that the backup-archive clients that you added to the Tivoli Storage
Manager server group are registered Tivoli Storage Manager client nodes.

866 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 Important: The node password that is assigned during node registration is
ignored by the authentication process when unified logon is enabled. Instead,
the Windows domain password is used.
7. From each client workstation, all users must ensure that they are members of
either the Administrators group or the Backup Operators group on their
workstations.
8. Reboot the client workstation to enable the group.
9. Edit the client options file ([Link]) to set up the backup-archive client to use
unified logon. Ensure that the following options are active (the asterisk is
deleted). For example:
commmethod namedpipe
namedpipename \\server_name\pipe\server1
nodename username
passwordaccess generate

Tip:
a. In the example, server_name is the NetBIOS name of the computer where
the Tivoli Storage Manager server is running.
b. In the example, nodename can be substituted with the name of the
workstation where the Tivoli Storage Manager server is installed.
c. The username must be the same as the Windows account name that the
user is logged in as.
10. To verify that unified logon is enabled, start the backup-archive client. You can
also perform a backup and restore.
Related tasks:
“Registering nodes with the server” on page 420

Chapter 28. Managing security 867

868 IBM Tivoli Storage Manager for Windows: Administrator's Guide
Chapter 29. Protecting and recovering the server
infrastructure and client data
The Tivoli Storage Manager infrastructure consists of the database and the setup
files that are required to recover the database and client data. The setup files
include, for example, the active log and the archive log. Client data includes data
that is backed up, archived, and migrated to primary storage pools.

About this task

Clustering is a high-availability solution that minimizes or eliminates many

potential sources of downtime.

Database backups, infrastructure setup files, and copies of client data can be stored
offsite, as shown in Figure 105.

On-site Off-site

Server Server storage

primary storage pools
Copy storage
pools
HSM

Backup

Archive Database
Database and
backups
recovery log
Disk storage pool
Server

Infrastructure
setup files

Tape storage pool

Figure 105. Recovery from a disaster

DRM: The disaster recovery manager (DRM ) can automate some disaster recovery tasks. A
note like this one identifies those tasks.

Related tasks:
“Storage pool hierarchies” on page 260
Related information:
Configuring clustered environments

© Copyright IBM Corp. 1993, 2013 869

Protecting the database and infrastructure setup files
To restore a damaged or lost database you must have a database backup. You must
also have copies of the files that are required to recover the database and client
data. Database backup media and setup files can be stored offsite for protection.

Restriction: Database mirroring is not possible with Tivoli Storage Manager.

However, you can mirror the database by using operating system or file system
capabilities. You can also use device redundancy such as RAID capabilities in the
storage that is used for the server database.

DRM: To store database backup media and setup files offsite, you can use disaster recovery
manager.

Related tasks:
Chapter 32, “Disaster recovery manager,” on page 1005

Backing up the server database

When the Tivoli Storage Manager server is first started, a full backup starts
automatically. After the initial full backup, the database is backed up automatically
by the database manager. To ensure that the database can be restored to an
acceptable point-in-time, you can schedule daily backups or backup the database
manually.

About this task

Automatic backups by the database manager are based on the following values
that are set by Tivoli Storage Manager:
v The active log space that was used since the last backup, which triggers a full
database backup
v The active log utilization ratio, which triggers an incremental database backup

You can back up the database to tape, FILE, or to remote virtual volumes.

To set up regular database backups, complete the following tasks:

v “Step 1: Defining device classes for database backups” on page 871
v “Step 2: Specifying virtual address space for database manager processes” on
page 871
v “Configuring concurrent multistreaming” on page 872
v “Step 4: Running database backups” on page 875
Related concepts:
“Active log” on page 647
“Database: Overview” on page 644
Related tasks:
“Using virtual volumes to store data on another server” on page 732
“Increasing the size of the active log” on page 678

870 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Step 1: Defining device classes for database backups
You can use existing device classes for database backups or you can define new
ones. You can also specify device classes for incremental backups that are different
from the device classes for full backups. For example, you might want to write full
backups to tape and incremental backups to disk.

About this task

Reserve the device class that you want to use for backups so that the server does
not attempt to back up the database if a device is not available. If a database
backup shares a device class with a lower priority operation, such as reclamation,
and all the devices are in use, the lower priority operation is automatically
canceled. The canceled operation frees a device for the database backup.

Restriction: Tivoli Storage Manager does not support database backup (loading
and unloading) to a Centera device.

Procedure

To specify the device class to be used for database backups, issue the SET
DBRECOVERY command. For example, to specify a device class named DBBACK,
issue the following command:
set dbrecovery dbback

Tips:
v When you issue the SET DBRECOVERY command, you can also specify the number
of number of concurrent data streams to use for the backup. Use the NUMSTREAMS
parameter.
v To change the device class, reissue the SET DBRECOVERY command.
v If you issue the BACKUP DB command with the TYPE=FULL parameter, and the
device class is not the one that is specified in the SET DBRECOVERY command, a
warning message is issued. However, the backup operation continues and is not
affected.
v Device class definitions are saved in the device configuration files.
Related concepts:
“Configuring concurrent multistreaming” on page 872
Related tasks:
“Protecting the device configuration file” on page 879

Step 2: Specifying virtual address space for database manager

processes
Specify the percentage of the virtual address space that is dedicated to database
manager processes.

About this task

By default, the percentage of the virtual address space that is dedicated to all
database manager processes is set to 70 - 80 percent of system random-access
memory.

Procedure

To change this setting, specify the DBMEMPERCENT server option. Ensure that the
value that you specify provides adequate memory for applications other than the

Chapter 29. Protecting and recovering the server infrastructure and client data 871
 Tivoli Storage Manager are running on the system.

Configuring concurrent multistreaming

Multiple, concurrent data streams reduce the time required to back up or restore
the database. You can specify the number of data streams that the IBM Tivoli
Storage Manager server uses for backup and restore operations.

For example, if you assign four drives to database backup processing, Tivoli
Storage Manager attempts to write to all four drives concurrently. For restore
operations, the server uses the information that is in the volume history file to
determine the number of data streams that were used during the backup
operation. The server attempts to use the same number of data streams during the
restore operation. For example, if the backup operation used four data streams, the
server attempts the restore operation using four data streams.

The following table summarizes server actions for different drive-availability

scenarios. An available drive is a drive that is online and that is ready for use by the
server.

Operation If the number of available If the number of available If the number of available drives is
drives exceeds the specified drives equals the specified less than the specified number of
number of streams, the server number of streams, the streams, the server uses
uses server uses
Backup The number of drives that is The number of drives that is All available drives.
equal to the specified number equal to the specified
of streams. number of streams.
Restore The number of drives that is The number of drives that is All available drives. At least one drive
equal to the number of equal to the number of is required for restore processing
streams that were used in the streams that were used in the
backup operation. A restore backup operation.
process never uses more
drives than the number of
streams that were used to
back up the database.

Suppose that you specify four data streams for database backup operations. To
indicate the maximum number of volumes that can be simultaneously mounted,
you specify 4 as the value of the MOUNTLIMIT parameter in the device class
definition. If only three drives are available at the time of the backup operation,
the operation runs using three drives. A message is issued that indicates that fewer
drives are being used for the backup operation than the number requested. If all
four drives for the device class are online, but one drive is in use by another
operation, the backup operation has a higher priority and preempts use of the
drive. If you specify four data streams, but the value of the MOUNTLIMIT parameter
is 2, only two streams are used.

Important: Although multiple, concurrent data streams can reduce the time that is
required for a backup operation, the amount of time that you can save depends on
the size of the database. In general, the benefit of using multiple, concurrent data
streams for database backup and restore operations is limited if the database is less
than 100 GB.

Another potential disadvantage is that more volumes are required for multistream
processing than for single-stream processing. For example, if the backup of an 850
GB database requires a single linear tape open (LTO) volume, switching to four
data streams requires four volumes. Furthermore, those volumes might be partially

872 IBM Tivoli Storage Manager for Windows: Administrator's Guide

filled, especially if you use high-capacity volumes and device compression. For
example, if the backup of an 850 GB database consumes only 30% of a tape
volume after compression, then four-stream processing will result in even larger
amounts of wasted space.

Planning for concurrent multistreaming during database backup and restore

operations:

The decision to use multiple, concurrent data streams for database backup and
restore operations depends on the size of the database, the cost of media, and
performance impacts.

Before you begin

When deciding whether to use data streaming, consider the following issues to
determine whether the benefits of concurrent data streaming are sufficient. If the
disadvantages of multiple, concurrent data streaming exceed the benefits, continue
to use single-stream processing.
v What is the size of your database? In general, the amount of time that you save
by using multiple, concurrent data streams decreases as the size of the database
decreases because of the extra time caused by additional tape mounts. If your
database is less than 100 GB, the amount of time that you save might be
relatively small.
In many environments with databases larger than 100 GB, two database-backup
streams can provide superior performance. However, depending on your
environment, additional streams might not provide enough I/O throughput
relative to the size of your database, the devices that you use, and the I/O
capability of your environment. Consider using three or four database-backup
streams only for environments in which the following conditions apply:
– The Tivoli Storage Manager database is located on very high-performing disk
subsystems.
– The database is spread across several different RAID arrays that use multiple
database directories.
v How many drives are available for the device class to be used for database
backup?
v Will server operations other than database backup operations compete for
drives?
v If drives are preempted by a database backup operation, what will be the effect
on server operations?
v What is the cost of the tape volumes that you use for database backup
operations? For example, suppose that the backup of an 850 GB database
requires a single high-capacity LTO volume. If you specify four streams, the
same backup operation requires four volumes.

Chapter 29. Protecting and recovering the server infrastructure and client data 873
 Specifying multiple data streams for database backup operations:

You can specify multiple data streams for automatic or manual database-backup
operations. For database restore operations, the server attempts to use the same
number of data streams that you specified for the backup operation.

Procedure

To specify multiple data streams, complete one of the following steps:

v For automatic database-backup operations, issue the SET DBRECOVERY command
and specify a value for the NUMSTREAMS parameter. For example, if you have a
device class DBBACK, issue the following command to specify two data streams:
set dbrecovery dbback numstreams=2
v For manual database-backup operations, issue the BACKUP DB command and
specify a value for the NUMSTREAMS parameter. The value of the NUMSTREAMS
parameter that you specify with the BACKUP DB command overrides the value for
the NUMSTREAMS parameter that you specify with the SET DBRECOVERY command.
For example, if you have a device class DBBACK, issue the following command
to specify three data streams:
backup db dbback numstreams=3

Example

Tips:
v To change the number of data streams for automatic database backup
operations, reissue the SET DBRECOVERY command and specify a different value
for the NUMSTREAMS parameter. For example, reissue the SET DBRECOVERY
command if you add additional drives to the target library or if drives are not
available because of maintenance or device failure. The new value specified by
the NUMSTREAMS parameter is used for the next backup operation.
v To display the number of data streams that are to be used for a database backup
operation, issue the QUERY DB command.
v During a database backup operation, the number of sessions that is displayed in
the output of the QUERY SESSION command or the SELECT command is equal to or
less than the number of specified data streams. For example, if you specified
four data streams, but only three drives are online, 3 sessions are displayed in
the output. If you issue the QUERY DRIVE command, the number of drives in use
is also 3.
v If you reduce the number of data streams after a database backup operation, this
information will not be available to the server when the database is restored. To
specify fewer data streams for the restore operation, take one or both of the
following actions in the device configuration file:
– Reduce the number of online and usable drive definitions by removing
DEFINE DRIVE commands.
– Update the value of the MOUNTLIMIT parameter of the DEFINE DEVCLASS
command.
– During database backup operations, stop other Tivoli Storage Manager
database activities. Other database activities compete for database I/O and
affect throughput during database backup operations that use multiple
streams.

874 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Step 4: Running database backups
Database backups can be full, incremental, or snapshot. An incremental backup
includes all the changes since the last full backup. You can schedule the backups to
occur automatically, or you can back up the database manually. A snapshot
database backup is a full database backup that does not interrupt the current full
and incremental backup series.

About this task

Ensure that you can recover the database to its most current state or to a specific
point-in-time by making both full and incremental database backups:
v To restore the database to its most current state, you need the last full backup,
the last incremental backup after that full backup, and the active and archive log
files.

Restriction: You cannot use snapshot database backups to restore a database to

its most current state.
v To restore a database to a point-in-time, you need the last full backup before the
point-in-time and the last incremental backup after the last full backup before
the point-in-time.
You can also use snapshot database backups to restore a database to a
point-in-time.

Tivoli Storage Manager can make full and incremental database backups to tape
while the server is running and available to clients. However, when deciding what
backups to do and when to do them, consider the following properties of backups:
v Full backups take longer than incremental backups.
v Full backups have shorter recovery times than incremental backups because you
must load only one set of volumes to restore the entire database.
v Full backups are required for the first backup and after extending the database
size.
v Only full backups prune archive log space in the archive log directory. If the
available active and archive log space gets low, full database backups occur
automatically. To help prevent space problems, schedule regular full backups
frequently.

Tips: To protect the database, use the following guidelines:

v Store the database backup media an off-site location for protection.
v Back up the database immediately after you back up the storage pools.
v Turn off migration and reclamation while you back up the database.
v Do not issue the MOVE DATA command while you back up the database.

Backing up the database manually:

To back up the database manually, issue the BACKUP DB command from an

administrative client or from the server console.

About this task

For a full database backup, specify TYPE=FULL. For an incremental database backup,
specify TYPE=INCREMENTAL. For example, to run a full database backup using a
device class LTOTAPE, three volumes, and three concurrent data streams, issue the
following command:

Chapter 29. Protecting and recovering the server infrastructure and client data 875
 backup db devclass=ltotape type=full volumenames=vol1,vol2,vol3
numstreams=3

Scheduling database backups:

Database backups require devices, media, and time. Consider scheduling backups
at specific times of the day and after major storage operations.

About this task

Consider scheduling database backups after the following types of operations:

v Regular client backup or archive operations
v Storage pool migration and reclamation
v Storage pool backups
v MOVE DATA or DELETE VOLUME command operations
For example, you might back up your storage pools daily and then immediately
back up the database.

Procedure

To schedule database backups, use the DEFINE SCHEDULE command. For a full
database backup, specify TYPE=FULL. For an incremental database backup, specify
TYPE=INCREMENTAL. For example, to set up a schedule to run a full backup to device
class FILE every day at 1:00 a.m., enter the following command:
define schedule daily_backup type=administrative
cmd="backup db deviceclass=file type=full" starttime=01:00

Running snapshot database backups:

A snapshot database backup is a full database backup that does not interrupt the
full and incremental backup series. Consider using snapshot database backups in
addition to full and incremental backups.

Procedure

To make a snapshot database backup, issue the BACKUP DB command. For example,
to make a snapshot database backup to the TAPECLASS device class, enter the
following command:
backup db type=dbsnapshot devclass=tapeclass

New volume history entries are created for the snapshot database volumes.

Restriction: To prevent the accidental loss of what might only way to recover the
server, you cannot delete the most current snapshot database using the DELETE
VOLHISTORY command.
Related concepts:
“Volume history file and volume reuse” on page 72
Related tasks:
“Protecting the volume history file” on page 878

876 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Protecting infrastructure setup files
Infrastructure setup files are prerequisites for recovering the Tivoli Storage
Manager database and client data. In most cases, these files cannot be recreated, so
you must ensure that copies are up-to-date and easily accessible.

Protecting the active, archive, and archive failover logs

The active log, which records transactions that are in progress on the server, is
necessary to restore the database to its most current state. The archive log contains
copies of closed log files that were in the active log. The archive log is not needed
for normal processing, but it is typically needed for recovery of the database.

About this task

For protection against database and log media failures, place the active log and the
archive log in different file systems. In addition, mirror both logs. Mirroring
simultaneously writes data to two independent disks. For example, suppose that a
sudden power outage causes a partial page write. The active log is corrupted and
is not readable. Without mirroring, recovery operations cannot complete when the
server is restarted. However, if the active log is mirrored and a partial write is
detected, the log mirror can be used to construct valid images of the missing data.

To protect the active log, the archive log, and the archive failover log, take the
following steps:

Procedure
v To specify the active log mirror, use the MIRRORLOGDIRECTORY parameter on the
DSMSERV FORMAT command. Mirror the active log in a file system that exists on a
different disk drive than the primary active log.

Tip: To specify the active log mirror, you can also use the MIRRORLOGDIRECTORY
server option.
v You cannot mirror the archive log through Tivoli Storage Manager. However,
you can mirror the archive log by using operating system or hardware
mirroring, such as that provided by RAID 5. Mirror the archive log directory
using file system or disk-drive subsystem-level facilities. Ensure that the
mirrored log is in a file system on a different physical hardware.
v The archive failover log, also called a secondary archive log, is the directory that
the server uses to store archive log files if the archive log directory is full.
Specifying an archive failover log directory can prevent problems that occur if
the archive log runs out of space. Use of an archive failover log is optional, but
it provides an extra level of protection.
Store the archive log and the archive failover log on different physical drives. To
mirror the archive failover log, use operating system or hardware mirroring.

What to do next

Tips:
v Consider mirroring the active log and the archive log if retention protection is
enabled. If a database restore is needed, you can restore the database to the
current point in time with no data loss.
v You can dynamically start or stop mirroring while Tivoli Storage Manager is
running.
v Despite its benefits, mirroring does not protect against a disaster or a hardware
failure that affects multiple drives or causes the loss of the entire system. In

Chapter 29. Protecting and recovering the server infrastructure and client data 877
 addition, mirroring doubles the amount of disk space that is required for logs.
Mirroring also results in decreased performance.
Related concepts:
“Active log” on page 647
“Archive log” on page 648
“Archive failover log” on page 649

Protecting the volume history file

To restore the database, the server needs the information that is in a volume
history file. You can specify duplicate volume history files. When the server
updates volume information in the database, it also updates each file.

About this task

The following volume history is stored in the Tivoli Storage Manager database and
updated in the volume history files:
v Sequential-access storage-pool volumes that were added, reused through
reclamation or move data operations, or deleted during delete volume or
reclamation operations
v Full and incremental database-backup volumes
v Export volumes for administrator, node, policy, and server data
v Snapshot database-backup volumes
v Backup set volumes

Procedure

To specify the file path and name for a volume history file, use the VOLUMEHISTORY
server option. To specify more than one path and name, use multiple
VOLUMEHISTORY entries. Tivoli Storage Manager stores duplicate volume histories in
all the files that are specified with VOLUMEHISTORY options. To find the required
volume-history information during a database restore operation, the server tries to
open volume history files in the order in which the VOLUMEHISTORY entries occur in
the server options file. If the server cannot read a file, the server tries to open the
next volume history file.
Ensure that volume history is protected by taking one or more of the following
steps:
v Store at least one copy of the volume history file offsite or on a disk separate
from the database.
v Store a printout of the file offsite.
v Store a copy of the file offsite with your database backups and device
configuration file.
v Store a remote copy of the file, for example, on an NFS-mounted file system.

Tip: To manually update the volume history file, you can use the BACKUP
VOLHISTORY command. Ensure that updates are complete by following these
guidelines:
v If you must halt the server, wait a few minutes after issuing the BACKUP
VOLHISTORY command.
v Specify multiple VOLUMEHISTORY options in the server options file.
v Review the volume history files to verify that the files were updated.

878 IBM Tivoli Storage Manager for Windows: Administrator's Guide

What to do next
DRM: DRM saves a copy of the volume history file in its disaster recovery plan file.

Related tasks:
“Deleting information about volume history” on page 615

Protecting the device configuration file

The device configuration file contains information that is required to read backup
data and restore the database. You can specify duplicate device configuration files.
When the server updates device configuration information in the database, it also
updates each file. A device configuration file cannot be recreated.

About this task

The following device configuration information is stored in the Tivoli Storage

Manager database and updated in the device configuration files:
v Devices class definitions
v Library definitions
v Drive definitions
v Path definitions
v Server definitions
v The database manager backup node ID
The device information must match the devices configured on the system where
the restore operation can be performed. You might have to edit those commands in
an existing file so that they match.

Procedure

To specify the file path and name for a device configuration file, use the DEVCONFIG
server option. To specify more than one path and name, use multiple DEVCONIG
entries. Tivoli Storage Manager stores duplicate device configuration information in
all the files that are specified with DEVCONFIG options.
To find the required device-configuration information during a database restore
operation, the server tries to open device configuration files in the order in which
the DEVCONFIG entries occur in the server options file. If the server cannot read a
file, the server tries to open the next device configuration file.
To ensure the availability of device configuration information, take one or more of
the following steps:
v Store at least one copy of the device configuration file offsite or on a disk
separate from the database.
v Store a printout of the file offsite.
v Store a copy of the file offsite with your database backups and volume history
file.
v Store a remote copy of the file, for example, on an NFS-mounted file system.

Tips:
v To manually update the device configuration file, use the BACKUP DEVCONFIG
command. Ensure that updates are complete by following these guidelines:
– If you must halt the server, wait a few minutes after issuing the BACKUP
DEVCONIG command.
– Specify multiple DEVCONFIG options in the server options file.

Chapter 29. Protecting and recovering the server infrastructure and client data 879
 – Review the device configuration files to verify that the files were updated.
– If you are using automated tape libraries, volume location information is
saved in the device configuration file. The file is updated whenever CHECKIN
LIBVOLUME, CHECKOUT LIBVOLUME, and AUDIT LIBRARY commands are issued,
and the information is saved as comments (/*....*/). This information is used
during restore or load operations to locate a volume in an automated library.
If a disaster occurs, you might have to restore Tivoli Storage Manager with devices
that are not included in the device configuration file.

What to do next
DRM: DRM automatically saves a copy of the device configuration file in its disaster
recovery plan file.

Related tasks:
“Updating the device configuration file” on page 906

Protecting the server options file

To restore the database, you need a copy of the server options file. The server
options file includes the file paths of the active log, the archive log, the active log
mirror, and the archive failover log. This information is required to restore the
database.

About this task

To ensure the availability of server options file, take one or more of the following
steps:
v Store at least one copy of the server options file offsite or on a disk separate
from the database.
v Store a printout of the file offsite.
v Store a copy of the file offsite with your database backups and device
configuration file.
v Store a remote copy of the file, for example, on an NFS-mounted file system.

DRM: DRM automatically saves a copy of the server options file in its disaster recovery
plan file.

Protecting information about the database and recovery logs

To restore the database, you need detailed information about the database and
recovery log. The recovery log includes the active log, the active log mirror, the
archive log, and the archive failover log. The recovery log contains records of
changes to the database.

About this task

You can determine the following information from the recovery log:
v The directory where the recovery log is located
v The amount of disk space required
If you lose the recovery log, you lose the changes that were made since the last
database backup.

880 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Procedure

Complete both of the following steps:

v To obtain detailed information about the database, issue the following
commands:
query db format=detailed
query dbspace
v To obtain information about the recovery log, issue the following command:
query log format=detailed

What to do next
DRM: DRM helps you save database and recovery log information.

Protecting the Secure Sockets Layer digital certificate file

As part of the process of setting up IBM Tivoli Storage Manager to use Secure
Sockets Layer (SSL) for client-server authentication, a digital certificate file,
[Link], is created.

The [Link] file includes the server's public key, which allows the client to
encrypt data. The digital certificate file cannot be stored in the server database
because the Global Security Kit (GSKit) requires a separate file in a certain format.
The [Link] file is generated by the V6.3 server for distribution to the V6.3
clients.

Keep backup copies of the [Link] and [Link] file in a secure location. If
both of the original files and any copies are lost or corrupted, you can generate a
new certificate file.

Attention: If client data object encryption is in use and the encryption key is not
available, data cannot be restored or retrieved under any circumstance. When
using ENABLECLIENTENCRYPTKEY for encryption, the encryption key is stored
on the server database. This means that for objects using this method, the server
database must exist and have the proper values for the objects for a proper restore
operation. Ensure that you back up the server database frequently to prevent data
loss.

For more information about encryption keys, see IBM Tivoli Storage Manager Using
the Application Programming Interface.
Related tasks:
“Troubleshooting the certificate key database” on page 837

Protecting the disaster recovery plan

The disaster recovery plan file contains the information required to recover a Tivoli
Storage Manager server to the point in time represented by the last database
backup operation that is completed before the plan is created.

You can use server-to-server communications to store copies of the recovery plan
on a remote target server, in addition to traditional disk-based files. Storing
recovery plan files on a target server provides the following advantages:
v A central repository for recovery plan files
v Automatic expiration of plan files
v Query capabilities for displaying information about plan files and their contents

Chapter 29. Protecting and recovering the server infrastructure and client data 881
 v Fast retrieval of a recovery plan file if a disaster occurs

You can also store the recovery plan locally, on CD, or in print.

DRM: DRM can query the server and generate a detailed recovery plan for your
installation.

Related tasks:
“Storing the disaster recovery plan locally” on page 1019
“Storing the disaster recovery plan on a target server” on page 1019
Related reference:
“The disaster recovery plan file” on page 1048

Protecting client data

If storage pools and volumes are not backed up, client data can be permanently
lost.

Protecting the data that is in primary storage pools

To protect client data, back up your primary storage pools to copy storage pools.
For fast client restores, copy active client backup data that is in primary storage
pools to active-data pools.

About this task

A typical Tivoli Storage Manager configuration includes a primary disk pool and
primary tape pool for data backup. Copy storage pools contain active and inactive
versions of data that is backed up from primary storage pools. Figure 106 shows a
configuration with an onsite FILE-type active-data pool and an offsite copy storage
pool.

Server Storage On-site storage

HSM
Backup Active backup
data only
Archive
Disk Storage
Pool (FILE)

Disk Storage Pools

Archive, backup Off-site Storage
(active and
inactive data)
and HSM data

Tape Storage Pool

Tape Storage Pool

Figure 106. Copy storage pools and active-data pools

Related concepts:
“Active-data pools” on page 241

882 IBM Tivoli Storage Manager for Windows: Administrator's Guide

“Copy storage pools” on page 241
“Primary storage pools” on page 240
Related tasks:
“Storage pool hierarchies” on page 260

Backing up primary storage pools

To protect client data, back up your primary storage pools to copy storage pools.
Copy active client backup data in primary storage pools to active-data pools for
fast client restores.

About this task

Tip: Backing up storage pools requires an additional 200 bytes of space in the
database for each file copy. As more files are added to the copy storage pools and
active-data pools, reevaluate your database size requirements.

Each of the commands in the following examples uses four parallel processes
(MAXPROCESS=4) to perform an incremental backup of the primary storage pool
to the copy storage pool or a copy to the active-data pool. Set the MAXPROCESS
parameter in the BACKUP STGPOOL command to the number of mount points or
drives that can be dedicated to this operation.

Procedure
v To back up data in a primary storage pool to a copy storage pool, use the BACKUP
STGPOOL command. For example, to back up a primary storage pool named
ARCHIVEPOOL to a copy storage pool named DISASTER-RECOVERY, issue the
following command:
backup stgpool archivepool disaster-recovery maxprocess=4

The only files backed up to the DISASTER-RECOVERY pool are files for which a
copy does not exist in the copy storage pool. The data format of the copy
storage pool and the primary storage pool can be NATIVE, NONBLOCK, or the
NDMP formats NETAPPDUMP, CELERRADUMP, or NDMPDUMP. The server
copies data from the primary storage pool only to a copy storage pool that has
the same format.

Restriction: A BACKUP STGPOOL command does not back up a shred storage pool
to a copy storage pool unless you explicitly permit it by setting the
SHREDTONOSHRED parameter to YES. If this value is not specified, the server issues
an error message and does not allow the backup. If this value is specified, the
server does not issue a warning when the BACKUP STGPOOL command for the
shred pool is run.

Tip: To further minimize the potential loss of data, you can mark the backup
volumes in the copy storage pool as OFFSITE and move them to an offsite
location. In this way, the backup volumes are preserved and are not reused or
mounted until they are brought on-site. Ensure that you mark the volumes as
OFFSITE before you back up the database. To avoid marking volumes as offsite
or physically move volumes:
– Specify a device class of SERVER in your database backup.
– Back up a primary storage pool to a copy storage pool or associated with a
device class of SERVER.

Chapter 29. Protecting and recovering the server infrastructure and client data 883
 v To copy active data, use the COPY ACTIVEDATA command. For example, to copy
active data from a primary storage pool named BACKUPPOOL to an active-data
pool named CLIENT-RESTORE, issue the following command:
copy activedata backuppool client-restore maxprocess=4

The primary storage pool must have a data format of NATIVE or NONBLOCK.
Copies from primary storage pools with any of the NDMP formats are not
permitted. The only files copied to the CLIENT-RESTORE pool are active backup
files for which a copy does not exist in the active-data pool.

Results

Because backups and active-data copies are made incrementally, you can cancel the
processes. If you reissue the BACKUP STGPOOL or COPY ACTIVEDATA command, the
backup or active-data copy continues from the point at which the process was
canceled.

Restrictions:
v If a backup is to be made to a copy storage pool and the file exists with the
same insertion date, no action is taken. Similarly, if a copy is to be made to an
active-data pool and the file exists with the same insertion data, no action is
taken.
v When a disk storage pool is backed up, cached files (copies of files that remain
on disk after being migrated to the next storage pool) are not backed up.
v Files in a copy storage pool or an active-data pool do not migrate to another
storage pool.
v After a file is backed up to a copy storage pool or a copy is made to an
active-data pool, the file might be deleted from the primary storage pool. When
an incremental backup of the primary storage pool occurs, the file is then
deleted from the copy storage pool. Inactive files in active-data pools are deleted
during the process of reclamation. If an aggregate being copied to an active-data
pool contains some inactive files, the aggregate is reconstructed into a new
aggregate without the inactive files.
Related concepts:
“Active-data pools” on page 241
“Copy storage pools” on page 241
“Primary storage pools” on page 240
“Securing sensitive client data” on page 539
Related tasks:
“Backing up the data in a storage hierarchy” on page 265
Chapter 19, “Automating server operations,” on page 619

884 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Example: Scheduling a backup with one copy storage pool:

Create a schedule for backing up two primary storage pools to the same copy
storage pool.

About this task

Assume that you have two primary storage pools: one random access storage pool
(DISKPOOL) and one tape storage pool (TAPEPOOL, with device class
TAPECLASS). Files stored in DISKPOOL are migrated to TAPEPOOL. You want to
back up the files in both primary storage pools to a copy storage pool.

Procedure

To schedule daily incremental backups of the primary storage pools, do the

following steps:
1. Define a copy storage pool called COPYPOOL, with the same device class as
TAPEPOOL, by issuing the following command:
define stgpool copypool tapeclass pooltype=copy maxscratch=50

Note:
a. Because scratch volumes are allowed in this copy storage pool, you do not
need to define volumes for the pool.
b. All storage volumes in COPYPOOL are located onsite.
2. Perform the initial backup of the primary storage pools by issuing the
following commands:
backup stgpool diskpool copypool maxprocess=2
backup stgpool tapepool copypool maxprocess=2
3. Define schedules to automatically run the commands for backing up the
primary storage pools. The commands to schedule are those that you issued in
step 2.

Results

Tips:
v To minimize tape mounts, you can take one or both of the following steps:
– Back up the disk storage pool first, then the tape storage pool.
– If you schedule storage pool backups and migrations and have enough disk
storage, back up or copy as many files as possible from the disk storage pool
to copy storage pools and active-data pools. After the backup and copy
operations are complete, migrate the files from the disk storage pools to
primary tape storage pools.
v if you have active-data pools, you can schedule the COPY ACTIVEDATA command
to copy the active data that is in primary storage pools to the active-data pools.
Related concepts:
“Active-data pools” on page 241
“Copy storage pools” on page 241
“Primary storage pools” on page 240
“Securing sensitive client data” on page 539
Related tasks:
“Backing up the data in a storage hierarchy” on page 265
Chapter 19, “Automating server operations,” on page 619
Chapter 29. Protecting and recovering the server infrastructure and client data 885
 Protecting data in a Centera storage pool:

Performing a storage pool backup for data stored in a Centera storage pool is not
supported. To ensure the safety of the data, therefore, consider using the
replication feature of the Centera storage device.

About this task

With this feature, you can copy data to a replication Centera storage device at a
different location. If the data in the primary Centera storage pool become
unavailable, you can access the replication Centera storage device by specifying its
IP address using the HLADDRESS parameter on the UPDATE DEVCLASS command for
the device class pointed to by the Centera storage pool. After the primary Centera
storage device is re-established, you can issue the UPDATE DEVCLASS command again
and change the value of the HLADDRESS parameter to point back to the primary
Centera storage device. You must restart the server each time you update the
HLADDRESS parameter on the UPDATE DEVCLASS command.
Related concepts:
“Files on sequential volumes (CENTERA)” on page 42

Simultaneous-write operations to copy storage pools and

active-data storage pools
You can set up a primary storage pool so that when a client backs up, archives, or
migrates a file, the file is written to the primary storage pool and simultaneously
stored into each copy storage pool specified for the primary storage pool.

About this task

You can also enable the simultaneous-write function so that active client backup
data is written to active-data pools at the same time it is written to the primary
storage pool. The active-data pools must be specified in the definition of the
primary storage pool, and the clients whose active data is to be saved must be
members of a policy domain that specifies the active-data pool as the destination
for active backup data.

The simultaneous-write function is not intended to replace regular backups of

storage pools. If you use the function to simultaneously write to copy storage
pools and active-data pools, ensure that the copy of each primary storage pool and
of the active-data in each primary storage pool is complete by regularly issuing the
BACKUP STGPOOL command.
Related tasks:
“Writing data simultaneously to primary, copy, and active-data pools” on page 331

Delaying reuse of volumes for recovery purposes

When you define or update a sequential access storage pool, you can use the
REUSEDELAY parameter. This parameter specifies the number of days that must
elapse before a volume can be reused or returned to scratch status after all files are
expired, deleted, or moved from the volume.

886 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 About this task

When you delay reuse of such volumes and they no longer contain any files, they
enter the pending state. Volumes remain in the pending state for the time that is
specified with the REUSEDELAY parameter for the storage pool to which the volume
belongs.

Delaying reuse of volumes can be helpful under certain conditions for disaster
recovery. When files are expired, deleted, or moved from a volume, they are not
erased from the volumes: The database references to these files are removed. Thus
the file data might still exist on sequential volumes if the volumes are not
immediately reused.

A disaster might force you to restore the database using a database backup that is
not the most recent backup. In this case, some files might not be recoverable
because the server cannot find them on current volumes. However, the files might
exist on volumes that are in pending state.

Procedure

You might be able to use the volumes in pending state to recover data by doing
the following steps:
1. Restore the database to a point-in-time before file expiration.
2. Use a primary, copy-storage, or active-data pool volume that is not rewritten
and that contains the expired file at the time of database backup.

Results

If you back up your primary storage pools, set the REUSEDELAY parameter for the
primary storage pools to 0 to efficiently reuse primary scratch volumes. For your
copy storage pools and active-data pools, delay the reuse of volumes for as long as
you keep your oldest database backup.
Related tasks:
“Scenario: Protecting the database and storage pools” on page 898
Related reference:
“Running expiration processing to delete expired files” on page 512

Auditing storage pool volumes

If there are inconsistencies between the information in the database about files in a
storage pool volume and the files themselves, you might not be able to access the
files.

About this task

Use this section to help you audit storage pool volumes for data integrity.

Task Required Privilege Class

Audit volumes in storage pools over which Restricted storage privilege
they have authority
Audit a volume in any storage pool System privilege, unrestricted storage
privilege

Chapter 29. Protecting and recovering the server infrastructure and client data 887
 To ensure that all files are accessible on volumes in a storage pool, audit any
volumes you suspect might have problems by using the AUDIT VOLUME command.
You have the option of auditing multiple volumes using a time range criteria, or
auditing all volumes in a storage pool.

Audit a volume when the following conditions are true:

v The volume is damaged.
v The volume was accessed for a long time, for example, six months
v A read or write error occurs while accessing the volume
v The database was restored to an earlier point-in-time, and the volume is either a
disk volume or a volume that was identified as being reused or deleted since the
database backup

If a storage pool has data validation enabled, run an audit for the volumes in the
storage pool to have the server validate the data.

Note: If Tivoli Storage Manager detects a damaged file on a Centera volume, then
a command is sent to Centera to delete the file. If Centera is unable to delete the
file because the retention period for the file is not expired, then the volume that
contains the file is not be deleted.

To display the results of a volume audit after it completes, use the QUERY ACTLOG
command.
Related tasks:
“Requesting information from the activity log” on page 800

Storage pool volume audit

When you audit a volume, a background process is started, and the results of the
audit are stored in the activity log.

During the auditing process, the server performs the following actions:
v Sends informational messages about processing to the server console.
v Prevents new files from being written to the volume.
v Generates a cyclic redundancy check, if data validation is enabled for the storage
pool.

You can specify whether you want the server to correct the database if
inconsistencies are detected. Tivoli Storage Manager corrects the database by
deleting database records that refer to files on the volume that cannot be accessed.
The default is to report inconsistencies that are found (files that cannot be
accessed), but to not correct the errors.

If files with read errors are detected, their handling depends on the following
conditions:
v The type of storage pool to which the volume is assigned
v The FIX parameter on the AUDIT VOLUME command
v The location of file copies (whether a copy of the file exists in a copy storage
pool)

888 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Errors in an audit of a primary storage pool volume:

When a volume in a primary storage pool is audited, the setting of the FIX
parameter determines how errors are handled.

The FIX parameter on an AUDIT VOLUME command can have the following effects:
FIX=NO
The server reports, but does not delete, any database records that refer to
files found with logical inconsistencies. If the AUDIT VOLUME command
detects a read error in a file, the file is marked as damaged in the database.
You can do one of the following actions:
v If a backup copy of the file is stored in a copy storage pool, you can
restore the file by using the RESTORE VOLUME or RESTORE STGPOOL
command.
v If the file is a cached copy, you can delete references to the file on this
volume by using the AUDIT VOLUME command again. Specify FIX=YES.
If the AUDIT VOLUME command does not detect a read error in a damaged
file, the file state is reset, and the file can be used. For example, if a dirty
tape head caused some files to be marked damaged, you can clean the
head and then audit the volume to make the files accessible again.
FIX=YES
Any inconsistencies are fixed as they are detected.
If the AUDIT VOLUME command detects a read error in a file:
v If the file is not a cached copy and a backup copy is stored in a copy
storage pool, the file is marked as damaged in the database. The file can
then be restored using the RESTORE VOLUME or RESTORE STGPOOL
command.
v If the file is not a cached copy and a backup copy is not stored in a copy
storage pool, all database records that refer to the file are deleted.
v If the file is a cached copy, the database records that refer to the cached
file are deleted. The primary file is stored on another volume.
If the AUDIT VOLUME command does not detect a read error in a damaged
file, the file state is reset, and the file can be used. For example, if a dirty
tape head caused some files to be marked damaged, you can clean the
head and then audit the volume to make the files accessible again.

Errors in an audit of copy storage pool volumes:

When a volume in a copy storage pool is audited, the setting of the FIX parameter
determines how errors are handled.

The FIX parameter on an AUDIT VOLUME command can have the following effects:
FIX=NO
The server reports the error and marks the file copy as damaged in the
database.
FIX=YES
The server deletes references to the file on the audited volume from the
database.

Chapter 29. Protecting and recovering the server infrastructure and client data 889
 Errors in an audit of active-data storage pool volumes:

When a volume in an active-data storage pool is audited, the setting of the FIX
parameter determines how errors are handled.

The FIX parameter on an AUDIT VOLUME command can have the following effects:
FIX=NO
The server reports the error and marks the file copy as damaged in the
database.
FIX=YES
The server deletes references to the file on the audited volume from the
database. The physical file is deleted from the active-data pool.

When auditing a volume in an active-data pool, the server skips inactive files in
aggregates that were removed by reclamation. These files are not reported as
skipped or marked as damaged.

Data validation during audit volume processing

Data validation for storage pools allows the server to validate that data sent to a
device during a write operation matches what the server later reads.

About this task

Data validation is helpful if you introduce new hardware devices. The validation
assures that the data is not corrupted as it moves through the hardware, and then
is written to the volume in the storage pool. You can use the DEFINE STGPOOL or
UPDATE STGPOOL commands to enable data validation for storage pools.

When you enable data validation for an existing storage pool, the server validates
data that is written from that time forward. The server does not validate existing
data which was written to the storage pool before data validation was enabled.

When data validation is enabled for storage pools, the server generates a cyclic
redundancy check (CRC) value and stores it with the data when it is written to the
storage pool. The server validates the data when it audits the volume, by
generating a cyclic redundancy check and comparing this value with the CRC
value stored with the data. If the CRC values do not match, then the server
processes the volume in the same manner as a standard audit volume operation.
This process can depend on the following conditions:
v The type of storage pool to which the volume is assigned
v The FIX parameter of the AUDIT VOLUME command
v The location of file copies (whether a copy of the file exists in a copy storage
pool or an active-data pool)

Check the activity log for details about the audit operation.

The server removes the CRC values before it returns the data to the client node.
Related reference:
“Errors in an audit of active-data storage pool volumes”
“Errors in an audit of copy storage pool volumes” on page 889
“Errors in an audit of a primary storage pool volume” on page 889

890 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 Choosing when to enable data validation:

Data validation is available for nodes and storage pools. The forms of validation
are independent of each other.

About this task

Figure 107 shows data validation:

v During a client session with the server
2
v During a client session with the storage agent
1 (the storage agent reads the
VALIDATEPROTOCOL setting for the client from the Tivoli Storage Manager
server)
v During a storage agent session with the server
3
v When a server (including a storage agent) sends data to the storage pool
4 or

5
You can enable data validation for one or more nodes, storage agents, or storage
pools. Figure 107 illustrates data transfer that is eligible for data validation within a
Tivoli Storage Manager environment. Your environment may contain some or all of
these objects.

Tivoli Storage
Manager
client
1 2

3 Tivoli Storage
Storage
Manager
Agent
server

4 5

Storage
Pool

Figure 107. Data transfer eligible for data validation

Table 72 provides information that relates to Figure 107. This information explains
the type of data being transferred and the appropriate command to issue.
Table 72. Setting data validation
Where to Set
Numbers in Data Type of Data
Figure 107 Validation Transferred Command Command Parameter Setting

1 Node File Data and See Note See Note
definition Metadata

2 Node File Data and REGISTER NODE UPDATE NODE VALIDATEPROTOCOL=ALL or
definition Metadata VALIDATEPROTOCOL=DATAONLY

3 Server Metadata DEFINE SERVER UPDATE VALIDATEPROTOCOL=ALL
definition SERVER
(storage agent
only)

Chapter 29. Protecting and recovering the server infrastructure and client data 891
Table 72. Setting data validation (continued)
Numbers in Where to Set
Figure 107 Data Type of Data
on page 891 Validation Transferred Command Command Parameter Setting

4 Storage pool File Data DEFINE STGPOOL UPDATE CRCDATA=YES1
definition STGPOOL
issued on the
Tivoli Storage
Manager
server

5 Storage pool File Data DEFINE STGPOOL UPDATE CRCDATA=YES1
definition STGPOOL
issued on the
Tivoli Storage
Manager
server
1
This parameter is not available for primary sequential access storage pools that use the following data formats:
NETAPPDUMP, CELERRADUMP, NDMPDUMP.

Note: The storage agent reads the VALIDATEPROTOCOL setting for the client from the
Tivoli Storage Manager server.

Figure 108 is similar to the previous figure, however note that the top section
encompassing
1,
2, and
3 is shaded. All three of these data validations are
related to the VALIDATEPROTOCOL parameter. What is significant about this validation
is that it is active only during the client session. After validation, the client and
server discard the CRC values generated in the current session. This is in contrast
to storage pool validation,
4 and
5, which is always active when the storage
pool CRCDATA setting is YES.

The validation of data transfer between the storage pool and the storage agent
4
is managed by the storage pool CRCDATA setting defined by the Tivoli Storage
Manager server. Even though the flow of data is between the storage agent and the
storage pool, data validation is determined by the storage pool definition.
Therefore, if you always want your storage pool data validated, set your primary
storage pool CRCDATA setting to YES.

Tivoli Storage
Manager
client
1 2

Storage 3 Tivoli Storage

Agent Manager
server

4 5

Storage
Pool

Figure 108. Protocol data validation versus storage pool data validation

892 IBM Tivoli Storage Manager for Windows: Administrator's Guide

If the network is unstable, you might decide to enable only data validation for
nodes. Tivoli Storage Manager generates a cyclic redundancy check when the data
is sent over the network to the server. Certain nodes might have more critical data
than others and might require the assurance of data validation. When you identify
the nodes that require data validation, you can choose to have only the user's data
validated or all the data validated. Tivoli Storage Manager validates both the file
data and the file metadata when you choose to validate all data.

When you enable data validation for a server-to-server exchange or between a

storage agent and server, the server must validate all data. You can enable data
validation by using the DEFINE SERVER or UPDATE SERVER command. For
server-to-server exchange, you can use virtual volumes. For data that is exchanged
between a storage agent and the server, refer to the Storage Agent User's Guide for
the storage agent's operating system.

If the network is fairly stable but your site is perhaps using new hardware devices,
you might decide to enable only data validation for storage pools. When the server
sends data to the storage pool, the server generates cyclic redundancy checking,
and stores the CRC value with the data. The server validates the CRC value when
the server audits the volume. Later, you might decide that data validation for
storage pools is no longer required after the devices prove to be stable.
Related tasks:
“Using virtual volumes to store data on another server” on page 732
“Auditing storage pool volumes” on page 887
Related reference:
“Validating a node's data during a client session” on page 536

Performance considerations for data validation:

Data validation affects performance because the server requires additional

processor power to calculate and compare CRC values.

Consider the impact on performance when you decide whether data validation is
necessary for storage pools. This method of validation is independent of validating
data during a client session with the server. When you choose to validate storage
pool data, there is no performance impact on the client.

If you enable CRC for storage pools on devices that later prove to be stable, you
can increase performance by updating the storage pool definition to disable data
validation.

Performing storage pool data validation:

Use the AUDIT VOLUME command to specify an audit for data written to volumes
within a range of days, or to run an audit for a storage pool.

About this task

You can manage when the validation of data in storage pools occurs by scheduling
the audit volume operation. You can choose a method suitable to your
environment, for example:
v Select volumes at random to audit. A random selection does not require
significant resources or cause much contention for server resources but can
provide assurance that the data is valid.

Chapter 29. Protecting and recovering the server infrastructure and client data 893
 v Schedule a daily audit of all volumes written in the last day. This method
validates data written to a storage pool on a daily basis.
v Audit volumes in storage pools only for client nodes that are considered to be
critical users.

Auditing a disk storage pool volume

When you audit a disk storage pool volume, you can specify that only summary
messages are sent to the activity log and server console.

About this task

To display the results of a volume audit after it completes, you can issue the QUERY
ACTLOG command.

Procedure

To specify that only summary messages for d:\tsm\admvol.1 are sent to the
activity log and server console, issue the following command:
audit volume d:\adsm\admvol.1 quiet=yes

The audit volume process is run in the background and the server returns the
following message:
ANR2313I Audit Volume (Inspect Only) process started for volume
D:\TSM\ADMVOL.1 (process ID 4).

To view the status of the audit volume process, issue the following command:
query process 4

Here is an example of the audit volume process report:

Process Process Description Status
Number
-------- ------------------------ ---------------------------------------------
4 Audit Volume Volume D:\ADSM\ADMVOL.1 (Storage
(Inspect Only) Pool BACKUPPOOL), Files Processed: 680,
Irretrievable Files Found: 0, Partial Files
Skipped: 0

Auditing multiple volumes in a sequential access storage pool

When you audit a sequential storage volume containing files that span multiple
volumes, the server selects all associated volumes.

The server then begins the audit process with the first volume on which the first
file is stored. For example, Figure 109 on page 895 shows five volumes defined to
ENGBACK2. In this example, File A spans VOL1 and VOL2, and File D spans
VOL2, VOL3, VOL4, and VOL5.

894 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 A
D
B
A D D
E
C
D
VOL1 VOL2 VOL3 VOL4 VOL5

Figure 109. Tape volumes with files a, b, c, d, and e

If you request that the server audit volume VOL3, the server first accesses volume
VOL2, because File D begins at VOL2. When volume VOL2 is accessed, the server
only audits File D. It does not audit the other files on this volume.

Because File D spans multiple volumes, the server accesses volumes VOL2, VOL3,
VOL4, and VOL5 to ensure that there are no inconsistencies between the database
and the storage pool volumes.

For volumes that require manual mount and demount operations, the audit
process can require significant manual intervention.

Auditing a single volume in a sequential access storage pool

To audit a single volume in a sequential storage pool, request that the server skip
any files that span multiple volumes.

About this task

This option is useful when the volume you want to audit contains part of a file,
the rest of which is stored on a different, damaged volume. For example, to audit
only volume VOL5 in the example in Figure 109 and have the server fix any
inconsistencies found between the database and the storage volume, enter:
audit volume vol5 fix=yes skippartial=yes

Auditing volumes by date written

You can limit the audit to volumes that were written in a certain time range.

About this task

When you use the parameters FROMDATE, TODATE, or both, the server limits the audit
to only the sequential media volumes that meet the date criteria, and automatically
includes all online disk volumes. When you include the STGPOOL parameter you
limit the number of volumes that might include disk volumes.

Procedure

Issue the AUDIT VOLUME command with the FROMDATE and TODATE parameters.
For example, to audit the volumes in storage pool BKPOOL1 for volumes written
from March 20, 2002 to March 22, 2002.
audit volume stgpool=bkppool1 fromdate=03/20/2002 todate=03/22/2002

The server audits all volumes that were written to starting at [Link] a.m. on
March 20 and ending at [Link] p.m. on March 22, 2002.

Chapter 29. Protecting and recovering the server infrastructure and client data 895
 Auditing volumes in a specific storage pool
You can limit the audit to volumes in a specified storage pool.

About this task

For example, you can audit the volumes in storage pool BKPOOL1 by issuing the
following command:
audit volume stgpool=bkppool1

Scheduling volume audits

You can schedule periodic volume audits.

About this task

For example, if your critical users store data in storage pool STPOOL3 and you
want all volumes in the storage pool audited every two days at 9:00 p.m., issue the
following command:
define schedule crcstg1 type=administrative
cmd=’audit volume stgpool=stgpool3’ active=yes starttime=21:00 period=2

Fixing damaged files

If files are marked as damaged, you can take steps to correct them.

About this task

A data error, which results in a file being unreadable, can be caused by such things
as a tape deteriorating or being overwritten or by a drive needing cleaning. If a
data error is detected when a client tries to restore, retrieve, or recall a file or
during a volume audit, the file is marked as damaged. If the same file is stored in
other copy storage pools or active-data pools, the status of those file copies is not
changed.

If a client tries to access a damaged file and an undamaged copy is available on an

on-site copy storage pool volume or active-data pool volume, the server sends the
user the undamaged copy.

Iffiles are marked as damaged, you can perform the following operations on them:
v Restore, retrieve, or recall the files
v Move the files by migration, reclamation, or the MOVE DATA command
v Back up during a BACKUP STGPOOL operation if the primary file is damaged
v Restore during a RESTORE STGPOOL or RESTORE VOLUME operation if the backup
copy in a copy storage pool or active-data pool volume is damaged
v Migrate or reclaim during migration and reclamation

896 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Ensuring the integrity of files
There are steps that you can take to ensure the data integrity of user files.

Procedure

To maintain the data integrity of user files, you can perform the following steps:
1. Detect damaged files before the users do. The AUDIT VOLUME command marks a
file as damaged if a read error is detected for the file. If an undamaged copy is
in an on-site copy storage pool or an active-data pool volume, it is used to
provide client access to the file.
2. Reset the damaged status of files if the error that caused the change to
damaged status was temporary. You can use the AUDIT VOLUME command to
correct situations when files are marked damaged due to a temporary hardware
problem, such as a dirty tape head. The server resets the damaged status of
files if the volume in which the files are stored is audited and no read errors
are detected.
3. Correct files that are marked as damaged. If a primary file copy is marked as
damaged and a usable copy exists in a copy storage pool or an active-data pool
volume, the primary file can be corrected using the RESTORE VOLUME or RESTORE
STGPOOL command.
4. Regularly run commands to identify files that are marked as damaged:
v The RESTORE STGPOOL command displays the name of each volume in the
restored storage pool that contains one or more damaged primary files. Use
this command with the preview option to identify primary volumes with
damaged files without actually performing the restore operation.
v The QUERY CONTENT command with the DAMAGED parameter displays damaged
files on a specific volume.
Related tasks:
“Data validation during audit volume processing” on page 890
“Restoring damaged files”

Restoring damaged files

If you use copy storage pools, you can restore damaged client files. You can also
check storage pools for damaged files and restore the files.

About this task

This section explains how to restore damaged files based on the scenario in
“Example: Scheduling a backup with one copy storage pool” on page 885.

If a client tries to access a file stored in TAPEPOOL and a read error occurs, the file
in TAPEPOOL is automatically marked as damaged. Future accesses to the file
automatically use the copy in COPYPOOL as long as the copy in TAPEPOOL is
marked as damaged.

To restore any damaged files in TAPEPOOL, you can define a schedule that issues
the following command periodically:
restore stgpool tapepool

You can check for and replace any files that develop data-integrity problems in
TAPEPOOL or in COPYPOOL. For example, every three months, query the
volumes in TAPEPOOL and COPYPOOL by entering the following commands:
query volume stgpool=tapepool

Chapter 29. Protecting and recovering the server infrastructure and client data 897
 query volume stgpool=copypool

Then issue the following command for each volume in TAPEPOOL and
COPYPOOL:
audit volume <volname> fix=yes

If a read error occurs on a file in TAPEPOOL, that file is marked damaged and an
error message is produced. If a read error occurs on file in COPYPOOL, that file is
deleted and a message is produced.

Restore damaged primary files by entering:

restore stgpool tapepool

Finally, create new copies in COPYPOOL by entering:

backup stgpool tapepool copypool

Scenario: Protecting the database and storage pools

To demonstrate how you might protect the database and storage pools, suppose
that your company takes steps to protect against the permanent loss of data that is
stored in the database and storage pools. You can modify this procedure to meet
your needs.

About this task

This scenario assumes a storage hierarchy that consists of the following storage
pools:
v Default random-access storage pools named BACKUPPOOL, ARCHIVEPOOL,
and SPACEMGPOOL
v A tape storage pool named TAPEPOOL
To provide extra levels of protection for client data, the scenario also specifies an
offsite copy storage pool and an onsite active-data pool.

The standard procedures for the company include the following activities:
v Weekly reclamation of its copy storage pool. Reclamation for the copy storage
pools is turned off at other times.

Note: In a copy storage pool definition, the REUSEDELAY parameter delays

volumes from being returned to scratch or being reused. Set the value high
enough to ensure that the database can be restored to an earlier point in time
and that database references to files in the storage pool are valid. For example,
to retain database backups for seven days and, therefore, sets REUSEDELAY to 7.
v Nightly back up of its primary storage pools to the copy storage pool. Every
night, copy the active client backup data in the primary storage pools to the
active-data pool
v A weekly full backup of the database and incremental backups on the other days
v Daily shipment of the database backup volumes and copy storage pool volumes
to an offsite location

Procedure

To protect client data, perform the following actions:

898 IBM Tivoli Storage Manager for Windows: Administrator's Guide

1. Create a copy storage pool named DISASTER-RECOVERY. Only scratch tapes
are used, and the maximum number of scratch volumes is set to 100. The copy
storage pool is defined by entering:
define stgpool disaster-recovery tapeclass pooltype=copy
maxscratch=100
2. Create an active-data pool named CLIENT-RESTORE and associates it with a
sequential-access disk device class (FILE). Only scratch volumes are used, and
the maximum number of scratch volumes is set to 50. The active-data pool is
defined by entering:
define stgpool client-restore diskclass pooltype=activedata
maxscratch=50
3. Perform the first backup of the primary storage pools. The first backup of a
primary storage pool is a full backup and, depending on the size of the storage
pool, can take a long time.
4. Define schedules for the following daily operations:
a. Run incremental backups of the primary storage pools each night. Issue the
following commands:
backup stgpool backuppool disaster-recovery maxprocess=2
copy activedata backuppool client-restore maxprocess=2
backup stgpool archivepool disaster-recovery maxprocess=2
backup stgpool spacemgpool disaster-recovery maxprocess=2
backup stgpool tapepool disaster-recovery maxprocess=2
copy activedata tapepool client-restore maxprocess=2
The BACKUP STGPOOL commands use multiple, parallel processes to perform
an incremental backup of each primary storage pool to the copy storage
pool. The COPY ACTIVEDATA commands use multiple, parallel processes to
copy the active versions of client backup data to the active-data pool. Only
those files for which a copy does not exist in the copy pool or active-data
pool are backed up.
Consider turning migration off during the rest of the day. You can add a
schedule to migrate from disk to tape. In this way, the backups are done
while the files are still on disk.
b. Change the access mode to offsite for copy storage pool volumes that have
read/write or read-only access, are onsite, and are at least partially filled. To
change the access mode, issue the following command:
update volume * access=offsite location=’vault site info’
wherestgpool=disaster-recovery whereaccess=readwrite,readonly
wherestatus=filling,full
c. Back up the database by using the BACKUP DB command. For example, issue
the following command:
backup db type=incremental devclass=tapeclass scratch=yes

Restriction: Do not run the MOVE DRMEDIA and BACKUP STGPOOL or BACKUP DB
commands concurrently. Ensure that the storage pool backup processes are
complete before you issue the MOVE DRMEDIA command.
5. Perform the following operations nightly after the scheduled operations
completes:
a. Back up the volume history and device configuration files. If they change,
back up the server options files and the database and recovery log setup
information.
b. Move the copy storage pool volumes marked offsite, the database backup
volumes, volume history files, device configuration files, server options
files, and the database and recovery log setup information to the offsite
location.

Chapter 29. Protecting and recovering the server infrastructure and client data 899
 c. Identify offsite volumes that must be returned onsite. For example, issue the
following command:
query volume stgpool=disaster-recovery access=offsite status=empty

These volumes, which became empty through expiration, reclamation, and

file space deletion, waited the delay time that is specified by the REUSEDELAY
parameter. The administrator periodically returns outdated backup database
volumes. These volumes are displayed with the QUERY VOLHISTORYcommand
and can be released for reuse with the DELETE VOLHISTORY command.
6. Bring the volumes that are identified in step 5c onsite and update their access
to read/write.

Recovering the database and client data

If a disaster occurs, recovering the database and client data are among the first
priorities. Depending on your disaster recovery strategy, the recovery site can be
your primary data center or a remote data center. Use the disaster recovery plan to
organize recovery activities.
Related tasks:
Chapter 32, “Disaster recovery manager,” on page 1005

Restoring the database

To restore the database, you must have the database backup volumes. The
database backup volumes can be full, incremental, or snapshot.

About this task

For database restore operations, the Tivoli Storage Manager server reads the
information that is in the volume history file to determine the number of data
streams to read. The server attempts to match the number of streams that were
used during the backup operation. For example, if the backup operation used four
streams, the Tivoli Storage Manager server attempts the restore operation using
four streams.

If you reduce the number of data streams after a database backup operation, this
information will not be available to the server when the database is restored. To
specify fewer data streams for the restore operation, take one or both of the
following actions in the device configuration file:
v Reduce the number of online and usable drive definitions by removing DEFINE
DRIVE commands.
v Update the value of the MOUNTLIMIT parameter of the DEFINE DEVCLASS
command.
Related concepts:
“Configuring concurrent multistreaming” on page 872

900 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Types of server database restores
Point-in-time restores are typically used for situations such as disaster recovery or
to remove the effects of errors that can cause inconsistencies in the database. If you
want to recover the database to the time when the database was lost, recover the
database to its current state.
Point-in-time restore
v Removes and recreates the active log directory and archive log directory
specified in [Link] file
v Restores the database image from backup volumes to the database
directories recorded in a database backup or to new directories
v Restores archive logs from backup volumes to the overflow directory
v Applies logs from the overflow directory up to specified point in time
Database restores using snapshot backups are a form of point-in-time
restore.
Most current restore
v Does not remove and recreate the active log directory or archive log
directory.
v Restores a database image from the backup volumes to the database
directories recorded in a database backup or to new directories.
v Restores archive logs from backup volumes to the overflow directory.
v Applies logs from overflow directory and archive logs from archive log
directory.

Restoring a server database to a point in time:

To restore a database to point in time, you need the latest full backup before the
point in time. You also need the latest incremental backup after that last full
backup. You can also use snapshot database backups to restore a database to a
specific point in time.

Before you begin

Before restoring the database, have available the following infrastructure setup
files:
v Server options file
v Volume history file:
Copy the volume history file pointed to by the server options file. The backup
copy must a different name. If the restore fails and you must try it again, you
might need the backup copy of the volume history file. After the database is
restored, any volume history information pointed to by the server options is lost.
This information is required to identify the volumes to be audited.
If your old volume history file shows that any of the copy storage pool volumes
that are required to restore your storage pools were reused (STGREUSE) or
deleted (STGDELETE), you might not be able to restore all your files. You can
avoid this problem by including the REUSEDELAY parameter when you define
your copy storage pools.
v Device configuration file:
You might need to modify the device configuration file based on the hardware
available at the recovery site. For example, the recovery site might require a
different device class, library, and drive definitions.
v Detailed query output about the database and recovery log

Chapter 29. Protecting and recovering the server infrastructure and client data 901
 If files were migrated, reclaimed, or moved after a backup, the files might be lost
and the space occupied by those files might be reused. You can minimize this loss
by using the REUSEDELAY parameter when defining or updating sequential-access
storage pools. This parameter delays volumes from being returned to scratch or
being reused.

Procedure

To restore the database to a point-in-time, complete the following steps:

1. If the database or recovery log directories were lost, recreate the directories. For
example:
mkdir e:\tsm\db001
mkdir f:\tsm\db001
mkdir g:\tsm\db001
mkdir h:\tsm\activelog
mkdir i:\tsm\archlog
mkdir j:\tsm\archfaillog
2. Use the DSMSERV RESTORE DB utility. For example, to restore the database to a
backup series that was created on April 19, 2009, enter:
dsmserv restore db todate=04/19/2009
The server completes the following actions:
a. Reads the volume history file to locate the last full backup that occurred on
or before the specified date and time.
b. Using the device configuration file, requests a mount of the first volume.
The first volume contains the beginning of the full backup.
c. Restores the backup data from the first volume.
d. Continues to request mounts and to restore data from the backup volumes
that contain the full backup and any incremental backups that occurred on
or before the specified date.
3. From the old volume history information that was generated by the QUERY
VOLHISTORY command, obtain a list of all the volumes that were reused
(STGREUSE), added (STGNEW), and deleted (STGDELETE) since the original
backup. Use this list to perform the remaining steps in this procedure. It might
also be necessary to update the device configurations in the restored database.
4. Issue AUDIT VOLUME command and specify the FIX=YES parameter to audit all
disk volumes, all reused volumes, and all deleted volumes.
The audit volume process identifies files recorded in the database that can no
longer be found on a volume. If a copy of the file is in a copy storage pool or
an active-data pool, the file on the audited volume is marked as damaged.
Otherwise, the file is deleted from the database and is lost.
5. If the audit detects any damaged files, issue the RESTORE STGPOOL command to
restore those files after you audit the volumes in the storage pool.
6. Mark as "destroyed" any volumes that cannot be located, and recover those
volumes from copy storage pool backups. If no backups are available, delete
the volumes from the database by using the DELETE VOLUME command with the
DISCARDDATA=YES parameter.
7. Redefine any storage pool volumes that were added since the database backup.

What to do next

After a restore, the volume inventories for Tivoli Storage Manager and for your
tape management system might be inconsistent. For example, after a database
backup, a new volume is added to Tivoli Storage Manager. The tape management

902 IBM Tivoli Storage Manager for Windows: Administrator's Guide

system inventory records the volume as belonging to Tivoli Storage Manager. If the
database is restored from the backup, Tivoli Storage Manager has no record of the
added volume, but the tape management system does. You must synchronize these
inventories.

Similarly, the volume inventories for Tivoli Storage Manager and for any
automated libraries might also be inconsistent. Issue the AUDIT LIBRARY command
to synchronize these inventories.
Related tasks:
“Updating the device configuration file” on page 906
“Restoring to a point-in-time in a shared library environment” on page 915
“Delaying reuse of volumes for recovery purposes” on page 886

Restoring a server database to its most current state:

You can use full and incremental backups to restore a database to its most current
state. Snapshot database backups are complete database copies of a point in time.

About this task

You can restore a database to its most current state if the last backup series that
was created for the database is available. A backup series consists of a full backup,
the latest incremental backup, and all active and archive logs for database changes
since the last backup in the series was run.

Attention: Recovering the database to its most current state is not possible if the
active or archive logs are lost.

Procedure

To restore a database to its most current state, issue the DSMSERV RESTORE DB
command. For example:
dsmserv restore db

If the original database and recovery log directories are available, use the DSMSERV
RESTORE DB utility to restore the database. However, if the database and recovery
log directories are lost, recreate them first, and then issue the DSMSERV RESTORE DB
utility.

Example: Restoring a library manager database

In this example, a library manager's corrupted database is restored. You can
modify the procedure to meet your needs.

About this task

In a Tivoli Storage Manager shared library environment, the server that manages
and controls the shared library is known as the library manager. The library
manager maintains a database of the volumes within the shared library.

Procedure

Complete the following steps to restore the corrupted database:

Chapter 29. Protecting and recovering the server infrastructure and client data 903
 1. Copy the volume history file to a temporary location and rename the file.
After the database is restored, any volume history information that is pointed
to by the server options is lost. You need this information to identify the
volumes to be audited.
2. Put the device configuration file and the server options file in the server
working directory. You can no longer recreate the device configuration file;
you must have a copy of the original.
3. Gather the outputs from your detailed queries about your database and
recovery log setup information.
4. Determine whether the original database and recovery log directories exist. If
the original database or recovery log directories were lost, recreate them using
the operating system mkdir command.

Note: The directories must have the same name as the original directories.
5. Use the DSMSERV RESTORE DB utility to restore the database to the current time.
6. Start the Tivoli Storage Manager server instance.
7. Issue an AUDIT LIBRARY command from each library client for each shared
library.
8. Create a list from the old volume history information (generated by the QUERY
VOLHISTORY command) that shows all of the volumes that were reused
(STGREUSE), added (STGNEW), and deleted (STGDELETE) since the original
backup. Use this list to perform the rest of this procedure.
9. Audit all disk volumes, all reused volumes, and any deleted volumes located
by the AUDIT VOLUME command using the FIX=YES parameter.
10. Issue the RESTORE STGPOOL command to restore those files detected as
damaged by the audit. Include the FIX=YES parameter on the AUDIT VOLUME
command to delete database entries for files not found in the copy storage
pool or active-data pool.
11. Mark any volumes that cannot be located as destroyed, and recover those
volumes from copy storage pool backups. Recovery from active-data pool
volumes is not suggested unless the loss of inactive data is acceptable. If no
backups are available, delete the volumes from the database by using the
DELETE VOLUME command with the DISCARDDATA=YES parameter.
12. Redefine any storage pool volumes that were added since the database
backup.

Results

Note: When a database is loaded or restored, the server-to-server communication

verification token is changed. The verification token is an attribute of the database
and is not stored in the database itself. Part of the token is the installation date and
time for the database. For servers that are defined for server-to-server
communications, issue an UPDATE SERVER command with FORCESYNC=YES.
Related tasks:
“Restoring to a point-in-time a library manager server” on page 915

904 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Example: Restoring a library client database
In this example, a library client's corrupted database might be restored. You can
modify this procedure to meet your needs.

About this task

In a Tivoli Storage Manager shared library environment, the servers that share a
library and rely on a library manager to coordinate and manage the library usage
are known as library clients. Each library client maintains a database of volume
usage and volume history. If the database of the library client becomes corrupted,
it might be restored by following these steps:

Procedure
1. Copy the volume history file to a temporary location and rename the file.
After the database is restored, any volume history information that is pointed
to by the server options is lost. You need this information to identify the
volumes to be audited.
2. Put the device configuration file and the server options file in the server
working directory. You can no longer recreate the device configuration file;
you must have a copy of the original.
3. Gather the outputs from your detailed queries about your database and
recovery log setup information.
4. Check to see if the original database and recovery log directories exist. If the
original database or recovery log directories were lost, recreate them using the
operating system mkdir command.

Note: The directories must have the same name as the original directories.
5. Use the DSMSERV RESTORE DB utility to restore the database to the current time.
6. Create a list from the old volume history information (generated by the QUERY
VOLHISTORY command) that shows all of the volumes that were reused
(STGREUSE), added (STGNEW), and deleted (STGDELETE) since the original
backup. Use this list to perform the rest of this procedure.
7. Audit all disk volumes, all reused volumes, and any deleted volumes located
by the AUDIT VOLUME command using the FIX=YES parameter.
8. Issue the RESTORE STGPOOL command to restore those files detected as
damaged by the audit. Include the FIX=YES parameter on the AUDIT VOLUME
command to delete database entries for files not found in the copy storage
pool.
9. Mark any volumes that cannot be located as destroyed, and recover those
volumes from copy storage pool backups. If no backups are available, delete
the volumes from the database by using the DELETE VOLUME command with the
DISCARDDATA=YES parameter.
10. Issue the AUDIT LIBRARY command for all shared libraries on this library client.
11. Redefine any storage pool volumes that were added since the database
backup.

Results

Note: When a database is loaded or restored, the server-to-server communication

verification token is changed. The verification token is an attribute of the database
and is not stored in the database itself. Part of the token is the installation date and
time for the database. For servers that are defined for server-to-server
communications, issue an UPDATE SERVER command with FORCESYNC=YES.

Chapter 29. Protecting and recovering the server infrastructure and client data 905
 Related tasks:
“Restoring to a point-in-time a library client server” on page 916

Updating the device configuration file

If a disaster occurs, you might have to restore Tivoli Storage Manager with devices
that are not included in the device configuration file.

About this task

If this occurs, you must update the device configuration files manually with
information about the new devices. Whenever you define, update, or delete device
information in the database, the device configuration file is automatically updated.
This information includes definitions for device classes, libraries, drives, and
servers.

Definitions for paths are included when SRCTYPE is set to SERVER.

Library volume location information is updated in the device configuration file

whenever CHECKIN LIBVOLUME, CHECKOUT LIBVOLUME, and AUDIT LIBRARY commands
are issued for SCSI libraries.

If an automated tape library is used at the recovery site, volume location

information in comments (/*....*/) in the device configuration file must be
modified. First, manually place the physical database backup volumes in the
automated library and note the element numbers where you place them. Then
manually edit the device configuration file to identify the locations of the database
backup volumes so that the server can find them to restore the database.

For virtual volumes, the device configuration file stores the password (in encrypted
form) for connecting to the remote server. If you regressed the server to an earlier
point-in-time, this password might not match what the remote server expects. In
this case, manually set the password in the device configuration file. Then ensure
that the password on the remote server matches the password in the device
configuration file.

Note: Set the password in clear text. After the server is operational again, you can
issue a BACKUP DEVCONFIG command to store the password in encrypted form.
Related tasks:
“Recovering with different hardware at the recovery site” on page 1040
“Automated SCSI library at the original and recovery sites” on page 1040
Related reference:
Automated SCSI library at the original site and a manual scsi library at the
recovery site

906 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Restoring storage pools and storage pool volumes
When you restore a storage pool, the server determines which files are in that
storage pool. Using file copies from a copy storage pool or an active-data pool,
server restores the files that were in the storage pool to the same or a different
storage pool. As part of the restore operation, inactive file versions are deleted
from the server database if the server determines that an inactive file needs to be
replaced but cannot find it in the active-data pool.

About this task

The RESTORE STGPOOL command restores specified primary storage pools that have
files with the following problems:
v The primary copy of the file had read errors during a previous operation. Files
with read errors are marked as damaged.
v The primary copy of the file on a volume that has an access mode of
DESTROYED..
v The primary file is in a storage pool that is UNAVAILABLE, and the operation is
for restore, retrieve, or recall of files to a user, or export of file data.

Restrictions:
v Cached copies of files in a disk storage pool are never restored. References to
any cached files were identified with read errors or cached files that are stored
on a destroyed volume are removed from the database during restore processing.
v Restoring from an active-data pool might cause some or all inactive files to be
deleted from the database if the server determines that an inactive file needs to
be replaced but cannot find it in the active-data pool. Do not consider
active-data pools for recovery of a primary pool unless the loss of inactive data
is acceptable.
v You cannot restore a storage pool defined with a CENTERA device class.
v Restoring from an active-data pool might cause some or all inactive files to be
deleted from the database if the server determines that an inactive file needs to
be replaced but cannot find it in the active-data pool.

Restore processing copies files from a copy storage pool or an active-data pool
onto new primary storage pool volumes. The server then deletes database
references to files on the original primary storage pool volumes. A primary storage
pool volume becomes empty if all files that were stored on that volume are
restored to other volumes. In this case, the server automatically deletes the empty
volume from the database.

After the files are restored, the old references to these files in the primary storage
pool are deleted from the database. Tivoli Storage Manager locates these files on
the volumes to which they were restored, rather than on the volumes on which
they were previously stored. If a destroyed volume becomes empty because all
files were restored to other locations, the destroyed volume is automatically
deleted from the database.

Tivoli Storage Manager uses database information to determine which files to

restore for a storage pool. As a result, restore processing does not require that the
original volumes be accessed. For example, if a primary storage pool volume is
damaged, you can use the RESTORE VOLUME command to recreate files that were
stored on that volume, even if the volume itself is not readable. However, if you
delete the damaged files (DISCARDDATA=YES on the DELETE VOLUME command),
the server removes references from the database to the files in the primary storage

Chapter 29. Protecting and recovering the server infrastructure and client data 907
 pool volume and to copies of the files in copy storage pool volumes and
active-data pool volumes. You cannot restore those files.

Procedure

To restore a storage pool, use the RESTORE STGPOOL command. To identify volumes
that contain damaged primary files, use the PREVIEW=YES parameter. During
restore processing, a message is issued for every volume in the restored storage
pool that contains damaged, noncached files. To identify the specific files that are
damaged on these volumes, use the QUERY CONTENT command.

What to do next
DRM: DRM can help you track your on-site and offsite primary and copy storage pool
volumes. DRM can also query the server and generate a current, detailed disaster recovery
plan for your installation.

Related tasks:
“Fixing damaged files” on page 896

Restoring files to a storage pool with collocation enabled

When restoring files to a primary storage pool that has collocation enabled, the
server restores the files by collocation group, by client node, or by client file space.

About this task

This process preserves the collocation of client files. However, if the copy storage
pool or active-data pool being used to restore files does not have collocation
enabled, restore processing can be slow.

Procedure

If you need to use a copy storage pool or an active-data pool that is not collocated
to restore files to a primary storage pool that is collocated, you can improve
performance by completing the following steps:
1. Restore the files first to a random access storage pool (on disk).
2. Allow or force the files to migrate to the target primary storage pool.
For the random access pool, set the target storage pool as the next storage pool.
Adjust the migration threshold to control when migration occurs to the target
storage pool.
Related tasks:
“Keeping client files together using collocation” on page 357

Fixing an incomplete storage pool restoration

If the restoration of storage pool volumes is incomplete, you can get more
information about the remaining files on those volumes.

About this task

The restoration might be incomplete for one or more of the following reasons:
v Either files were never backed up, or the backup copies were marked as
damaged.
v A copy storage pool or active-data pool was specified on the RESTORE STGPOOL
command, but files were backed up to a different copy storage pool or

908 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 active-data pool. If you suspect this problem, use the RESTORE STGPOOL command
again without specifying a copy storage pool or active-data pool from which to
restore files. You can specify the PREVIEW parameter on the second RESTORE
STGPOOL command, if you do not actually want to restore files.
v Volumes in the copy storage pool or active-data pool needed to perform the
restore operation are offsite or unavailable. Check the activity log for messages
that occurred during restore processing.
v Backup file copies in copy storage pools or active-data pools were moved or
deleted by other processes during restore processing. To prevent this problem,
do not issue the following commands for copy storage pool volumes or
active-data pool volumes while restore processing is in progress:
– MOVE DATA
– DELETE VOLUME and with the DISCARDDATA parameter to YES
– AUDIT VOLUME with FIX parameter set to YES
– MIGRATE STGPOOL
– RECLAIM STGPOOL
v You can prevent reclamation processing for your copy storage pools and
active-data pools by setting the RECLAIM parameter to 100 with the UPDATE
STGPOOL command.

Restoring storage pool volumes

Restore processing copies files from a copy storage pool or an active-data pool
onto new primary storage pool volumes.

About this task

Tivoli Storage Manager uses database information to determine which files to

restore for a volume. As a result, restore processing does not require that the
original volumes be accessed. For example, if a primary storage pool volume is
damaged, you can use the RESTORE VOLUME command to recreate files that were
stored on that volume, even if the volume itself is not readable. However, if you
delete the damaged files (DELETE VOLUME command with the DISCARDDATA parameter
set to YES), the server removes references from the database to the files in the
primary storage pool volume and to copies of the files in copy storage pool
volumes and active-data pool volumes. You cannot restore those files.

After files are restored, the server deletes database references to files on the
original primary storage pool volumes. Tivoli Storage Manager now locates these
files on the volumes to which they were restored, rather than on the volume on
which they were previously stored. A primary storage pool volume becomes empty
if all files that were stored on that volume are restored to other volumes. In this
case, the server automatically deletes the empty volume from the database.

Procedure

To recreate files for one or more volumes that were lost or damaged, use the
RESTORE VOLUME command. The RESTORE VOLUME command changes the access mode
of the volumes being restored to destroyed. When the restoration is complete (when
all files on the volume are restored to other locations), the destroyed volume is
empty and is then automatically deleted from the database.

Chapter 29. Protecting and recovering the server infrastructure and client data 909
 What to do next

Attention:
v Cached copies of files in a disk storage pool are never restored. References to
any cached files that are on a volume that is being restored are removed from
the database during restore processing.
v You can also recreate active versions of client backup files in storage pool
volumes by using duplicate copies in active-data pools. However, do not
consider active-data pools for recovery of a volume unless the loss of inactive
data is acceptable. If the server determines that an inactive file must be replaced
but cannot find it in the active-data pool, restoring from an active-data pool
might cause some or all inactive files to be deleted from the database.
v You cannot restore volumes in a storage pool defined with a CENTERA device
class.

Scenario: Recovering a lost or damaged storage pool volume:

In this scenario, an operator inadvertently destroys a tape volume (DSM087)

belonging to the TAPEPOOL storage pool. A Tivoli Storage Manager administrator
recovers the data stored on the destroyed volume by using the offsite copy storage
pool.

Procedure

The Tivoli Storage Manager administrator completes the following steps:

1. Determine the copy pool volumes that contain the backup copies of the files
that were stored on the volume that was destroyed. Issue the following
command:
restore volume dsm087 preview=yes
This command produces a list of offsite volumes that contain the backed up
copies of the files that were on tape volume DSM087.
2. Set the access mode of the copy volumes identified as UNAVAILABLE to
prevent reclamation.

Note: This precaution prevents the movement of files stored on these volumes
until volume DSM087 is restored.
3. Bring the identified volumes to the on-site location and set their access mode to
READONLY to prevent accidental writes. If these offsite volumes are being
used in an automated library, the volumes must be checked into the library
when they are brought back on-site.
4. Restore the destroyed files. Issue this command:
restore volume dsm087
This command sets the access mode of DSM087 to DESTROYED and attempts
to restore all the files that were stored on volume DSM087. The files are not
restored to volume DSM087, but to another volume in the TAPEPOOL storage
pool. All references to the files on DSM087 are deleted from the database and
the volume itself is deleted from the database.
5. Set the access mode of the volumes used to restore DSM087 to OFFSITE using
the UPDATE VOLUME command.
6. Set the access mode of the restored volumes that are now on-site, to
READWRITE.

910 IBM Tivoli Storage Manager for Windows: Administrator's Guide

7. Return the volumes to the offsite location. If the offsite volumes used for the
restoration were checked into an automated library, these volumes must be
checked out of the automated library when the restoration process is complete.

Fixing an incomplete volume restoration:

When the restoration of a volume might be incomplete, you can get more
information about the remaining files on volumes for which restoration was
incomplete.

About this task

The restoration might be incomplete for one or more of the following reasons:
v Files were either never backed up or the backup copies are marked as damaged.
v A copy storage pool or active-data pool was specified on the RESTORE VOLUME
command, but files were backed up to a different copy storage pool or a
different active-data pool. If you suspect this problem, use the RESTORE VOLUME
command again without specifying a copy storage pool or active-data pool from
which to restore files. You can specify the PREVIEW parameter on the second
RESTORE VOLUME command, if you do not actually want to restore files.
v Volumes in the copy storage pool or active-data pool needed to perform the
restore operation are offsite or unavailable. Check the activity log for messages
that occurred during restore processing.
v Backup file copies in copy storage pools or active-data pools were moved or
deleted by other processes during restore processing. To prevent this problem,
do not issue the following commands for copy storage pool volumes or
active-data pool volumes while restore processing is in progress:
– MOVE DATA
– DELETE VOLUME with the DISCARDDATA parameter set to YES
– AUDIT VOLUME with the FIX parameter set to YES
– MIGRATE STGPOOL
– RECLAIM STGPOOL
You can prevent reclamation processing for your copy storage pools and
active-data pools by setting the RECLAIM parameter to 100 with the UPDATE
STGPOOL command.

Marking volumes as destroyed:

The destroyed volume access mode designates primary volumes for which files are
to be restored.

About this task

This mode permits the restoration of entire volumes. If a volume is designated as

destroyed, the server does not mount that volume for either read or write access.
You can designate a volume as destroyed with either of two commands:
v The RESTORE VOLUME command automatically changes the access mode of the
specified volumes to destroyed.
v Issue the UPDATE VOLUME command with the ACCESS parameter set to
DESTROYED.

Chapter 29. Protecting and recovering the server infrastructure and client data 911
 The destroyed designation is important during restore processing, particularly
when the RESTORE STGPOOL command is used to restore many primary storage pool
volumes after a major disaster.

Procedure

To mark a volume as destroyed, complete the following steps:

1. Designate as destroyed only those volumes that must be restored. If a volume
is known to be usable after a disaster, do not set its access mode to destroyed.
2. After you identify the primary volumes to be restored and set their access
mode to destroyed, you can add new volumes to the storage pool. The new
volumes are used to contain the files as they are restored from the copy storage
pool volumes or active-data pool volumes. The new volumes can also be used
for new files that users back up, archive, or migrate.
3. The destroyed designation permits tracking the files that must still be restored
from copy storage pools or active-data pools. If restore processing ends before
completion, you can restart the restore. Only the files that still reside on
destroyed volumes would need to be restored.

Restore and retrieve files by priority:

When you restore a file that has multiple copies that are stored on different
sequential storage pools, the Tivoli Storage Manager server selects the storage pool
volume to restore the file from by priority. The server also restores or retrieves files
from the client by priority. The priority is based on the status of the volume.

If you did not specify preemption settings for the server, and a mounted volume is
in use by another process, restore operations is pending until the other processes
complete. The restore operation is pending even when an unmounted volume is
available. In Tivoli Storage Manager V7.1, the server prioritizes idle volumes before
unmounted volumes to enable a restore operation to complete faster.

The server selects a storage pool to restore or retrieve a file from in the following
order:
1. Mounted file volume
2. Unmounted file volume
3. Unmounted VTL volume
4. Idle (mounted, but not in use) volume
5. Unmounted, tape library volume
6. Unmounted, stand-alone tape volume
7. In use, VTL or tape volume
8. Unavailable volume
9. Offsite volume
10. Irretrievable file volume

For example, there are three copies of a file stored in three storage pool volumes
that are called V1, V2, and V3. The storage pool volumes have the following status:
v V1 is idle and is stored in a storage tape pool called P1
v V2 is in use and is stored in a storage tape pool called P2
v V3 is unmounted and is stored in a storage tape pool called P3

912 IBM Tivoli Storage Manager for Windows: Administrator's Guide

The file is restored from the V1 volume in the P1 storage pool because the storage
pool volume has a higher priority.

Preventing duplication of restored files

When Tivoli Storage Manager restores data, it might cause some duplication of
restored files. Duplication can occur if primary volumes are not available, and
Tivoli Storage Manager does not have a complete copy storage pool or active-data
pool from which to perform the restore.

About this task

If duplication occurs, Tivoli Storage Manager uses volumes from multiple copy
storage pools or active-data pools to restore the data. This process can result in
duplicate data being restored. To prevent this duplication, keep one complete set of
copy storage pools and one complete set of active-data pools available to the
server. Alternatively, ensure that only one copy storage pool or one active-data
pool has an access of read/write during the restore operation.

Duplication of restored files occurs only when these conditions exist:

v Primary volumes are unavailable or offsite.
v Multiple copy storage pools or active-data pools are available.
v Copy storage pools and active-data pools do not contain all of the files that are
in the primary storage pools.

The following example explains this scenario:

The primary storage pool Main contains volumes Main1, Main2, and Main3.
v Main1 contains files File11, File12, File13
v Main2 contains files File14, File15, File16
v Main3 contains files File17, File18, File19

The copy storage pool DuplicateA contains volumes DupA1, DupA2, and DupA3.
v DupA1 contains copies of File11, File12
v DupA2 contains copies of File13, File14
v DupA3 contains copies of File15, File16, File17, File18 (File19 is missing because
BACKUP STGPOOL was run on the primary pool before the primary pool
contained File 19.)

The copy storage pool DuplicateB contains volumes DupB1 and DupB2.
v DupB1 contains copies of File11, File12
v DupB2 contains copies of File13, File14, File15, File16, File17, File18, File19

If you do not designate copy storage pool DuplicateB as the only copy storage
pool to have read/write access for the restore operation, then Tivoli Storage
Manager can choose the copy storage pool DuplicateA, and use volumes DupA1,
DupA2, and DupA3. Because copy storage pool DuplicateA does not include file
File19, Tivoli Storage Manager would then use volume DupB2 from the copy
storage pool DuplicateB. The program does not track the restoration of individual
files, so File15, File16, File17, and File18 are restored a second time, and duplicate
copies are generated when volume DupB2 is processed.

Chapter 29. Protecting and recovering the server infrastructure and client data 913
 Restoring and recovering an LDAP server
If you use an LDAP directory server to authenticate passwords, you might need to
restore its contents at some time.

There are ways to avoid locking your ID and not being able to logon to the server
or rendering data unavailable.
v Give system privilege class to the console administrator ID.
v Make sure that at least one administrator with system privilege class can access
the server with LOCAL authentication.
v Do not back up the LDAP directory server to the IBM Tivoli Storage Manager
server. An administrator who backs up the Windows Active Directory or the
IBM Tivoli Directory Server to the Tivoli Storage Manager server might render
them unusable. The Tivoli Storage Manager server requires an external directory
for the initial administrator authentication. Backing up the directory server to
the Tivoli Storage Manager server locks the administrator ID and renders them
unable to logon to the LDAP directory server.

Recovering an LDAP server with replication or other methods:

You must configure the LDAP settings on a target server before replicating,
exporting, or importing nodes and administrators onto it.

You must run the SET LDAPUSER and SET LDAPPASSWORD commands, and define the
LDAPURL option on the target server. If it is not configured properly, you can
replicate, export, import, or use Enterprise Configuration on the target server. But
all nodes and administrators that are transferred from the source to the target with
the LDAP server are then changed to use LDAP authentication. Nodes and
administrators that changed to LDAP authentication on the target server become
inaccessible.

You can configure the target server for LDAP authentication after replicating or
exporting to it, but the data is unavailable until that occurs. After configuring the
LDAP settings at the target server level, the node or administrator entries must be
set up on the LDAP server. Either share the LDAP server between the source and
the target server, or replicate the source LDAP server to the target server. All
applicable nodes and administrators are transferred to the target.

If the transfer is unsuccessful, the LDAP administrator must manually add the
node and administrator passwords onto the LDAP server. Or you can issue the
UPDATE NODE or UPDATE ADMIN commands on the IBM Tivoli Storage Manager server.

Audit the LDAP server

If the IBM Tivoli Storage Manager gets out-of-sync with the LDAP directory server,
you might notice some unexpected errors. To put the data in-sync, issue the AUDIT
LDAPDIRECTORY command.

After you issue the AUDIT LDAPDIRECTORY FIX=YES command, the following events
occur:
v All nodes and administrators that were removed from the LDAP directory
server are listed for you.
v All nodes and administrators that are missing from the LDAP directory server
are listed for you. You can correct these missing entries by issuing the UPDATE
NODE or UPDATE ADMIN command.

914 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 v A summary is provided to you on how many entries were examined on the
LDAP directory server.
v Error messages are issued to you if the Tivoli Storage Manager server cannot
connect to the LDAP directory server. The configuration is controlled by the
LDAPURLoption settings.

If you want to see what is out-of-sync without removing any nodes or

administrators, issue the AUDIT LDAPDIRECTORY FIX=NO command. After issuing this
command, you receive reports on the discrepancies between the Tivoli Storage
Manager server database and the LDAP directory server.

If multiple Tivoli Storage Manager servers share an LDAP directory server, avoid
issuing the AUDIT LDAPDIRECTORY FIX=YES command.

Restoring to a point-in-time in a shared library environment

A point-in-time restore for a library manager server or a library client server
requires additional steps to ensure the consistency of the volume inventories of the
affected servers.

Restoring to a point-in-time a library manager server

A point-in-time restore of a library manager server can create inconsistencies
between the volume inventories of the library manager and library client servers.
Steps must be taken to prevent this problem.

About this task

The restore operation removes all library client server transactions that occurred
after the point in time from the volume inventory of the library manager server.
However, the volume inventory of the library client server still contains those
transactions. New transactions can then be written to these volumes, resulting in a
loss of client data.

Procedure

Complete the following steps after the restore:

1. Halt further transactions on the library manager server: Disable all schedules,
migration, and reclamations on the library client and library manager servers.
2. Audit all libraries on all library client servers. The audits re-enter those volume
transactions that were removed by the restore on the library manager server.
Audit the library clients from the oldest to the newest servers. Use the volume
history file from the library client and library manager servers to resolve any
conflicts.
3. Delete the volumes from the library clients that do not own the volumes.
4. Resume transactions by enabling all schedules, migration, and reclamations on
the library client and library manager servers.

Chapter 29. Protecting and recovering the server infrastructure and client data 915
 Restoring to a point-in-time a library client server
A point-in-time restore of a library client server can cause volumes to be removed
from the volume inventory of a library client server and later overwritten.

About this task

If a library client server acquired scratch volumes after the point-in-time to which
the server is restored, these volumes would be set to private in the volume
inventories of the library client and library manager servers. After the restore, the
volume inventory of the library client server can be regressed to a point-in-time
before the volumes were acquired, thus removing them from the inventory. These
volumes would still exist in the volume inventory of the library manager server as
private volumes owned by the client.

The restored volume inventory of the library client server and the volume
inventory of the library manager server would be inconsistent. The volume
inventory of the library client server must be synchronized with the volume
inventory of the library manager server in order to return those volumes to scratch
and enable them to be overwritten. To synchronize the inventories, complete the
following steps:

Procedure
1. Audit the library on the library client server to synchronize the volume
inventories of the library client and library manager servers.
2. To resolve any remaining volume ownership concerns, review the volume
history and issue the UPDATE VOLUME command as needed.

Example: Recovering to a point-in-time

An administrator restores a library manager server and a library manager client to
the point-in-time of the last backup.

About this task

The processor on which Tivoli Storage Manager is located, the database, and all
on-site storage pool volumes are destroyed by fire. You can use either full and
incremental backups or snapshot database backups to restore a database to a
point-in-time.

DRM: DRM can help you perform these steps.

Do the following steps:

Procedure
1. Install Tivoli Storage Manager on the replacement processor with the same
server options and the same size database and recovery log as on the
destroyed system.
2. Move the latest backup and all of the DISASTER-RECOVERY volumes on-site
from the offsite location.

Note: Do not change the access mode of these volumes until after you
complete step 7 on page 917.
3. If a current, undamaged volume history file exists, save it.

916 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 4. Restore the volume history and device configuration files, the server options,
and the database and recovery log setup. For example, the recovery site might
require different device class, library, and drive definitions.
5. Restore the database from the latest backup level by issuing the DSMSERV
RESTORE DB utility.
6. Change the access mode of all the existing primary storage pool volumes in
the damaged storage pools to DESTROYED. For example, issue the following
commands:
update volume * access=destroyed wherestgpool=backuppool
update volume * access=destroyed wherestgpool=archivepool
update volume * access=destroyed wherestgpool=spacemgpool
update volume * access=destroyed wherestgpool=tapepool
7. Issue the QUERY VOLUME command to identify any volumes in the
DISASTER-RECOVERY storage pool that were on-site at the time of the
disaster. Any volumes that were on-site would were destroyed in the disaster
and could not be used for restore processing. Delete each of these volumes
from the database by using the DELETE VOLUME command with the
DISCARDDATA option. Any files backed up to these volumes cannot be
restored.
8. Change the access mode of the remaining volumes in the
DISASTER-RECOVERY pool to READWRITE. Issue the following command:
update volume * access=readwrite wherestgpool=disaster-recovery

Clients can now access files. If a client tries to access a file that was stored on
a destroyed volume, the retrieval request goes to the copy storage pool. In this
way, clients can restore their files without waiting for the primary storage pool
to be restored. When you update volumes brought from offsite to change their
access, you greatly speed recovery time.
9. Define new volumes in the primary storage pool so the files on the damaged
volumes can be restored to the new volumes. With the new volumes, clients
can also back up, archive, or migrate files to the server. If you use only scratch
volumes in the storage pool, you are not required to complete this step.
10. Restore files in the primary storage pool from the copies in the
DISASTER-RECOVERY pool. To restore files from DISASTER-RECOVERY
pool, issue the following commands:
restore stgpool backuppool maxprocess=2
restore stgpool tapepool maxprocess=2
restore stgpool archivepool maxprocess=2
restore stgpool spacemgpool maxprocess=2

These commands use multiple parallel processes to restore files to primary

storage pools. After all the files are restored for a destroyed volume, that
volume is automatically deleted from the database.
11. To repopulate the active-data pool, copy active versions of backup data from a
primary storage pool to an active-data pool. For example, issue the following
commands:
copy activedata backuppool client-restore maxprocess=2
copy activedata tapepool client-restore maxprocess=2
12. To ensure against another loss of data, immediately back up all storage
volumes and the database. Then resume normal activity, including weekly
disaster backups and movement of data to the offsite location.
Related tasks:
“Updating the device configuration file” on page 906
“Recovering the database and client data” on page 900

Chapter 29. Protecting and recovering the server infrastructure and client data 917
 “Fixing an incomplete storage pool restoration” on page 908

918 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Chapter 30. Replicating client node data
Node replication is the process of incrementally copying, or replicating, data that
belongs to backup-archive client nodes. Data is replicated from one IBM Tivoli
Storage Manager server to another Tivoli Storage Manager server.

The server from which client node data is replicated is called a source replication
server. The server to which client node data is replicated is called a target replication
server. A server can function as the source server of replicated data for some client
nodes and as the target server of replicated data for other client nodes.

The purpose of replication is to maintain the same level of files on the source and
the target replication servers. As part of replication processing, client node data
that was deleted from the source replication server is also deleted from the target
replication server. When client node data is replicated, only the data that is not on
the target replication server is copied.

If a disaster occurs and the source replication server is temporarily unavailable,

client nodes can recover their data from the target replication server. If the source
replication server cannot be recovered, you can convert client nodes to store data
on the target replication server. When there is an outage, the source replication
server can automatically fail over to a target replication server for data recovery.

You can replicate the following types of client node data:

v Active and inactive backup data together, or only active backup data
v Archive data
v Data that was migrated to a source replication server by Tivoli Storage Manager
for Space Management clients

Use node replication for data recovery at a disaster recovery site and to maintain
the same level of files on the source and target replication servers. Node replication
is used for the following objectives:
v Controlling network throughput by scheduling node replication at specific times
v Recovering data from large-scale site loss

To ensure high server availability, you can use node replication with clustering. In
a clustered environment, a client is less likely to fail over to another server. If you
replicate data from several source replication servers to one target replication
server, there is a high dependency on the target replication server. A clustered
environment eases the dependency on the target replication server.

The following table specifies replication compatibility for Tivoli Storage Manager
source and target replication servers.
Table 73. Replication compatibility of Tivoli Storage Manager servers
Compatibility with target replication server
Source replication server version version
V6.3.0 - V6.3.2 V6.3.0 - V6.3.2
V6.3.3 V6.3.3 or later
V6.3.4 or later V6.3.4 or later

© Copyright IBM Corp. 1993, 2013 919

 Table 73. Replication compatibility of Tivoli Storage Manager servers (continued)
Compatibility with target replication server
Source replication server version version
V7.1 V7.1

You can replicate data for client nodes that are V6.3 or earlier. You can also
replicate data that was stored on a Tivoli Storage Manager V6.2 or earlier server
before you upgraded it to V6.3. You can also replicate data that was stored on a
Tivoli Storage Manager V6.2 or earlier server before you upgraded it to V6.3.

For information about implementing a replication configuration, see “Setting up

the default replication configuration” on page 946.
Related concepts:
Managing passwords and logon procedures
“Clustered environment overview” on page 991

Automatic failover for data recovery overview

Automatic failover for data recovery occurs if the source replication server is
unavailable because of a disaster or a system outage.

During normal operations, when the Tivoli Storage Manager V7.1 client logs in to a
source replication server, it receives connection information for the target failover
server. The client node stores the failover connection information in the client
options file. During client restore operations, the Tivoli Storage Manager server
automatically changes clients to the target replication server and back again. Only
one failover server can exist per node at any time. The server information is stored
in the options file. The failover server can be modified only if the default
replication server is modified and another replication is completed for the node.

To use automatic failover for replicated client nodes, the source replication server,
the target replication server, and the client must be at the v7.1 level or later. If any
of the servers are at an earlier level, automatic failover is disabled and you must
rely on manual failover.

If the client cannot connect to the source replication server, it uses the failover
connection information and attempts to log on to the target failover server. The
client logs on to the target replication server and is allowed only to recover data.
The client cannot store data during failover processing.

When a new client operation is started, the client attempts to connect to the source
replication server. The client resumes operations on the source server if the source
replication server is available.

When you configure a node for replication by using the REGISTER NODE command
or the UPDATE NODE command, the failover server is not specified.
Related tasks:
“Restoring, retrieving, and recalling data manually from a target replication server”
on page 988

920 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Source and target node-replication servers
A source replication server is a Tivoli Storage Manager server that has client node
data. You replicate the client node data to another server, called the target
replication server. The client node data on the target replication server is managed
by policies on the source replication server. You cannot replicate nodes from a
Tivoli Storage Manager V6.3.3 or later source replication server to a target
replication server that is running on an earlier version of Tivoli Storage Manager.

Replication server configurations

A source replication server can replicate client node data to only one target
replication server at a time. However, a single target replication server can be the
target for multiple source replication servers. A client node cannot back up,
archive, or migrate data to a server where the node is a target node for replication.

The example configuration in the following figure illustrates the relationships

between source and target replication servers in a network.

NODE4

NODE1
CHICAGO_SRV

PHOENIX_SRV DALLAS_SRV
NODE4 data
NODE1 and NODE2 data

NODE3 data

NODE5 data

NODE2 ATLANTA_SRV

NODE3

NODE5

Figure 110. Networked configuration of replication servers

Figure 110 depicts the following data movements:

v PHOENIX_SRV and DALLAS_SRV replicate client node data to each other. Each
server is both a source and a target for replicated client-node data.

Chapter 30. Replicating client node data 921

 v NODE1 and NODE2 are defined to PHOENIX_SRV. The nodes are configured
for replication. During store operations, NODE1 and NODE2 send their data to
PHOENIX_SRV. The data is then replicated to DALLAS_SRV.
v NODE3 is defined to DALLAS_SRV. The node is configured for replication.
During store operations, NODE3 sends its data to DALLAS_SRV. The data is
then replicated to PHOENIX_SRV.
v Data belonging to NODE4 and NODE5 is replicated to DALLAS_SRV.
Data that is replicated to a target server cannot be replicated again. For example,
the data that belongs to NODE4 cannot be replicated to PHOENIX_SRV from
DALLAS_SRV.

Policy management for node replication

The management class and storage hierarchy that are used by a client node on a
target replication server depend on whether the domain for the node exists on the
target server.

When a client node is registered on a target replication server, the domain for the
node is sent to the target server. If the target server does not have a domain with
the same name, the node on the target server is placed in the standard domain on
the target server and bound to the default management class.

To maintain the same number of file versions on the source and the target
replication servers, the source replication server manages file expiration and
deletion. If a file on the source replication server is marked for deletion, but not
yet deleted by the expiration processing, the target replication server deletes the
file during the next replication process. Expiration processing on the target
replication server is disabled for replicated data. The file on the target replication
server is deleted by the source replication server after the file is expired and
deleted on the source.

If a client node is removed from replication on the target replication server, the
policies on the target replication server are enabled. Data on the target replication
server is then managed by the policies that are on the target replication server, and
expiration processing can delete expired files.

Important: Policies that are defined on replication servers and that are dissimilar
can cause undesirable side-effects. As newer versions of backup files are replicated,
versions that exceed the value of the VEREXISTS parameter for the copy group are
marked for immediate deletion. If the node that owns the files is configured for
replication, expiration does not delete the files. However, because these files are
marked for immediate deletion, they are not available for the client to restore. The
files remain in the storage pool until replication deletes them based on the policy
on the source replication server.

Tips:
v Policies and storage pool hierarchies on the source and target replication servers
can be different. You can use deduplicated storage pools on the source
replication server, on the target replication server, or both. However, to maintain
synchronization of the data on source and target replication servers, configure
the management classes on the source and target replication servers to manage
data similarly. To coordinate policies, consider using Tivoli Storage Manager
enterprise configuration.
v Ensure that sufficient space is available in the storage pool on the target
replication server.

922 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 v If a replicated file on the target replication server is deleted, it is eligible for
replication the next time replication runs for the client node that owns the file.

Node replication processing

Replication processing involves the interaction between replication rules, states,
and modes. Data deduplication and the updating of client node attributes are also
factors in node replication processing.

Replication rules
Replication rules control what data is replicated and the order in which it is
replicated.

Replication rule definitions

Replication rules are assigned to data types. Data types include backup data,
archive data, and space-managed data that is migrated by Tivoli Storage Manager
for Space Management clients.

The Tivoli Storage Manager server has the following predefined set of replication
rules. You cannot create replication rules.
ALL_DATA
Replicates backup, archive, or space-managed data. The data is replicated
with a normal priority. For example, you can assign the ALL_DATA rule to
backup data and archive data, and assign a different rule to
space-managed data.
ACTIVE_DATA
Replicates only active backup data. The data is replicated with a normal
priority. You can assign this rule only to the backup data type.
ALL_DATA_HIGH_PRIORITY
Replicates backup, archive, or space-managed data. The data is replicated
with a high priority. In a replication process that includes both
normal-priority and high-priority data, high-priority data is replicated first.
ACTIVE_DATA_HIGH_PRIORITY
Replicates active backup data. The data is replicated with a high priority.
You can assign this rule only to the backup data type.
DEFAULT
Replicates data according to the rule that is assigned to the data type at the
next higher level in the replication-rule hierarchy. The replication-rule
hierarchy comprises file space rules, individual client-node rules, and
server rules. Server rules apply collectively to all nodes that are defined to
a source replication server and that are configured for replication.
Rules that are assigned to data types in file spaces take precedence over
rules that are assigned to data types for individual nodes. Rules that are
assigned to data types for individual nodes take precedence over server
rules. For example, if the DEFAULT replication rule is assigned to back up
data in a file space, the server checks the replication rule for backup data
that is assigned to the client node. If the client node rule for backup data is
DEFAULT, the server checks the server rule for backup data.
The DEFAULT rule is valid only for data types at the file space level and
the client node level. It is not valid for data types at the server level.

Chapter 30. Replicating client node data 923

 NONE
Backup, archive, or space-managed data is not replicated. For example, if
you do not want to replicate archive data, assign the NONE rule to the
archive data type.

Tip: When you set up the default replication configuration, you do not have to
assign or change replication rules. Tivoli Storage Manager automatically assigns
the DEFAULT replication rule to all data types in the file spaces and in the client
nodes that you configured. The system-wide replication rules are automatically set
to ALL_DATA. You can change file space, client-node, and system-wide rules after
you set up the default configuration.

If a file space is added to a client node that is configured for replication, the file
space rules for data types are initially set to DEFAULT. If you do not change the
file space rules, the client node and server rules determine whether data in the file
space is replicated.

Replication rule attributes

Replication rules have attributes. You can modify the replication state of a rule.

A replication rule has the following attributes:

Rule name
Specifies the name of a replication rule, such as ALL_DATA,
ALL_DATA_HIGH_PRIORITY, and so on.
Priority
Specifies the order in which data is replicated. Data can be replicated with
normal priority or high priority. In a replication process that includes both
normal-priority and high-priority data, high-priority data is replicated first.
The replication rules ALL_DATA and ACTIVE_DATA are normal priority.
The replication rules ALL_DATA_HIGH_PRIORITY and
ACTIVE_DATA_HIGH_PRIORITY are high priority.
Active-data only?
Specifies that the rule replicates only active backup data. The value of this
attribute is YES for the rules ACTIVE_DATA and
ACTIVE_DATA_HIGH_PRIORITY.
State Specifies whether a replication rule is enabled or disabled. Before data is
replicated, the state of the rule that applies to the data is checked. If the
state is enabled, the data is replicated. If the state is disabled, the data is
skipped during replication.
You can change the state of a replication rule.

To display the attributes of replication rules, issue the QUERY REPLRULE command.

Replication rule hierarchy

Replication rules are applied to data types in file spaces and in individual client
nodes. Replication rules are also applied collectively at the server level to all client
nodes that are defined to the server and that are configured for replication. File
space rules take precedence over rules for individual nodes. Rules for individual
nodes take precedence over server rules.

In a client node that is configured for replication, each file space has three
replication rules. One rule applies to backup data in the file space. The other rules
apply to archive data and to space-managed data. The rules for the file space exist
regardless of whether the file space has backup, archive, or space-managed data. If

924 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 a file space does not contain a particular type of data, the rule for the data type in
the file space is ignored during replication processing.

Similarly, each client node that is configured for replication has replication rules for
backup data, archive data, and space-managed data. Client node rules apply to all
the file spaces that belong to a node. Replication rules also exist at the server level
that applies collectively to every client node that is configured for replication on a
source replication server.

The following figure shows the replication rule hierarchy.

File space File space File space

BACKUP DATA RULE BACKUP DATA RULE BACKUP DATA RULE

ARCHIVE DATA RULE ARCHIVE DATA RULE ARCHIVE DATA RULE

SPACE-MANAGED DATA RULE SPACE-MANAGED DATA RULE SPACE-MANAGED DATA RULE

BACKUP DATA RULE BACKUP DATA RULE

ARCHIVE DATA RULE ARCHIVE DATA RULE

SPACE-MANAGED DATA RULE SPACE-MANAGED DATA RULE

NODE NODE

Source replication server

BACKUP DATA RULE

ARCHIVE DATA RULE

SPACE-MANAGED DATA RULE

Figure 111. Replication-rule hierarchy

During replication processing, file space rules take precedence over rules for
individual nodes. Rules for individual nodes take precedence over server rules.
The replication rule that has precedence is called the controlling replication rule.

Example: Processing replication rules

In this example, replication rules are assigned to data types in file spaces, client
nodes, and the server.

The following figure shows an example replication configuration that comprises

two client nodes and three file spaces. This figure shows how Tivoli Storage
Manager processes replication rules.

Chapter 30. Replicating client node data 925

 File space /a File space /b File space /a

BACKUP DATA RULE = BACKUP DATA RULE = BACKUP DATA RULE =

DEFAULT ACTIVE_DATA ALL_DATA_HIGH_PRIORITY
ARCHIVE DATA RULE = ARCHIVE DATA RULE = ARCHIVE DATA RULE =
DEFAULT DEFAULT DEFAULT
SPACE-MANAGED DATA RULE = SPACE-MANAGED DATA RULE = SPACE-MANAGED DATA RULE =
DEFAULT DEFAULT DEFAULT

BACKUP DATA RULE = BACKUP DATA RULE =

DEFAULT DEFAULT
ARCHIVE DATA RULE = ARCHIVE DATA RULE =
DEFAULT DEFAULT
SPACE-MANAGED DATA RULE = SPACE-MANAGED DATA RULE =
NONE DEFAULT
NODE 1 NODE 2

Source replication server

BACKUP DATA RULE =
ALL_DATA
ARCHIVE DATA RULE =
ALL_DATA_HIGH_PRIORITY
SPACE-MANAGED DATA RULE =
ALL_DATA

REPLICATE NODE1, NODE2 command

Replication process

High priority data

File space /a, NODE1: File space /b, NODE1: File space /a, NODE2:
ALL BACKUP DATA
ARCHIVE DATA ARCHIVE DATA ARCHIVE DATA

Normal priority data

File space /a, NODE1: File space /b, NODE1: File space /a, NODE2:

ALL BACKUP DATA ACTIVE BACKUP DATA SPACE-MANAGED DATA

Target replication
server

Figure 112. Replication configuration example

When the REPLICATE NODE command is issued, a single replication process begins.
The source replication server identifies client nodes that are configured for
replication and the rules that apply to the file spaces in nodes that are enabled.

926 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 For example, the file spaces in NODE1 and NODE2 contain important archive
data. Replication of that data is a high priority. The source replication server checks
the rules for archive data in file spaces. Because the rules are set to DEFAULT, the
source replication server checks the client-node replication rules for archive data.
These rules are also set to DEFAULT. As a result, the source replication server
checks the server replication rule for archive data. Because it is set to
ALL_DATA_HIGH_PRIORITY, archive data in file spaces that belong to NODE1
and NODE2 is replicated before normal-priority data.

The backup data in file space /a that belongs to NODE2 is also high priority. The
file space rule for backup data, which is ALL_DATA_HIGH_PRIORITY, takes
precedence over the client node rule of DEFAULT and the server rule of
ALL_DATA.

Tips:
v Figure 112 on page 926 shows one possible configuration to achieve the specified
results. In general, multiple configurations can exist that accomplish the same
purpose.
For example, to replicate archive data first, you can assign the
ALL_DATA_HIGH_PRIORITY replication rule to the archive data type in each
file space that belongs to NODE1 and NODE2.
v Figure 112 on page 926 shows one replication process. To replicate certain client
nodes ahead of other client nodes, you can issue multiple REPLICATE NODE
commands in sequence, either manually or in a maintenance script. Each
command can specify a different client node or different file spaces in an
individual client node. For example, suppose NODE1 contains a large amount of
data and you want to conserve bandwidth. To replicate client node data
sequentially, you can specify NODE1 in a single REPLICATE NODE command and
NODE2 in another REPLICATE NODE command.
Related concepts:
“Replication rule hierarchy” on page 924
“Replication rule definitions” on page 923

Replication state
Replication state indicates whether replication is enabled or disabled. When you
disable replication, replication does not occur until you enable it.

Replication states apply to the following objects:

Replication rules
The state of a replication rule can be enabled or disabled. If a replication
rule is disabled, data to which the rule is assigned is not replicated.
For example, suppose that you have 500 client nodes that are defined to a
source replication server. The replication rule that is assigned to archive
data in some file spaces is ALL_DATA_HIGH_PRIORITY. The server rule
that is assigned to archive data is also ALL_DATA_HIGH_PRIORITY. If
you disable the ALL_DATA_HIGH_PRIORITY rule, archive data in any file
space that is assigned that rule is skipped during replication processing.
Individual client nodes
When you disable replication for a client node, backup, archive, and
space-managed data in file spaces that belong to the client node is skipped
during replication processing. Disabling replication for a client node can be

Chapter 30. Replicating client node data 927

 useful, for example, if the data that belongs to the node is corrupted or
contaminated. After the problem is fixed, you can enable replication.
For replication to occur, the client node must be enabled on the source and
the target replication servers.
Data types in file spaces
You can enable or disable replication of one or more data types in a file
space. For example, you can disable replication of backup data in a file
space.
Disabling replication of a data type in a file space can be useful if, for
example, a file space has large amounts of backup data and archive data.
You can enable the replication of backup data and disable replication of
archive data. After the backup data is replicated, you can enable replication
of archive data.
You cannot disable replication for an entire file space. You can only disable
replication for an entire file space by setting the state of each data type to
disabled.
During replication, the replication state of a file-space data type is checked
only on the source replication server. The replication state of a file space
data type on the target replication server is irrelevant.

Figure 113 on page 929 shows the interaction of replication states and replication
rules. In the example, NODE1 has a single file space /a that contains archive data.
Assume that the replication state of NODE1 on the target replication server is
ENABLED and that replication processing for all nodes is enabled.

928 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 REPLICATE NODE NODE1

Replication DISABLED NODE1 data is

state of NODE1 not replicated
Archive data ?
in /a
is not replicated
ENABLED
NONE

ALL_DATA
File space ENABLED Replication Archive data
ALL_DATA_HIGH_PRIORITY DISABLED
rule for state of archive in /a
archive data data type is not replicated
? in /a
?
DEFAULT

ALL_DATA
ALL_DATA_HIGH_PRIORITY Node rule NONE Archive data
for in /a
archive data is not replicated
?

DEFAULT

System-wide NONE Archive data

rule for in /a
archive data is not replicated
?

ALL_DATA
ALL_DATA_HIGH_PRIORITY

State DISABLED Archive data

of replication in /a
rule is not replicated
?

ENABLED

Archive data
in /a
is replicated

Replication
processing
ends

Figure 113. Replication state example

During replication processing, the following replication states are checked:

1. NODE1
2. Archive data type in file space /a

Chapter 30. Replicating client node data 929

 3. Replication rule for archive data

To determine the replication state of a file space, issue the QUERY FILESPACE
command. To determine the replication state of a client node, issue the QUERY NODE
command, and to determine the replication state of a rule, issue the QUERY
REPLRULE command.

Replication mode
Replication mode is part of a client node definition and indicates whether a client
node is set up to send or receive replicated data. The replication mode can also
indicate whether the data that belongs to a client node is to be synchronized the
first time that replication occurs. Data synchronization applies only to client nodes
whose data was exported from the source replication server and imported on the
target replication server.

The following modes are possible for a client node whose data is not being
synchronized:
SEND Indicates that the client node is set up to send data to the target replication
server. The SEND replication mode applies only to the client node
definition on a source replication server.
RECEIVE
Indicates that the client node is set up to receive replicated data from the
source replication server. The RECEIVE replication mode applies only to
the client node definition on a target replication server.
NONE
The client node is not configured for replication. To be configured for
replication, the client node must be enabled or disabled for replication.

If the data that belongs to a client node was previously exported from a source
replication server and imported on a target replication server, the data must be
synchronized. Synchronization is also required after a database restore to preserve
the client node data that is on the target replication server. When the data that
belongs to a client node is synchronized, entries in the databases of the source and
target replication servers are updated.

The following special settings for replication mode are required to synchronize
data.

Restriction: To synchronize data, the date of the imported data on the target
replication server must be the original creation date.
SYNCSEND
Indicates that data that belongs to the client node on the source replication
server is to be synchronized with the client node data on the target
replication server. The SYNCSEND mode applies only to the client node
definition on a source replication server.
When data synchronization is complete, the replication mode for the node
on the source replication server is set to SEND.
SYNCRECEIVE
Indicates that data that belongs to the client node on the target replication
server is synchronized with the client node data on the source replication
server. This SYNCRECEIVE mode applies only to the client node definition
on the target replication server.

930 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 When data synchronization is complete, the replication mode for the node
on the target replication server is set to RECEIVE.

Replication of deduplicated data

Data deduplication is a method for eliminating redundant data that is stored in
sequential-access disk (FILE) primary storage pools, copy storage pools, and
active-data storage pools. Before the data is replicated, the source replication server
determines whether storage pools are set up for data deduplication.

Restriction: During replication processing, the simultaneous-write function is

disabled on the target replication server when you store data to a primary storage
pool that is enabled for data deduplication. Data that is replicated consists of only
files or extents of data that do not exist on the target replication server.

The following table shows the results when storage pools on source and target
replication servers are enabled for data deduplication. The destination storage pool
is specified in the backup or archive copy-group definition of the management
class for each file. If the destination storage pool does not have enough space and
data is migrated to the next storage pool, the entire file is sent, whether the next
storage pool is set up for deduplication.

And the destination storage

If the storage pool on the pool on the target
source replication server is replication server is The result is
Enabled for data Enabled for data Only extents that are not
deduplication deduplication stored in the destination
storage pool on the target
replication server are
transferred.
Enabled for data Not enabled for data Files are reassembled by the
deduplication deduplication source replication server and
replicated in their entirety to
the destination storage pool.
Not enabled for data Enabled for data The source replication server
deduplication deduplication determines whether any
extents were identified for
files that were previously
stored in deduplicated
storage pools. Any files that
were never in a deduplicated
storage pool are replicated in
their entirety. For files that
had extents that were
previously identified, only
extents that do not exist in
the destination storage pool
are transferred.
Not enabled for data Not enabled for data Files are replicated in their
deduplication deduplication entirety to the destination
storage pool.

Tip: If you have a primary storage pool that is enabled for deduplication on a
source replication server, you can estimate a size for a new deduplicated storage
pool on the target replication server. Issue the QUERY STGPOOL command for the
primary deduplicated storage pool on the source replication server. Obtain the

Chapter 30. Replicating client node data 931

 value for the amount of storage space that was saved in the storage pool as a
result of server-side data deduplication. This value is represented by the field
Duplicate Data Not Stored in the command output. Subtract this value from the
estimated capacity of the storage pool.
Related tasks:
“Deduplicating data” on page 285
“Writing data simultaneously to primary, copy, and active-data pools” on page 331

Client node attributes that are updated during replication

During node replication, the source replication server updates the client node
attributes that are part of the node definition that is on the target replication server.

A Tivoli Storage Manager V7.1 server can replicate any node definition that is set
as a proxy agent for a replicating node. The associated replication node definition
must be replicated to the target replication server for failover to occur.

The following client node attributes are updated during node replication:
v Aggregation
v Automatic file space rename
v Archive delete authority
v Backup delete authority
v Backup initiation
v Cipher strength
v Client option sets
v Compression option
v Contact
v Data-read path
v Data-write path
v Email address
v File space access rules that are created with the client SET ACCESS command
v High-level address
v Low-level address
v Node lock state
v Operating system
v Option set name
v Password
v Password expiration period
v Proxy node
v Role override
v Session initiation
v Transaction group maximum
v URL
v Validate protocol

932 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 Attention:
v Ensure that all servers in your replication configuration do not create client
option set conflicts on target replication servers. When an option set is created,
all client nodes that reference that option set use the options that are defined.
Client nodes that are replicated from different source replication servers can use
the same option set. Only the first option set that is replicated is defined on the
target replication server.
v The node password for the client node is updated during replication on the
source replication server by one server and on the target replication server by a
different server. Because authentication can happen on a Lightweight Directory
Access Protocol (LDAP) directory server or the Tivoli Storage Manager server,
data can be lost. If dual authentication occurs, the password is not updated
during replication.

The following client node attributes are not updated during node replication:
v Domain name (might not exist on target replication server)
v Node conversion type
v Target level for automatic client deployment

For details about client node attributes, go to Tivoli Documentation Central at

[Link]
Documentation Central, and select Tivoli Storage Manager. Open the information
center that matches the version of Tivoli Storage Manager that you installed. You
can then search for the attribute.

Node replication restrictions

Restrictions can affect planning and implementation. For example, Tivoli Storage
Manager applies the replication rule for archive data to the data that was migrated
by the HSM for Windows client.

The following restrictions apply to node replication:

Store operations to a target replication server
If a client node is configured for replication, you cannot back up, archive,
or migrate its data to the server that is the target replication server for the
replicated data that belongs to the node.
Failover to a target replication server
You can use automatic failover only with Tivoli Storage Manager V7.1
servers and clients. Only one failover server for each node can be used.
The failover server is always set to the last target server to which the node
was replicated. The client can recover data from the target replication
server, but it cannot store data during failover processing. Issue the QUERY
SESSION command to determine whether a client node is in failover mode.
Client node definition on the target replication server
If you plan to add a node for replication, the client node definition cannot
exist on the target replication server. If the client node definition does exist
on the target replication server, you must remove or rename the node
before replication can occur.
However, if the data that belongs to a client node was exported from the
source replication server and imported on the target replication server, you
do not have to delete the client node definition on the target replication
server. To replicate, the data must be synchronized between the source and
target replication servers. Synchronization occurs during replication.

Chapter 30. Replicating client node data 933

 To synchronize data, the data must be imported with the value of the
DATES parameter that is set to ABSOLUTE on the IMPORT NODE command.
Import and export operations
Data that belongs to a replicated client node cannot be imported to the
target replication server for replication. However, you can export the data
that belongs to the client node from the source replication server to other
servers. To export, you can use media or server-to-server virtual volumes.
Replication rules are not exported.
Data migrated by the HSM for Windows client
The Tivoli Storage Manager for HSM for Windows client provides
hierarchical storage management (HSM) for Windows NTFS file systems.
When the HSM for Windows client stores data on the Tivoli Storage
Manager server, the data is stored as archive data, not as space-managed
data.
During replication processing, Tivoli Storage Manager applies the
replication rule for archive data to the data that was migrated by the HSM
for Windows client. For example, suppose that a backup-archive client has
a file space that contains two directories. The data in one directory is
archived to the Tivoli Storage Manager server. The data in the other
directory is migrated by the HSM for Windows client, but it is stored as
archive data. Both sources of data are associated with the same file space
on the server.
If you set the file space replication rule for archive data to ALL_DATA and
the file-space replication rule for space-managed data to NONE, the rule
for space-managed data is ignored during replication processing. All the
data in the file space is replicated to the target replication server according
to the rule for archive data.
Objects that cannot be replicated
The following objects cannot be replicated to a target replication server:
v Replication rules
v Server node definitions
v Network-attached storage data in nonnative storage pools
v Client schedules
v Backup sets

Tips:
v If you want to convert client nodes for store operations to a target
replication server, you can manually duplicate client schedules that are
on the source replication server.
v You can generate backup sets on the target replication server for a
replicated client node.
Retention protection
You cannot configure servers for replication on which archive retention
protection is enabled.
Replication and file groups
When you are replicating files from one server to another, it is possible that
some of the files that are being replicated belong to a group of files that
are managed as a single logical entity. If a replication process ends without
replicating all the files in a group, client nodes are unable to restore,
retrieve, or recall the file group. When replication runs again, the source
replication server attempts to replicate the missing files.

934 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 Renaming a node
If a node is configured for replication, it cannot be renamed.
Backing up a single client node to two source replication servers
If you have been backing up, archiving, or migrating a client node to two
different servers, do not set up replication of the node from both source
replication servers to the same target replication server. Replicating from
two source servers might create different versions of the same file on the
target server. Replicating from two source servers can cause unpredictable
results when you restore, retrieve, or recall the file.
Password propagation to the target replication server
When client node data is replicated for the first time, the source server
sends the node definition, including the password, to the target replication
server. During subsequent replications, if the node password is updated,
the source server attempts to send the updated password to the target
replication server.
Whether these attempts succeed depends on the node authentication
method and on the combination of methods that are used on the source
and target replication servers. A conflict occurs if a node password is
authenticated on the source server by one server and on the target
replication server by a different server. Because authentication can happen
on an LDAP (Lightweight Directory Access Protocol) directory server or
the Tivoli Storage Manager server, data can be lost. If this kind of dual
authentication occurs, the password is not updated during replication.
Simultaneous-write function
During replication processing, the simultaneous-write function is disabled
on the target replication server when you store data to a primary storage
pool that is enabled for data deduplication. Data that is replicated consists
of only files or extents of data that do not exist on the target replication
server.
Related tasks:
“Writing data simultaneously to primary, copy, and active-data pools” on page 331

Task tips for node replication

Task-tip topics identify the commands that you use for tasks that are related to
node replication. Task-tip topics also provide links to topics with detailed
information about a task.

Change replication rules

Customize a node replication configuration by changing the rules that apply to
individual file spaces, to individual client nodes, and to collectively all client nodes
that are configured for replication.

If you want to change For more information,

replication rules... Use these commands... see...
For individual file UPDATE FILESPACE “Changing replication
spaces. For example, rules for individual file
change an ALL_DATA spaces” on page 952
rule for backup data to
ACTIVE_DATA.

Chapter 30. Replicating client node data 935

 If you want to change For more information,
replication rules... Use these commands... see...
For individual client UPDATE NODE “Changing replication
nodes. For example, rules for individual client
change the replication nodes” on page 954
rule that applies to
backup data in all of the
file spaces that belong to
an individual node.
For all client nodes. For SET BKREPLRULEDEFAULT, SET “Changing server
example, change a ARREPLRULEDEFAULT, and SET replication rules” on page
server replication rule of SPREPLRULEDEFAULT 956
ALL_DATA to NONE.

Add and remove client nodes for replication

Add client nodes for replication processing and remove them from replication
processing.

If you want to... Use these commands... For more information, see...
Add client nodes for REGISTER NODE and UPDATE “Adding client nodes for
replication. NODE replication processing” on
page 962
Remove client nodes from REMOVE REPLNODE “Removing client nodes from
replication. Removing a node replication processing” on
from replication converts a page 963 and “Converting
replicating node to a client nodes for store
non-replicating node. A operations on a target
non-replicating node can replication server” on page
back up, archive, and 989
migrate data directly to the
target replication server.

Manage replication servers

Manage source and target replication servers. For example, change a target
replication server or disable or enable inbound or outbound sessions.
Table 74. Commands for managing replication servers
If you want to... Use these commands... For more information, see...
Set up Secure Sockets Layer DEFINE SERVER and UPDATE “Configuring a server for
(SSL) communications SERVER SSL communications” on
between a source and target page 968
replication server
Change a target replication SET REPLSERVER “Selecting a new target
server replication server” on page
965
Remove a target replication SET REPLSERVER and REMOVE “Removing a replication
server REPLSERVER server” on page 967
Control the number of node REPLICATE NODE “Controlling throughput for
replication sessions node replication” on page
976

936 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 Table 74. Commands for managing replication servers (continued)
If you want to... Use these commands... For more information, see...
Disable or enable inbound or DISABLE SESSIONS and ENABLE “Disabling and enabling
outbound sessions from a SESSIONS outbound or inbound
source or target replication sessions” on page 980
server.
Disable or enable outbound DISABLE REPLICATION and “Disabling and enabling
replication processing from a ENABLE REPLICATION outbound node replication
source replication server processing” on page 981
Remove a replication REMOVE REPLNODE and SET “Removing a node
configuration REPLSERVER replication configuration” on
page 989
Specify, at a server level, the SET FAILOVERHLADDRESS “Configuring the target
IP address that the client replication server” on page
uses to communicate with 966
the failover server, if it is
different from the IP address
that replication uses

Validate a configuration and preview results

Before you replicate client node data, you can confirm that your replication rules
are set up correctly. You can also preview results. For example, you can display a
list of files to be replicated.
Table 75. Commands for validating a replication configuration and previewing results
If you want to... Follow these steps... For more information, see...
Determine whether server, VALIDATE REPLICATION “Validating a replication
client node, and file-space configuration” on page 970
replication rules are set up
correctly before you replicate
data. You can also verify the
connection between the
source and target replication
servers.
Preview replication results. REPLICATE NODE “Previewing node replication
results” on page 970
Display information about QUERY REPLSERVER “Removing a replication
the replicating server. server” on page 967

Manage replication processing

Manage replication processing. For example, start a replication process or disable a
replication rule.
Table 76. Commands for managing replication processing
If you want to... Use these commands... For more information, see...
Replicate data. You can REPLICATE NODE and DEFINE “Replicating data by
replicate data by individual SCHEDULE command” on page 971
file space, by priority, and by
data type.
Temporarily disable UPDATE FILESPACE “Disabling and enabling
replication for a data type in replication of data types in a
a file space. file space” on page 978

Chapter 30. Replicating client node data 937

 Table 76. Commands for managing replication processing (continued)
If you want to... Use these commands... For more information, see...
Temporarily disable UPDATE NODE “Disabling and enabling
replication for an individual replication for individual
client node. client nodes” on page 979
Temporarily disable UPDATE REPLRULE “Disabling and enabling
replication of data that is replication rules” on page
assigned a particular 982
replication rule.
Temporarily disable inbound DISABLE SESSIONS and ENABLE “Disabling and enabling
and outbound server SESSIONS outbound or inbound
sessions, including sessions” on page 980
replication sessions for all
client nodes.
Temporarily disable DISABLE REPLICATION and “Disabling and enabling
outbound replication ENABLE REPLICATION outbound node replication
processing from a source processing” on page 981
replication server.
Prevent replication of UPDATE FILESPACE “Purging replicated data in a
backup, archive, or file space” on page 983
space-managed data in a file
space on a source replication
server, and delete the data
from the target replication
server.
Cancel all replication CANCEL REPLICATION “Canceling replication
processes. processes” on page 985

Monitor replication processing and verify results

Preview results before replicating data, monitor replication processing, and verify
results after replication ends.
Table 77. Commands for monitoring replication processing
If you want to... Use these commands... For more information, see...
Specify the number of days SET REPLRETENTION “Retaining replication
to retain replication records records” on page 988
in the Tivoli Storage
Manager database.
Display information about QUERY FILESPACE “Displaying information
the replication settings for a about node replication
file space. settings for file spaces” on
page 985
Display information about QUERY NODE “Displaying information
the replication settings for a about node replication
client node. settings for client nodes” on
page 985
Display information about QUERY REPLRULE “Displaying information
replication rules. about node replication rules”
on page 985
Display records of running QUERY REPLICATION “Displaying information
and ended replication about node replication
processes. processes” on page 985

938 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 Table 77. Commands for monitoring replication processing (continued)
If you want to... Use these commands... For more information, see...
Determine whether QUERY REPLNODE “Measuring the effectiveness
replication to the target of a replication
replication server is keeping configuration” on page 986
pace with the number of files
that are eligible for
replication on the source
replication server.
Measure the effects of data QUERY REPLICATION “Measuring the effects of
deduplication. data deduplication on node
replication processing” on
page 987

Planning for node replication

Use a checklist of planning considerations to streamline implementation. Planning
considerations include items such as the types of data that you want to replicate
and the replication rules that you want to apply.

About this task

As you plan, remember that a target replication server must be accessible from a
source replication server by using an IP connection. The connection must provide
sufficient bandwidth to accommodate the volume of data to be replicated. If the
connection is insufficient and becomes a bottleneck for replication, keeping the
data on the two servers synchronized can be a problem. Keep in mind that you
can use client-side data deduplication with node replication to reduce network
bandwidth requirements and storage requirements.

The destination storage pool on a target replication server must have sufficient
space to store replicated data.

Include the following steps in your planning process:

Procedure
v Identify the client nodes and, if necessary, the files spaces that belong to those
nodes, that have data to be replicated.
v Identify the types of data that you want to replicate. With Tivoli Storage
Manager, you can replicate backup data, archive data, and space-managed data.
If you decide to replicate backup data, you must decide whether to replicate
only active backup data or to replicate active and inactive backup data. In
environments with large amounts of backup data, it might make sense to
replicate only active backup data. The type of data that you replicate can vary
by node or by file space.
v Determine the order in which data is to be replicated. If you want to replicate
some data ahead of other data in a replication process, you can assign a
high-priority replication rule to the data that you want to replicate first. A
high-priority classification does not increase the speed of replication.
v Set up a default replication configuration. In the default configuration, all
backup, archive, and space-managed data is replicated in all file spaces in all
replication-enabled client nodes. You can also change the replication rules at the
file space, client node, and server levels. In general, you can apply different
replication rules at different levels and achieve the same results.

Chapter 30. Replicating client node data 939

 v Identify source and target replication servers. If you have multiple source
replication servers, consider whether you want to replicate data to one or more
target replication servers. A single target replication server can have multiple
source replication servers. However, a source replication server can replicate
data to only one target replication server at a time.
v Consider using the enterprise configuration functions of Tivoli Storage Manager
to set up, manage, and synchronize policy configurations on the source and the
target replication servers. For example, you can change a policy configuration on
a source replication server and have the changes that are automatically
distributed to the target replication server.
v Consider how you want to replicate data. You can replicate one or more client
nodes or defined groups of client nodes in a single process. You can schedule
replication processes at different times to control network throughput.
v Decide whether you want to implement replication gradually or at one time. You
can start with a simple configuration that comprises a single source replication
server and a single target replication server. After you experiment with that
configuration, you can build a broader network of replicating servers.
v If you use an LDAP directory server to authenticate passwords, any target
replication servers must be configured for LDAP passwords. Data that is
replicated from a node that authenticates with an LDAP directory server is
inaccessible if the target replication server is not configured. If your target
replication server is not configured, replicated data from an LDAP node can still
be written to the replication server. However, the target replication server must
be configured to use an LDAP directory so that you can access the data.
v Ensure that you allocate extra space for the active log and the archive log during
replication processing.

Determining server database requirements for node

replication
Node replication requires more Tivoli Storage Manager database space to track the
files that are replicated.

About this task

To determine whether the database can manage more space requirements, you
must estimate how much more database space that node replication will use.

Requirement: Place the database and database logs on separate disks that have a
high performance capability. Use a separate disk or mount point for the following
options:
v Other applications that use the database and logs
v System tasks, such as system paging

Procedure
1. Determine number of files for each node and data type that is in use. Issue the
QUERY OCCUPANCY command for each node and data type that you plan to
replicate. For example, you can display information about the file spaces that
are assigned to the node named PAYROLL by issuing the following command:
query occupancy payroll
2. Determine how much more database space is required by using the value for
the total number of files that are used by all nodes and data types. Use the
following formula to calculate the amount of database space that is required:

940 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 Total_number_of_files_from_all_nodes_and_data_types * 300 (the number of
additional bytes needed for each replicated file)

Important: You must increase the available database space when the additional
required space approaches or exceeds the size of your database. Ensure that
you examine both replication servers and their databases and increase the
database size if necessary.
3. Increase the size of the database by the additional database space required and
include an additional 10% of the database size.

Estimating the total amount of data to be replicated

Determine the amount of data that is replicated initially and incrementally daily.
Initial replications typically take longer to complete than incremental daily
replications.

Procedure
1. Estimate the initial amount of data to be replicated. Decide which nodes, file
spaces, and data types must be replicated. To calculate the total amount of
physical space that is occupied for each file space and data type that must be
replicated, issue the QUERY OCCUPANCY command for each node. For example,
you can display information about the file spaces assigned to the node named
PAYROLL by issuing the following command:
query occupancy payroll

Tip: Tune the performance of replication to the data type. For example, if you
do not plan to replicate a data type in a file space, exclude the number of files
for that data type.
2. Determine the amount of data that is backed up daily by the client nodes.
Complete the following steps to estimate the amount of data that is replicated
incrementally daily:
a. When client nodes complete a store operation, the client logs completion
messages with the server. The completion messages report statistics or
information for a client operation that was transferred to the server.
Message ANE4961I shows the number of bytes that are transferred during a
client operation.
b. To determine the average amount of bytes that are backed up daily by a
particular node, add the number of bytes that are shown in the ANE4961I
messages over several days. Then calculate the average amount of bytes to
be replicated for all nodes to determine how much data is replicated daily.

Estimating network bandwidth required for replication

To ensure that the initial replication is successful, determine whether you have the
network bandwidth, processor availability, and the time for replication. Estimating
network bandwidth is based on the amount of data that must be replicated.

Procedure
1. To calculate the required network bandwidth, determine the following
information:
v Total amount of data (TD) to be replicated, in gigabytes. To calculate the total
amount of data, see “Estimating the total amount of data to be replicated.”
v Length of replication window time (RWT) in hours. The length of the
replication window time is the time that you schedule during server
maintenance for replication to occur.

Chapter 30. Replicating client node data 941

 v Data deduplication ratio (DR), if you are using the data deduplication
function. Issue the QUERY STGPOOL command and use the FORMAT=DETAIL
value of to determine the data deduplication ratio value. If you are not using
the data deduplication function, use 100 as the data deduplication ratio
value.
2. Calculate the bandwidth required by using the following formula:
(TD * (100 / DR) * 8192) / (RWT * 3600) = Required_Network_Bandwidth
(Mbps/second)

What to do next

If the value for required network bandwidth exceeds the capabilities of your
network, you must adjust the values in the formula. Reduce the TD value or
increase the replication time, to reduce the value for Required_Network_Bandwidth. If
you cannot adjust the TD or the RWT time values, adjust or replace your existing
network to reduce the additional workload.

Calculating the time that is required for replication

You can estimate how many hours it will take to replicate data. You must know
the amount of data you want to replicate and the bytes per hour that the network
can process.

Procedure
1. Determine the time that is required for replication by using the number of
bytes you want to replicate and the bytes per hour value. Use the following
formula to calculate how many hours it takes to complete replication:
Total_bytes_to_be_Replicated / Bytes_Per_Hour = Hours_to_Complete_Replication.
To calculate the Total_bytes_to_be_Replicated and the Bytes_Per_Hour values,
complete the steps in “Tuning replication processing” on page 977.
2. For the initial replication, determine how many hours it takes for the
replication to occur over the network during a daily window for replication by
using the following calculation:
Days_to_Complete_Replication = Hours_to_Complete_Replication / 24

Requirement: For daily incremental replications, the

Hours_to_Complete_Replication value must be larger than the amount of time you
schedule for replication.

What to do next

When you determine how long it takes for replication to finish, you can decide
which method you use to complete the initial replication. The method that you use
for the initial replication is based on the data, time, and bandwidth values that you
calculated.
Related tasks:
“Selecting a method for the initial replication” on page 943

942 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Selecting a method for the initial replication
Based on the test replication results and the total amount of data that you want to
replicate, determine a method to use for the initial replication.
Related tasks:
“Calculating the time that is required for replication” on page 942
“Tuning replication processing” on page 977

Method 1: Exporting and importing data by using node

replication synchronization
Use this method if you have a large amount of data to replicate and you cannot
wait for initial replication to complete. During replication, the data that is exported
from the source replication server and imported to the target replication server is
synchronized, and the data is displayed as if it is replicated.

Procedure
1. Export the data from the source replication server for the nodes that you want
to replicate. For example, to export client node information and all client files
for NODE1 directly to SERVERB, issue the following command:
export node node1 filedata=all toserver=serverb

Tip: You can also export client data directly to another server so that it can be
immediately imported. For example, to export client node information and all
client files for NODE1 directly to SERVERB, issue the following command:
export node node1 filedata=all toserver=serverb

The data is automatically imported when it is exported to another server.

2. Import the data to the target replication server. For example, to import client
node information from tape volumes TAPE01, TAPE02, and TAPE03, issue the
following command:
import node volumenames=tape01,tape02,tape03 dates=absolute

Requirement: Ensure that you use the DATES=ABSOLUTE parameter on the

IMPORT NODE command when you import the data to the target replication
server. The ABSOLUTE value sets the insertion dates for the objects on the
target replication server to be the same as the insertion dates on the source
replication server. When synchronization occurs, the servers use the insertion
dates to identify files that are the same and updates them as replicated files.
3. Configure the nodes on the source and target replication servers. Issue the
UPDATE NODE command and specify the following parameters:
v On the source replication server: REPLMODE=SYNCSEND
v On the target replication server: REPLMODE=SYNCRECEIVE
4. When replication is started, the nodes are configured for synchronization and
no data is transferred to the server. The source and target replication servers
exchange information about the data and add entries to the database so that
the data is managed as it is replicated. After the data synchronization is
complete, the nodes are automatically configured by the Tivoli Storage Manager
server to replicate data daily. When synchronization is complete on the source
replication server, the parameter is automatically set to REPLMODE=SEND. When
synchronization is complete on the target replication server, the parameter is
automatically set to REPLMODE=RECEIVE.

Chapter 30. Replicating client node data 943

 Method 2: Replicating subsets of nodes
Use this method to replicate subsets of nodes incrementally, based on specific
characteristics of the data of the nodes. You can use any characteristic to create
subsets of nodes. You can use this method if you have time to complete initial
replication but the network cannot manage replicating all nodes at one time.

About this task

When you decide how many nodes to add to a group, consider the amount of data
that is replicated daily by the nodes.

Procedure
1. Prioritize the subset of nodes that have critical data. Replicate critical data first,
by issuing the REPLICATE NODE command.
2. Continue to replicate the high-priority nodes daily while incrementally adding
the replication of other subsets of nodes that contain important, but not critical,
data.
3. Repeat this process until all subsets of all nodes that must be replicated
complete their initial replication.
Related concepts:
“Node replication processing” on page 923
“Replication rules” on page 923

Method 3: Replicating active data first

Use this method to set up an initial replication of only active data to avoid
overloading the network or servers if you replicate both active and inactive data.
Active versions of backup data are more important than the inactive versions.

Procedure
1. Configure the replication rules to replicate active data.
2. Replicate the active data. For example, issue the following UPDATE NODE
command to replicate active data for a node named PAYROLL:
update node payroll bkreplrule=active_data
3. After the initial replication of the active data is completed, configure the
replication rules to replicate all versions of the data by issuing the following
UPDATE NODE command:
update node payroll bkreplrule=active_data_high_priority

Results

During the next scheduled replication, any new active versions, including all
inactive versions, are replicated. The files that were active but are now inactive are
not replicated again.

944 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 Method 4: Replicating all nodes
Use this method if you determine you have time for the initial replication process
to complete, and you can configure the nodes and begin or schedule a replication.

Procedure
1. Configure all nodes that must be replicated by issuing the UPDATE NODE
command and specify the REPLSTATE=ENABLED parameter.
2. Issue the REPLICATE NODE command to start replication processing.
3. Monitor the progress of the replication by issuing the QUERY PROCESS command.
Summary information is displayed, which includes the amount of data that is
replicated and the process duration.

Remember: If you do not have time to complete replication, you can cancel it
after it has started, by issuing the CANCEL REPLICATION command.
4. Use the summary information to determine whether the values of the
controlled test match the actual replication values. You calculate the values of
the controlled test in “Tuning replication processing” on page 977. For example,
to display information about a replication process 23, issue the following
command:
query process 23

The following summary information is displayed:

Process Process Status
Number Description
-------- -------------- ---------
23 REPLICATE NODE Replicating Node(s) IRONMAN.
File spaces complete: 0.
File spaces examining and replicating: 1.
File spaces replicating: 0.
File spaces not started: 3.
Files current: 11,920.
Files replicated: 0 of 0.
Files updated: 0 of 0.
Files deleted: 0 of 0.
Bytes Replicated: 0 of 0.
Elapsed Time: 0 Day(s),
0 Hour(s), 1 Minute(s).

Scheduling incremental replication after the initial replication

After you complete the initial replication, schedule subsequent incremental
replications frequently. Scheduling replication frequently ensures that the data on
the target replication server is maintained at an acceptable recovery point that you
specify in your data recovery plan. Daily incremental replications typically do not
require as much time to complete as the initial replication.

About this task

If you are unable to complete the replication process in the amount of time that
you scheduled, increase the number of data sessions that transfer data to the target
replication server. Replication performance improves when more deduplicated data
is stored on the target replication server. When more extents are stored on the
target replication server, more duplicates are found for an extent.

Chapter 30. Replicating client node data 945

 Procedure

If you are replicating data from storage pools that are enabled for data
deduplication, run processes in the following order:
1. To identify duplicates, issue the IDENTIFY DUPLICATES command. Break files
into extents to reduce the amount of data that is sent to the target replication
server when replication occurs.
2. To start replication processing, issue the REPLICATE NODE command. Only file
extents that do not exist on the target replication server are sent during
replication, which reduces the required bandwidth and improves performance.
3. To start reclamation processing, issue the RECLAIM STGPOOL command.
Reclamation removes and links duplicated extents.

Setting up the default replication configuration

The default configuration replicates all backup, archive, and space-managed data
in all the file spaces in all the replication-enabled nodes.

About this task

The following figure shows the replication rules that are created in the default
configuration. Backup data includes both active and inactive backup data.

946 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 File space A File space B File space A

BACKUP DATA RULE = BACKUP DATA RULE = BACKUP DATA RULE =

DEFAULT DEFAULT DEFAULT
ARCHIVE DATA RULE = ARCHIVE DATA RULE = ARCHIVE DATA RULE =
DEFAULT DEFAULT DEFAULT
SPACE-MANAGED DATA RULE = SPACE-MANAGED DATA RULE = SPACE-MANAGED DATA RULE =
DEFAULT DEFAULT DEFAULT

BACKUP DATA RULE = BACKUP DATA RULE =

DEFAULT DEFAULT
ARCHIVE DATA RULE = ARCHIVE DATA RULE =
DEFAULT DEFAULT
SPACE-MANAGED DATA RULE = SPACE-MANAGED DATA RULE =
DEFAULT Client node 2 DEFAULT
Client node 1

Source replication server

BACKUP DATA RULE=
ALL_DATA
ARCHIVE DATA RULE =
ALL_DATA
SPACE-MANAGED DATA RULE =
ALL_DATA

REPLICATE NODE command

Replication process
Normal priority data

File space A, Client 1: File space B, Client 1: File space A, Client 2:

BACKUP DATA BACKUP DATA BACKUP DATA
ARCHIVE DATA ARCHIVE DATA ARCHIVE DATA
SPACE-MANAGED DATA SPACE-MANAGED DATA SPACE-MANAGED DATA

Target replication
server

To set up the default replication scenario, complete the following tasks:

“Step 1: Setting up server-to-server communications” on page 948
“Step 2: Specifying a target replication server” on page 950
“Step 3: Configuring client nodes for replication” on page 950

After you complete the default configuration, you can change the replication rules
to meet your specific replication requirements.

Chapter 30. Replicating client node data 947

 Step 1: Setting up server-to-server communications
Define the source and the target replication servers to each other. When you define
servers, you can also specify Secure Sockets Layer (SSL) communications.

About this task

Server definitions are required for the source replication server to communicate
with the target replication server and for the target replication server to report
status to the source replication server.

Before you begin this procedure, complete the following prerequisites:

v Ensure that the target replication server is accessible from the source replication
server with an IP connection.
v Ensure that there is enough network bandwidth to accommodate the volume of
data to be replicated. If bandwidth is insufficient and the flow of data is
impeded, the ability to keep the two servers synchronized is affected.
v If you plan to use SSL communication, the SSL infrastructure must be in place
on the source and target replication servers. Required SSL certificates must be in
the key database file that belongs to each server.

Important: You can specify only one target replication server for a source
replication server. However, you can specify one or more source replication servers
for a single target replication server. Source and target replication servers must be
V6.3 or later.

The method that you use to set up servers depends on whether the server
definitions exist and on whether you are using the cross-define function to
automatically define one server to another.

Procedure

To set up server-to-server communications, complete one of the following actions:

v If server definitions do not exist and you want to cross-define servers, complete
the following steps:
1. On the target replication server, issue the following commands:
set servername target_server_name
set serverpassword target_server_password
set serverhladdress _target_server_ip_address
set serverlladdress target_server_tcp_port
set crossdefine on

Remember: If you want an SSL connection, the value for the SET
SERVERLLADDRESS command on the target replication server must be an SSL
port. The value of the SET SERVERNAME command must match the server
name in the server definition.
2. On the source replication server, issue the following commands:
set servername source_server_name
set serverpassword source_server_password
set serverhladdress source_server_ip_address
set serverlladdress source_server_tcp_port

Remember: If you want an SSL connection, the value for the SET
SERVERLLADDRESS command on the source replication server must be an SSL
port. The value of the SET SERVERNAME command must match the server
name in the server definition.

948 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 3. On the source replication server connect to the target replication server by
using the DEFINE SERVER command. If you want an SSL connection, specify
SSL=YES. For example:
define server target_server_name hladdress=target_sever_ip_address
lladdress=target_server_tcp_port serverpassword=
target_server_password crossdefine=yes ssl=yes

A server definition is created on the source replication server, and the source
replication server is connected to the target replication server. A definition for
the target replication server is created that points to the source replication
server.
v If server definitions do not exist and you are not using the cross-define function,
complete the following steps:
1. Issue the following commands on both the source and target replication
servers:
set servername server_name
set serverpassword server_password
set serverhladdress ip_address
set serverlladdress tcp_port

Remember: If you want an SSL connection, the value for the SET
SERVERLLADDRESS command on the replication servers must be an SSL port.
The value of the SET SERVERNAME command must match the server name in
the server definition.
2. Issue the DEFINE SERVER command on each server. Do not specify the
CROSSDEFINE parameter. If you want an SSL connection, specify SSL=YES. For
example:
– On the source replication server:
define server target_server_name hladdress=target_server_ip_address
lladdress=target_server_tcp_port serverpassword=target_server_password
ssl=yes
– On the target replication server:
define server source_server_name hladdress=source_server_ip_address
lladdress=source_server_tcp_port serverpassword=source_server_password
ssl=yes
v If definitions exist for both the source and target replication servers, issue the
UPDATE SERVER command on each server. Do not specify the CROSSDEFINE
parameter. You can use the QUERY STATUS command to determine the server
names. If you want an SSL connection, specify SSL=YES. For example:
– On the source replication server:
update server target_server_name hla=target_server_ip_address
lladdress=target_server_tcp_port serverpassword=target_server_password
ssl=yes
– On the target replication server:
update server source_server_name hladdress=source_server_ip_address
lladdress=source_server_tcp_port serverpassword=
source_server_password
ssl=yes

Chapter 30. Replicating client node data 949

 Step 2: Specifying a target replication server
For each source replication server, you can specify only a single target replication
server. Multiple source replication servers can share the same target replication
server.

Before you begin

Before you begin this procedure, issue the PING SERVER command. The command
verifies that the definitions for the source and target replication servers are valid
and that the servers are connected.

Procedure

To specify a target replication server, issue the SET REPLSERVER command on the
source replication server. For example, to specify a server named PHOENIX_SRV
as the target replication server, issue the following command:
set replserver phoenix_srv

Results

Issuing the SET REPLSERVER command also sets replication rules to ALL_DATA. To
display replication rules, you can issue the QUERY STATUS command.
Related concepts:
“Replication server configurations” on page 921

Step 3: Configuring client nodes for replication

When a node is configured for replication for the first time, it cannot exist on the
target replication server unless the data of the node was exported from the source
replication server and imported on the target replication server.

Before you begin

Before you begin this procedure:

v Ensure that the source replication server and target replication server are set up
for server-to-server communications and that they can communicate with each
other.
v Ensure that a target replication server was specified.
v If the data for a client node was exported from a source replication server and
imported to the target replication server, you must synchronize the data. To
synchronize data, the data must be imported with the value of the DATES
parameter on the IMPORT NODE command that is set to ABSOLUTE. The source
replication server uses the file names and the insertion dates that files were
imported to identify identical files.

Restrictions:
v If a client node definition does not exist on the target replication server, do not
create it. The definition for the client node on the target replication server is
created automatically when the data of the node is replicated the first time.
v If a client node definition exists on both the source and target replication servers,
but the data that belongs to the client node was not exported and imported, you
must rename or remove the client node on the target replication server before
data can be replicated.

950 IBM Tivoli Storage Manager for Windows: Administrator's Guide

v If you previously removed a client node from replication on the source
replication server, but not on the target replication server, you do not have to
rename or remove the node on the target replication server.

About this task

To configure a client node for replication, take one of the following actions,
depending on whether data of the node was exported from the source replication
server and imported on the target replication server:

Procedure
v If the data of the node was not exported from the source replication server and
imported on the target replication server, complete one of the following steps:
– If the client node is not already registered on a source replication server, issue
the REGISTER NODE command on the source replication server. Specify
REPLSTATE=ENABLED or REPLSTATE=DISABLED.
For example, to enable a new client node, NODE1, for replication, issue the
following command:
register node node1 password replstate=enabled
– If the client node is already registered on a source replication server, issue the
UPDATE NODE command on the source replication server. Specify
REPLSTATE=ENABLED or REPLSTATE=DISABLED.
For example, to enable an existing client node, NODE1, for replication, issue
the following command:
update node node1 replstate=enabled
v If the data of the node was exported from the source replication server and
imported to the target replication server, complete the following steps:
1. On the source replication server, issue the UPDATE NODE command:
a. Specify REPLSTATE=ENABLED or REPLSTATE=DISABLED.
b. Specify REPLMODE=SYNCSEND.
2. On the target replication server, issue the UPDATE NODE command and specify
REPLMODE=SYNCRECEIVE.
Data is synchronized during replication. After the replication process is
complete, the REPLMODE parameter in the client node definition on the source
replication server is set to SEND. The REPLMODE parameter in the client node
definition on the target replication server is set to RECEIVE, and the REPLSTATE
parameter is set to ENABLED.

Results

If you set the replication state of the client node to DISABLED, the replication
mode is set to SEND, but replication does not occur. If you set the replication state
of the client node to ENABLED, the client node definition is created on the target
replication server when replication occurs for the first time. In addition, the
replication mode of the client node on the target replication server is set to
RECEIVE, and the replication state is set to ENABLED.

If you add a file space to a client node that is configured for replication, the file
space replication rules for data types are automatically set to DEFAULT. To change
file space replication rules, issue the UPDATE FILESPACE command.

To determine the replication mode and the replication state that a client node is in,
issue the QUERY NODE command.

Chapter 30. Replicating client node data 951

 What to do next

The default configuration is complete after client nodes are configured for
replication. You are now ready to replicate. If you do not change the default
replication rules, all backup, archive, and space-managed data in all
replication-enabled client nodes is replicated.
Related concepts:
“Replication mode” on page 930
“Replication state” on page 927

Customizing a node replication configuration

With Tivoli Storage Manager, you can change the replication configuration,
including the type and priority of data that is replicated.

Changing replication rules

You can change replication rules that apply to data types in individual file spaces,
to individual client nodes, and collectively to all client nodes that are configured
for replication.

Changing replication rules for individual file spaces

When replication processing begins, the source server first checks the replication
rules that apply to data types in file spaces. File space rules take precedence over
client node rules and server rules.

About this task

Rules for file spaces are either normal priority or high priority. In a replication
process that includes both normal-priority and high-priority data, high-priority
data is replicated first. If you issue the REPLICATE NODE command for two or more
clients, all high priority data for all file spaces in the specified nodes is processed
before normal priority data.

Before you select a rule, consider the order in which you want the data to be
replicated. For example, suppose that a file space contains active backup data and
archive data. Replication of the active backup data is a higher priority than the
archive data. To prioritize the active backup data, specify DATATYPE=BACKUP
REPLRULE=ACTIVE_DATA_HIGH_PRIORITY. To prioritize the archive data, issue the
UPDATE FILESPACE command again, and specify DATATYPE=ARCHIVE
REPLRULE=ALL_DATA.

You can specify the following rules:

ALL_DATA
Replicates backup, archive, or space-managed data. The data is replicated
with a normal priority.
ACTIVE_DATA
Replicates only the active backup data. The data is replicated with a
normal priority.

952 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 Attention:
v If you specify ACTIVE_DATA, inactive backup data in the file space is
not replicated, and inactive backup data in the file space on the target
replication server is deleted.
v If you specify ACTIVE_DATA, you cannot specify ARCHIVE or
SPACEMANAGED as values for the parameter DATATYPE in the same
command instance.
ALL_DATA_HIGH_PRIORITY
Replicates backup, archive, or space-managed data. The data is replicated
with a high priority.
ACTIVE_DATA_HIGH_PRIORITY
Replicates active backup data. The data is replicated with a high priority.

Attention:
v If you specify ACTIVE_DATA_HIGH_PRIORITY, inactive backup data in
the file space is not replicated, and inactive backup data in the file space
on the target replication server is deleted.
v If you specify ACTIVE_DATA_HIGH_PRIORITY, you cannot specify
ARCHIVE or SPACEMANAGED as values for the parameter DATATYPE in
the same command instance.
DEFAULT
Replicates data according to the client node rule for the data type.
For example, suppose that you want to replicate the archive data in all the
file spaces that belong to a client node. Replication of the archive data is a
high priority. One method to accomplish this task is to specify
DATATYPE=ARCHIVE REPLRULE=DEFAULT for each file space. Ensure that the
client replication rule for archive data is set to
ALL_DATA_HIGH_PRIORITY or to DEFAULT. If the client replication rule
is DEFAULT, the server replication rule for archive data must be set to
ALL_DATA_HIGH_PRIORITY.
NONE
Data is not replicated.
For example, if you do not want to replicate the space-managed data in a
file space, specify NONE.

To display the replication rules for a file space, issue the QUERY FILESPACE
command. Specify FORMAT=DETAILED.

To change file-space replication rules, issue one or more of the following

commands on the source replication server:

Procedure
v To change a replication rule for backup data, issue the UPDATE FILESPACE
command and specify DATATYPE=BACKUP. Specify the REPLRULE parameter. For
example, to specify the ACTIVE_DATA rule for backup data in file space /a on
NODE1, issue the following command:
update filespace node1 /a datatype=backup replrule=active_data
v To specify a replication rule for archive data, issue the UPDATE FILESPACE
command and specify DATATYPE=ARCHIVE. Specify the REPLRULE parameter. For
example, to specify the ALL_DATA_HIGH_PRIORITY rule for archive data in
file space /a on NODE1, issue the following command:

Chapter 30. Replicating client node data 953

 update filespace node1 /a datatype=archive replrule=all_data_high_priority
v To specify a replication rule for space-managed data, issue the UPDATE FILESPACE
command and specify DATATYPE=SPACEMANAGED. Specify the REPLRULE parameter.
For example, to specify the NONE rule for file space /a on NODE1, issue the
following command:
update filespace node1 /a datatype=spacemanaged replrule=none

Example

In the following example, assume that you have two client nodes, NODE1 and
NODE2. The nodes have the following file spaces:
v NODE1: /a, /b, /c
v NODE2: /a, /b, /c, /d, /e

All the file space rules are set to DEFAULT. The backup, archive, and
space-managed rules for NODE1 and NODE2 are also set to DEFAULT. The server
rules are set to ALL_DATA. To replicate the backup data in file space /a in NODE1
and file space /c in NODE2 first, before replicating data in the other file spaces,
issue the following commands:
update filespace node1 /a datatype=backup replrule=all_data_high_priority
update filespace node2 /c datatype=backup replrule=all_data_high_priority

The data that belongs to the two nodes is replicated in the following order:
1. High Priority: Data in file space /a that belongs to NODE1 and data in file
space /c in NODE2
2. Normal priority: Data in file spaces /b and /c that belongs to NODE1 and data
in file spaces /a, /b, /d, and /e that belongs to NODE2

What to do next

Important: Data types in new file spaces that are added to a client node after the
node is configured for replication are automatically assigned the DEFAULT
replication rule.
Related concepts:
“Replication rules” on page 923

Changing replication rules for individual client nodes

Client-node replication rules apply to every file space that belongs to a client node.
To replicate data according to a client node rule for a data type, the replication rule
that applies to the data type in a file space must be set to DEFAULT.

About this task

Rules for client nodes are either normal priority or high priority. In a replication
process that includes both normal-priority and high-priority data, high-priority
data is replicated first. If you issue the REPLICATE NODE command for two or more
clients, all high priority data for all file spaces in the specified nodes is processed
before normal priority data.

Before you select a rule, consider the order in which you want the data to be
replicated. For example, suppose that a client node contains active backup data
and archive data. Replication of the active backup data is a higher priority than

954 IBM Tivoli Storage Manager for Windows: Administrator's Guide

replication of the archive data. To prioritize the active backup data, specify the
ACTIVE_DATA_HIGH_PRIORITY replication rule for backup data. Specify the
ALL_DATA rule for archive data.

You can specify the following rules:

ALL_DATA
Replicates backup, archive, or space-managed data. The data is replicated
with a normal priority.
ACTIVE_DATA
Replicates only the active backup data. The data is replicated with a
normal priority.

Attention:
v If you specify ACTIVE_DATA, inactive backup data that belongs to the
client node is not replicated.
v If the replication rule for backup data in any file spaces that belong to
the client node is DEFAULT, inactive backup data in those file spaces on
the target replication server is deleted.
ALL_DATA_HIGH_PRIORITY
Replicates backup, archive, or space-managed data. The data is replicated
with a high priority.
ACTIVE_DATA_HIGH_PRIORITY
Replicates only the active backup data in client nodes. The data is
replicated with a high priority.

Attention:
v If you specify ACTIVE_DATA_HIGH_PRIORITY, inactive backup data
that belongs to the client node is not replicated.
v If the replication rule for backup data in any file spaces that belong to
the client node is DEFAULT, inactive backup data in those file spaces on
the target replication server is deleted.
DEFAULT
Replicates data according to the server rule for the data type.
For example, suppose that you want to replicate the archive data in all
client nodes that are configured for replication. Replication of the archive
data is a high priority. One method to accomplish this task is to set the
file-space and client-node replication rules for archive data to DEFAULT.
Set the server rule for archive data to ALL_DATA_HIGH_PRIORITY.
NONE
Data is not replicated. For example, if you do not want to replicate the
space-managed data in a client node, specify the NONE replication rule for
space-managed data.

To display the replication rules that apply to all file spaces that belong to a node,
issue the QUERY NODE command and specify FORMAT=DETAILED.

Remember: File spaces are not displayed for client nodes that are registered on the
source replication server but that have not completed store operations. File spaces
are created only after the client stores data to the source replication server.
Replication rules for data types in file spaces are automatically assigned values of
DEFAULT.

Chapter 30. Replicating client node data 955

 To change replication rules for a node, issue one or more of the following
commands on the source replication server:

Procedure
v To change a replication rule for backup data, issue the UPDATE NODE command
and specify the BKREPLRULEDEFAULT parameter. For example, to specify the
ACTIVE_DATA rule for backup data in NODE1, issue the following command:
update node node1 bkreplruledefault=active_data
v To change a replication rule for archive data, issue the UPDATE NODE command
and specify the ARREPLRULEDEFAULT parameter. For example, to specify the
ALL_DATA_HIGH_PRIORITY rule for archive data in NODE1, issue the
following command:
update node node1 arreplruledefault=all_data_high_priority
v To change a replication rule for space-managed data, issue the UPDATE NODE
command and specify the SPREPLRULEDEFAULT parameter. For example, to specify
the NONE rule for space-managed data in NODE1, issue the following
command:
update node node1 spreplruledefault=none
Related concepts:
“Replication rules” on page 923

Changing server replication rules

Server replication rules apply to all client nodes that are defined to a source
replication server and that are configured for replication. To replicate data
according to a server rule for a data type, the replication rule that applies to the
data type in client nodes and file spaces must be set to DEFAULT.

About this task

Server rules are either normal priority or high priority. In a replication process that
includes both normal-priority and high-priority data, high-priority data is
replicated first. If you issue the REPLICATE NODE command for two or more clients,
all high priority data for all file spaces in the specified nodes is processed before
normal priority data.

Before you select a rule, consider the order in which you want the data to be
replicated. For example, suppose that your client nodes contain active backup data
and archive data. Replication of the active backup data is a high priority. To
prioritize the active backup data, specify the ACTIVE_DATA_HIGH_PRIORITY
replication rule. Specify the ALL_DATA rule for archive data.

You can specify the following rules:

ALL_DATA
Replicates backup, archive, or space-managed data. The data is replicated
with a normal priority.
ACTIVE_DATA
Replicates only the active backup data. The data is replicated with a
normal priority.

956 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 Attention:
v If you specify ACTIVE_DATA, inactive backup data that belongs to
client nodes is not replicated.
v If the replication rules for backup data in any file spaces and any client
nodes is DEFAULT, inactive backup data in those file spaces on the
target replication server is deleted. For example, suppose the rules for
backup data in file space /a in NODE1 and file space /c in NODE2 are
DEFAULT. The rules for backup data in NODE1 and NODE2 are also
DEFAULT. If you specify ACTIVE_DATA as the server rule, inactive data
in file spaces /a and /c is deleted.
ALL_DATA_HIGH_PRIORITY
Replicates backup, archive, or space-managed data. The data is replicated
with a high priority.
ACTIVE_DATA_HIGH_PRIORITY
Replicates only the active backup data in client nodes. The data is
replicated with a high priority.

Attention:
v If you specify ACTIVE_DATA_HIGH_PRIORITY, inactive backup data
that belongs to client nodes is not replicated.
v If the replication rules for backup data in any file spaces and any client
nodes is DEFAULT, inactive backup data in those file spaces on the
target replication server is deleted. For example, suppose the rules for
backup data in file space /a in NODE1 and file space /c in NODE2 are
DEFAULT. The rules for backup data in NODE1 and NODE2 are also
DEFAULT. If you specify ACTIVE_DATA as the server rule, inactive data
in file spaces /a and /c is deleted.
NONE
Data is not replicated. For example, if you do not want to replicate the
space-managed data in your client nodes, specify NONE.

Restriction: The replication rule DEFAULT is not a valid server rule.

To change server replication rules, issue one or more of the following commands
on the source replication server:

Procedure
v To change the server replication rule that applies to backup data, issue the SET
BKREPLRULEDEFAULT command on the source replication server. For example, to
specify the ACTIVE_DATA rule for backup data, issue the following command:
set bkreplruledefault active_data
v To change the server replication rule that applies to archive data, issue the SET
ARREPLRULEDEFAULT command on the source replication server. For example, to
specify the ALL_DATA_HIGH_PRIORITY rule for archive data, issue the
following command:
set arreplruledefault all_data_high_priority
v To change the server replication rule that applies to space-managed data, issue
the SET SPREPLRULEDEFAULT command on the source replication server. For
example, to specify the NONE rule for space-managed data, issue the following
command:
set spreplruledefault none
Related concepts:

Chapter 30. Replicating client node data 957

 “Replication rules” on page 923

Scenario: Customizing the default replication configuration

You set up the default replication configuration for two client nodes, NODE1 and
NODE2. You decide that you want to replicate only certain types of data in their
file spaces and that you want to replicate some data types before other data types.

About this task

NODE1 has two file spaces, /a and /b. NODE2 has one file space, /a. File space
and client replication rules for backup, archive, and space-managed data are set to
DEFAULT. Server replication rules are set to ALL_DATA. You have the following
goals:
v Replicate only the active backup data in file space /a that belongs to NODE1.
v Do not replicate any space-managed data in any of the file spaces that belong to
NODE1.
v Replicate the archive data in all file spaces that belong to NODE1 and NODE2.
Make the replication of the data a high priority.
v Replicate the active and inactive backup data in file space /a that belongs to
NODE2. Make replication of the data a high priority.

To accomplish these goals, replication rules must be set as shown in the following
figure:

958 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 File space /a File space /b File space /a

Backup data rule: Backup data rule: Backup data rule:

ACTIVE_DATA DEFAULT ALL_DATA_HIGH_PRIORITY
Archive data rule: Archive data rule: Archive data rule:
DEFAULT DEFAULT DEFAULT
Space-managed data rule: Space-managed data rule: Space-managed data rule:
DEFAULT DEFAULT DEFAULT

Backup data rule: Backup data rule:

DEFAULT DEFAULT
Archive data rule: Archive data rule:
DEFAULT DEFAULT
Space-managed data rule: Space-managed data rule:
NONE DEFAULT
NODE 1 NODE 2

Source replication server

Backup data rule:
ALL_DATA
Archive data rule:
ALL_DATA_HIGH_PRIORITY
Space-managed data rule:
ALL_DATA

REPLICATE NODE command

Replication process

High priority data

File space /a, NODE 1: File space /b, NODE 1: File space /a, NODE 2:
ALL BACKUP DATA
ARCHIVE DATA ARCHIVE DATA ARCHIVE DATA

Normal priority data

File space /a, NODE 1: File space /b, NODE 1: File space /a, NODE 2:

ACTIVE BACKUP DATA ALL BACKUP DATA SPACE-MANAGED DATA

Target replication
server

Figure 114. Customized replication configuration

To change to the configuration illustrated in the figure, complete the following

steps:

Chapter 30. Replicating client node data 959

 Procedure
v To change the server replication rule for archive data, issue the SET
ARREPLRULEDEFAULT command on the source replication server:
set arreplruledefault all_data_high_priority
v To change the replication rule for space-managed data for NODE1, issue the
UPDATE NODE command:
update node node1 spreplruledefault=none
v To change the replication rules for backup data for file spaces that belong to
NODE1 and NODE2, issue the UPDATE FILESPACE command:
update filespace /a node1 fsreplrule=bkreplrule replrule=
active_data
update filespace /a node2 fsreplrule=bkreplrule replrule=
all_data_high_priority

Results

When the REPLICATE NODE command is issued from a command line or in a

maintenance script, high-priority data is replicated first. High-priority data
includes archive data that belongs to NODE1 and NODE2 and backup data that
belongs to NODE2. No preference is given to which node’s data is replicated first.

After high-priority data is replicated, normal-priority data is replicated.

Normal-priority data includes the active data in file space /a and the backup data
in file space /b that belong to NODE1. The backup data and space-managed data
that belongs to NODE2 is replicated along with the normal-priority data that
belongs to NODE1. No preference is given to which node’s data is replicated first.

Tips:
v In Figure 114 on page 959, all the data in all the files spaces of both client nodes
is replicated in one process. However, if the amount of node data is large and
you do not have enough bandwidth to replicate data in a single process, you can
use one of the following methods:
– Schedule or manually issue separate REPLICATE NODE commands at different
times for NODE1 and NODE2.
– Replicate high-priority and normal-priority data separately at different times
by specifying the PRIORITY parameter on the REPLICATE NODE command.
– Replicate different data types at different times by specifying the DATATYPE
parameter on the REPLICATE NODE command.
– Combine replication by priority and by data type by specifying both the
PRIORITY and DATATYPE parameters on the REPLICATE NODE command.
v To verify the replication rules that apply to the file spaces in the client nodes,
issue the VALIDATE REPLICATION command. You can also use this command to
verify that the source replication server can communicate with the target
replication server. To preview results, issue the REPLICATE NODE command and
specify PREVIEW=YES.
Related concepts:
“Replication rules” on page 923

960 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Scenario: Converting to node replication from import and
export operations
You are exporting client node data from one or more servers and importing the
data to another server for disaster recovery. You want to achieve the same results
by replicating data. In this scenario, the server that was exporting data is to be the
source replication server. The server where the data was imported is to be the
target replication server.

Before you begin

Client node data that was exported and imported must be synchronized between
the source and target replication servers. You set up client nodes to synchronize
their data as part of the process of configuring nodes for replication. Data is
synchronized the first time that replication occurs. To synchronize data, the data
must be imported to the disaster recovery server by using ABSOLUTE as the value
for the DATES parameter on the IMPORT NODE command.

Procedure

To convert from import and export operations to replication operations, complete

the following steps:
1. Set up server definitions for the source and target replication servers: “Step 1:
Setting up server-to-server communications” on page 948
2. Specify the server to which you are importing data as the target replication
server: “Step 2: Specifying a target replication server” on page 950
3. On the source replication server, issue the UPDATE NODE command:
a. Specify REPLSTATE=ENABLED or REPLSTATE=DISABLED. Disabling replication
means that replication does not occur until you enable it.
b. Specify REPLMODE=SYNCSEND.
To update the definitions for all the nodes that are defined to the source
replication server, issue the following command.
update node * replstate=enabled replmode=syncsend
4. On the target replication server, issue the UPDATE NODE command and specify
REPLMODE=SYNCRECEIVE and REPLSTATE=ENABLED:
update node * replmode=syncreceive replstate=enabled
5. If necessary, modify the file-space, client-node, and server replication rules:
“Changing replication rules” on page 952.
6. Replicate the data that belongs to your nodes: “Replicating data by command”
on page 971
After all data is synchronized, the REPLMODE parameter in the client node
definition on the source replication server is set to SEND. The REPLMODE
parameter in the client node definition on the target replication server is set to
RECEIVE.

What to do next

Important: You cannot display information about running replication processes for
client nodes that are being converted from import and export operations to
replication operations. The conversion process might run for a long time, but it
occurs only once for a client node that is being converted.

Chapter 30. Replicating client node data 961

 After you set up a basic replication configuration, you can change file space,
client-node, and server replication rules. To replicate data, issue the REPLICATE NODE
command in an administrative schedule or on a command line.
Related concepts:
“Replication mode” on page 930
“Replication state” on page 927

Adding and removing client nodes for replication

You can add or remove client nodes for replication at any time after you set up a
replication configuration.

Adding client nodes for replication processing

If the data that belongs to a client node was previously exported from the source
replication server and imported to the target replication server, you must set up
the node to synchronize its data. You set up a client node to synchronize its data as
part of the process of adding the node for replication.

Before you begin

Before you add a client node for replication, ask the following questions:
v Was the data that belongs to the client node previously exported from the server
that is to be the source replication server for replicated data?
v If the data was exported, was it imported on the server that is now the target
replication server for replicated data?
v When you imported the data, did you specify DATES=ABSOLUTE on the IMPORT
NODE command?

If you answered "yes" to all of the preceding questions, you must set up to
synchronize the data on the source and target replication servers. The following
procedure explains how to set up synchronization when you add client nodes for
replication. Synchronization occurs during replication.

Restrictions:
v If a client node definition does not exist on the target replication server, do not
create it. The definition for the client node on the target replication server is
created automatically when the data of the node is replicated the first time.
v If a client node definition exists on both the source and target replication servers,
but the data that belongs to the client node was not exported and imported, you
must rename or remove the client node on the target replication server before
data can be replicated.
v If you previously removed a client node from replication on the source
replication server, but not on the target replication server, you do not have to
rename or remove the node on the target replication server.

Procedure

To add a node for replication, complete one of the following actions:

v If the data of the node was not exported from the source replication server and
imported on the target replication server, complete one of the following steps:
– If the client node is not already registered on a source replication server, issue
the REGISTER NODE command on the source replication server. Specify
REPLSTATE=ENABLED or REPLSTATE=DISABLED.

962 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 For example, to enable a new client node, NODE1, for replication, issue the
following command:
register node node1 password replstate=enabled
– If the client node is already registered on a source replication server, issue the
UPDATE NODE command on the source replication server. Specify
REPLSTATE=ENABLED or REPLSTATE=DISABLED.
For example, to enable an existing client node, NODE1, for replication, issue
the following command:
update node node1 replstate=enabled
v If you exported node data from the source replication server and imported the
data to the target replication server or if you reconfigured replication for a node
after you removed the node from replication on the source replication server,
complete the following steps:
1. On the source replication server, issue the UPDATE NODE command:
a. Specify REPLSTATE=ENABLED or REPLSTATE=DISABLED
b. Specify REPLMODE=SYNCSEND.
2. On the target replication server, issue the UPDATE NODE command:
a. Specify REPLSTATE=ENABLED or REPLSTATE=DISABLED.
b. Specify REPLMODE=SYNCRECEIVE.

Results

If you set the replication state of the client node to DISABLED, the replication
mode is set to SEND, but replication does not occur. If you set the replication state
of the client node to ENABLED, the client node definition is created on the target
replication server when replication occurs for the first time. In addition, the
replication mode of the client node on the target replication server is set to
RECEIVE, and the replication state is set to ENABLED.

If you add a file space to a client node that is configured for replication, the file
space replication rules for data types are automatically set to DEFAULT.

What to do next

After you add client nodes for replication, ensure that they are included in any
existing administrative schedules for replication. Alternatively, you can create a
schedule for replication that includes the new client nodes.
Related concepts:
“Replication mode” on page 930
“Replication state” on page 927

Removing client nodes from replication processing

Removing a client node from replication changes the node to a non-replicated
node. If you remove a client node from replication on a target replication server,
the client node can back up, archive, or migrate data to the server.

About this task

Removing a client node from replication deletes only information about replication
from the server database. Removing a node from replication does not delete the
data that belongs to the node that was replicated.

Chapter 30. Replicating client node data 963

 Procedure

To completely remove a client node from replication, issue the REMOVE REPLNODE
command on the source and target replication servers that have the node that is
configured for replication. For example, to remove NODE1 and NODE2 from
replication, issue the following command:
remove replnode node1,node2

What to do next

To verify that the node was removed, issue the QUERY NODE command on the source
and the target replication servers. For example, to verify that NODE1 and NODE2
were removed, issue the following command:
query node node1,node2 format=detailed

If the node was removed, the fields Replication State and Replication Mode are
blank. If you do not want to keep the node data that is stored on the target
replication server, you can delete it using the DELETE FILESPACE command for the
file spaces that belong to the node. If you do not want to keep the node definition
on the target replication server, delete the node definition by issuing the REMOVE
NODE command.

If you remove a client node from replication, rename the node, or delete the node
data, and then remove the node. You can add the node for replication later. All the
node data is replicated to the target replication server.

Resetting the replication mode

If you set the replication mode incorrectly when you configure a client node with
data that was exported and imported, you must remove the client node from
replication before reconfiguring the node.

About this task

For example, suppose that you updated the definition of a client node whose data
you wanted to replicate. The data that belongs to the node was previously
exported from the source replication server and imported to the target replication
server. You specified ENABLED as the setting of the REPLSTATE parameter.
However, you did not specify SYNCSEND as the replication mode on the source
replication server. As a result, the REPLMODE parameter was automatically set to
SEND, and data that belongs to the node was not synchronized or replicated.

Procedure

To reconfigure the client node for replication, complete the following steps:
1. Issue the REMOVE REPLNODE command for the client node. For example, to
remove a client node, NODE1, from replication, issue the following command:
remove replnode node1
Issuing the REMOVE REPLNODE command resets the replication state and the
replication mode for the client node to NONE.
2. Issue the UPDATE NODE command with the correct parameters and values.
For example, to enable NODE1 for replication and synchronize the data that
belongs to the node, complete the following steps:
a. On the source replication server, issue the following command:
update node node1 replstate=enabled replmode=syncsend

964 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 b. On the target replication server, issue the following command:
update node node1 replstate=enabled replmode=syncreceive

Results

After synchronization and replication are complete, the REPLMODE parameter in the
client node definition on the source replication server is set to SEND. The REPLMODE
parameter in the client node definition on the target replication server is set to
RECEIVE.
Related concepts:
“Replication mode” on page 930
“Replication state” on page 927

Managing source and target replication servers

You can change or remove target replication servers. You can also protect
communications during node replication operations with Secure Sockets Layer.

Adding a source replication server

To add a source replication server, you follow the same procedure that you
followed when you set up the default replication configuration. As part of the
procedure, you set up server-to-server communications, specify a target replication
server, and configure client nodes for replication.

About this task

You can add a source replication server to an existing configuration. For example,
suppose that you have a replication configuration comprising a single source
replication server and a single target replication server. You can add another source
server that replicates data to the existing target server.
Related concepts:
“Replication server configurations” on page 921

Selecting a new target replication server

After you set up a basic replication configuration, you can change the target
replication server.

Before you begin

Before you begin this procedure:

v Ensure that the source and the target replication servers have server definitions
for each other. The following properties must be set in the definitions:
– Server name
– Server password
– Server high-level address
– Server low-level address
v Delete any administrative schedules on the source replication server that specify
the target replication server. Create new schedules that specify the new target
replication server.

Chapter 30. Replicating client node data 965

 Procedure

To change a target replication server, issue the SET REPLSERVER command on the
source replication server. Specify the name of the new target replication server. For
example, to specify NEW_TGTSRV as the new target replication server, issue the
following command:
set replserver new_tgtsrv

Results

If you change target replication servers during replication processing, replication to

the old target replication server continues until the processing ends or is canceled.
When replication occurs for the first time with the new target replication server,
files are replicated again.

The following example describes what occurs when you change or add target
replication servers. Suppose that DRSERVER is the target replication server for
PRODSERVER. PRODSERVER has one client, NODE1.
1. Files A, B, and C that belong to NODE1 are replicated to TGTSRV.
2. You change the target replication server to NEW_TGTSRV.
3. NODE1 backs up files D, E, and F to SRCSRV.
4. Replication occurs for NODE 1. Files, A, B, and C, which were replicated to
TGTSRV, are replicated to NEW_TGTSRV. New files D, E, and F are also
replicated to NEW_TGTSRV.
5. You reset the target replication server to TGTSRV.
6. Replication occurs for NODE1. Files D, E, and F are replicated to TGTSRV.

Files A, B, C, D, E, and F now exist on SRCSRV, TGTSRV, and NEW_TGTSRV.

Related concepts:
“Replication server configurations” on page 921

Configuring the target replication server

You can configure the target replication server for automatic failover for data
recovery. It is only necessary to configure for automatic failover if you use
separate, dedicated networks for server-to-server communication and client access.

Procedure
1. Determine the default replication server by using QUERY STATUS command. For
example, issue the following command from the target replication server to
determine the default replication server:
QUERY STATUS
2. Identify the high-level address that is used by the source replication server for
replication. For example, issue the following command from the source
replication server:
QUERY SERVER
3. Identify the high-level address that the client uses to connect to the source
replication server. For example, issue the following command from the target
replication server:
QUERY SERVER

966 IBM Tivoli Storage Manager for Windows: Administrator's Guide

4. If the source replication server and the client use a different high-level address,
issue the SET FAILOVERHLADDRESS command on the target replication server. For
example, to set the high-level address for the target replication server, issue the
following command:
SET FAILOVERHladdress high_level_address

Restriction: You can specify only one failover server for each node target
replication server. The failover server information is stored in the client options
file.

Removing a replication server

If you do not want to replicate data from a source replication server to its target
replication server, you can switch or remove a target replication server in a
replication process.

About this task

You use the REMOVE REPLSERVER command to remove a server from the replication
environment. You must issue the command on both the source and the target
replication servers. If you use this command on only one of the replication servers,
the data is not replicated.

Procedure
1. Identify all nodes that replicated data on the target server that is being
removed. On the source replication server, issue the QUERY REPLNODE command.
2. On the source replication server, issue the QUERY REPLSERVER command to
identify the GUID for the target server you want to remove.
3. Take note of the GUID for the source server. You specify the GUID later in the
procedure.
4. On the source replication server, issue the REMOVE REPLSERVER command, and
specify the GUID for the target server. For example, to remove a replication
server and all replication state information that is associated with a server,
issue the following command:
remove replserver 11.9c.54.e0.8a.b5.11.d6.b3.c3.[Link].c1.5b
5. Remove all of the replication state information for data that is replicated from
the source server. On the target replication server, issue the REMOVE REPLSERVER
command, and specify the GUID for the source server.
6. Remove all of the nodes that were replicated from the source server. On the
target replication server, issue the REMOVE REPLNODE command.

Results

The node definitions on the source server that were replicated to the specified
target server are in the replication SEND mode. You can continue to replicate these
nodes to other replication servers by using the SET REPLSERVER command to
modify the default replication server value.

Chapter 30. Replicating client node data 967

 Scenario: Automatic failover
You are replicating client node data from a source replication server to a target
server. You want to ensure that client restore operations can automatically fail over
to the target replication server, if the source replication server is unavailable.

In the following scenario, replication operations are set up to automatically fail

over from the source replication server to the target replication server:
1. A Tivoli Storage Manager V7.1 client connects to a Tivoli Storage Manager V7.1
source replication server. The source replication server provides the connection
information for the target replication server to the client.
2. If the source replication server is unavailable, the client uses the connection
information for the target replication server and logs on to the target replication
server.
3. If you replicated client node data to the target replication server, the client can
restore the data. The client cannot store data to the target replication server.

Configuring a server for SSL communications

You can configure source and target replication servers to use Secure Sockets Layer
(SSL) instead of TCP/IP.

Before you begin

A server that uses SSL can obtain a unique certificate that is signed by a certificate
authority (CA), or the server can use a self-signed certificate. Before you start the
source and target replication servers, install the certificates and add them to the
key database files. Required SSL certificates must be in the key database file that
belongs to each server. SSL support is active if the server options file contains the
SSLTCPPORT or SSLTCPADMINPORT option or if a server is defined with SSL=YES at
startup.

If the client is configured for SSL communication, the client uses the SSL port to
communicate with the target server during failover.

Because the certificate is provided by a trusted authority, it is accepted by the

server and, therefore, communication between servers can begin. You can query a
server-defined password by issuing the QUERY SSLKEYRINGPW command. Because the
server uses random characters, it is better to update the password by issuing the
following command:
SET SSLKEYRINGPW newpw UPDATE=Y

The server and its database are updated with the new password. After you update
the password, shut down the server, add the certificates, and start the server.

About this task

To determine whether a server uses SSL, issue the QUERY SERVER command.

Procedure

To update a server definition for SSL, issue the UPDATE SERVER command. For
example, to update the server definition for server PHOENIX_SRV, issue the
following command:
update server phoenix_srv ssl=yes

968 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 What to do next

Restriction: For event servers, library servers, and target replication servers, the
name of the source replication server must match the value of the SET SERVERNAME
command on the target server. Because the source replication server uses the name
of the target replication server for SSL communications, the name is not available
for the TCP/IP server definition that is required for enterprise events or library
sharing.

If you enable SSL communications and are using the following functions, you must
create separate source and target definitions that use TCP/IP for the corresponding
server-to-server communications:
v Enterprise configuration
v Command routing
v Virtual volumes
v LAN-free

Configuring SSL-enabled servers for non-SSL communications:

If you use SSL with node replication, you must create separate server definitions
for enterprise configuration, command routing, virtual volumes, and LAN-free
communications.

About this task

Suppose that you want to use a source replication server to replicate data and to
route commands. In the option file of the target replication server, the value of the
TCPPORT option is 1500. The value of the SSLTCPPORT option is 1542.

Procedure

To create separate definitions, issue the following commands. For enterprise

configuration, command routing, virtual volumes, and LAN-free communications,
you can use the server name NOSSL. The server definition name must have a
value that corresponds to the value of the SET SERVERNAME command on the target
replication server.
define server nossl hladdress=[Link] lladdress=1500 ssl=no
serverpasword=xxxxx

You can use the server name SSL for node replication:
define server ssl hladdress=[Link] lladdress=1542 ssl=yes
serverpassword=xxxxx

Verifying a node replication setup before processing

Before you start a replication process, take steps to ensure that your configuration
is set up correctly.

Chapter 30. Replicating client node data 969

 Validating a replication configuration
Before you start a replication process, you can display the controlling replication
rules for data types in the file spaces of client nodes that are configured for
replication. You can also verify the connection to the target replication server.

About this task

A controlling rule is the rule that the source replication server uses to replicate data
in a file space. For example, suppose the replication rule for backup data in file
space /a is DEFAULT. If the client-node rule for backup data is ALL_DATA, the
controlling rule for the backup data in file space /a is ALL_DATA.

Procedure

To validate a replication configuration, complete one of the following steps:

v To display a list of file spaces and the controlling replication rules that apply to
the file spaces, issue the VALIDATE REPLICATION command. For example, to
display the controlling replication rules that apply to data types in the file
spaces for NODE1 and NODE2, issue the following command:
validate replication node1,node2

All file spaces are displayed regardless of whether the state of the data types in
the file spaces is enabled or disabled.
v To display the controlling replication rules and verify the connection with the
target replication server, issue the following command:
validate replication node1,node2 verifyconnection=yes

Previewing node replication results

Preview results before you start a replication process.

Procedure

To preview results, complete one of the following steps:

v To display the following information, issue the REPLICATE NODE command and
specify PREVIEW=YES:
– The names of client nodes to be replicated
– The number of files to be replicated, deleted, or updated
– The estimated time to complete the process
– The tape volumes to be mounted
For example, to display the estimated time to complete the replication process
for a defined group of nodes, NODEGRP1, issue the following command:
replicate node nodegrp1 preview=yes
v To display a list of files to be replicated, issue the REPLICATE NODE command and
specify PREVIEW=YES LISTFILES=YES. For example, to display a list of files to be
replicated for NODEGRP1, issue the following command:
replicate node nodegrp1 preview=yes listfiles=yes

Specifying the LISTFILES parameter signifies that the WAIT parameter is set to
YES and that you cannot the issue the WAIT parameter from the server console.

970 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 Results

Remember: Information is displayed only if the following conditions are true:

v Data types in file spaces are enabled for replication. For example, suppose that a
file space contains backup and archive files and that replication of backup data
is enabled. Replication of archive data is disabled. In this case, only information
about backup files is displayed.
v Client nodes are enabled for replication.
v If the controlling replication rule is a server rule, the server rule must be
enabled.

Managing data replication

With Tivoli Storage Manager, you can have various methods to control node
replication processing.

Replicating data by command

You can replicate data that belongs to one or more client nodes or the data that
belongs to a defined group of client nodes. To replicate data automatically and at a
scheduled frequency, define an administrative schedule.

Procedure

You can replicate data in one of the following ways:

v By issuing the REPLICATE NODE command on the command line
v By using the DEFINE SCHEDULE command to create an administrative schedule,
and specify when to start replication processing
To use an administrative schedule, you create it by using the DEFINE SCHEDULE
command. When you start replication by using an administrative schedule, node
replication processing occurs at a specified frequency. Regardless of how you start
replication, the command must be issued on the source replication server.
For example, suppose that you have three client nodes, NODE1, NODE2, and
NODE3. To replicate data from all nodes daily at 6:00 a.m. in a schedule named
REPL_SCHED, issue the following command on the source replication server:
define schedule repl_sched type=administrative cmd="replicate node
node*" active=yes"
description="replicate clients with name noden" starttime=[Link] period=1
perunits=days

Results

High-priority data is replicated before data with a normal priority. To schedule

replication processes, you use the DEFINE SCHEDULE command.

If the data that belongs to a client node is being replicated, any attempt to replicate
the data by issuing another REPLICATE NODE command fails. For example, suppose
the backup data that belongs to a client node is scheduled for replication at 6:00
a.m. Replication of the archive data is scheduled for 8:00 a.m. Replication of the
backup-data must complete before replication of the archive data starts.

Example

If you have many client nodes and are replicating a large amount of data, you can
replicate data more efficiently by issuing several REPLICATE NODE commands in

Chapter 30. Replicating client node data 971

 separate schedules. For example, replicate the data that belongs to the most
important client nodes first in a single command. After the data that belongs to
those client nodes is replicated, replicate the data that belongs to the other nodes.

Tip: To ensure that replication for first group of client nodes finishes before the
replication for the other nodes starts, specify WAIT=YES on the first REPLICATE NODE
command. For example, if you want to replicate the data that belongs to NODE1
and NODE2 before the data that belongs to NODE3 and NODE4, issue the
following commands:
replicate node node1,node2 wait=yes
replicate node node3,node4

You can use the REPLBATCHSIZE and REPLSIZETHRESH server options to

increase the speed of a node replication process between two replicating servers.
The REPLBATCHSIZE option specifies the number of client files that are to be
replicated in a batch, within the same server transaction. The REPLSIZETHRESH
option specifies, in megabytes, a threshold for the amount of data that is replicated
as a batch, within the same server transaction.

Replicating data in individual file spaces

You can replicate the data in one or more file spaces that belong to individual
client nodes or defined groups of client nodes. Replicating data by file space can
be useful if, for example, you must replicate new data in a file space that belongs
to a certain client node, and you cannot wait for a scheduled replication.

About this task

Data is replicated for a file space only when the following conditions are true:
v The replication state for data types in file spaces are enabled. For example, if the
replication state for archive data in a file space is enabled, archive data in the
file space is replicated.
v The controlling rule for the data type in the file space cannot be NONE. For
example, suppose the replication rule for archive data in a file space is
DEFAULT. If the file space rules and client node rules for archive data are both
DEFAULT and the server rule for archive data is NONE, archive data in the file
space is not replicated.

Procedure

To replicate data by file space, issue the REPLICATE NODE command and specify the
file space name or file space identifier. For example, to replicate data in file space
/a in NODE1, issue the following command:
replicate node node1 /a

What to do next

Tip: With the REPLICATE NODE command, you can also replicate data by priority
and by data type. To achieve greater control over replication processing, you can
combine replication by file space, data type, and priority.

To obtain information about the node replication process while it is running, issue
the QUERY PROCESS command:
query process

For node replication purposes, each file space contains three logical file spaces:

972 IBM Tivoli Storage Manager for Windows: Administrator's Guide

v One for backup objects
v One for archive objects
v One for space-managed objects
By default, the QUERY PROCESS command reports results for each logical file space.
Other factors also affect the output of the QUERY PROCESS command:
v If a file space has a replication rule that is set to NONE, the file space is not
included in the count of file spaces that are being processed.
v If you specify data types in the REPLICATE NODE command, only those data types
are included in the count of file spaces that are being processed, minus any file
spaces that are specifically excluded.

Example: Backing up four file spaces with three object types

In this example, NODE1 has four file spaces with three object types. The QUERY
PROCESS command generates the following output for node replication:
Process Process Process Status
Number Description
--------- --------------- --------------
40 Replicate Node Replicating node(s) NODE1 File spaces
complete: 11. File spaces identifying
and replicating: 0. File spaces
replicating: 1. File spaces not
started: 0. Files current: 0. Files
replicated: 747,915 of 913,581. Files
updated: 0 of 0. Files deleted: 0 of 0.
Amount replicated: 546 GB of 732 GB.
Amount transferred: 389 GB. Elapsed time:
0 Day(s), 4 Hour(s), 59 Minute(s).

Because the example includes four file spaces with three object types, 12 logical file
spaces are processed for replication. The QUERY PROCESS command output shows
that 11 logical file spaces completed replication.
Related concepts:
“Node replication processing” on page 923

Replicating data by type

You can replicate one or more data types for individual nodes or defined groups of
client nodes. Replicating data by type can be useful if, for example, you have new
backup data, and you cannot wait for a scheduled replication.

About this task

Remember: Replication of a data in a file space occurs only if the following

conditions are true:
v Replication for data types in file spaces must be enabled. For example, if
replication for archive data in a file space is disabled, archive data in the file
space is not replicated.
v The controlling rule for the data type in the file space cannot be NONE. For
example, suppose the replication rule for archive data in a file space is
DEFAULT. If the file space rules and client node rules for archive data are both
DEFAULT and the server rule for archive data is NONE, archive data in the file
space is not replicated.

To replicate data by type, complete one of the following steps:

Chapter 30. Replicating client node data 973

 Procedure
v To replicate active and inactive backup data, issue the REPLICATE NODE command
and specify DATATYPE=BACKUP. For example, to replicate the active and inactive
backup data that belongs to NODE1, issue the following command:
replicate node node1 datatype=backup
v To replicate only active backup data, issue the REPLICATE NODE command and
specify DATATYPE=BACKUPACTIVE. For example, to replicate the active backup data
that belongs to NODE1, issue the following command:
replicate node node1 datatype=backupactive
v To replicate archive data, issue the REPLICATE NODE command and specify
DATATYPE=ARCHIVE. For example, to replicate the archive data that belongs to
NODE1, issue the following command:
replicate node node1 datatype=archive
v To replicate space-managed data, issue the REPLICATE NODE command and
specify DATATYPE=SPACEMANAGED. For example, to replicate the space-managed
backup data that belongs to NODE1, issue the following command:
replicate node node1 datatype=spacemanaged
v To replicate multiple data types, separate the data types by commas with no
intervening spaces. For example, to replicate active backup data and archive
data that belongs to NODE1, issue the following command:
replicate node node1 datatype=backupactive,archive

Example

If you do not specify a type on the REPLICATE NODE command, all data types are
replicated.

What to do next

Tip: Using the REPLICATE NODE command, you can also replicate data by file space
and by priority. To achieve greater control over replication processing, you can
combine replication by data type, file space, and priority.
Related concepts:
“Node replication processing” on page 923

Replicating data by priority

You can replicate high-priority data or normal-priority data that belongs to
individual client nodes or defined groups of client nodes. Replicating data by
priority can be useful if, for example, a client node recently acquired high-priority
data, and you cannot wait for a scheduled replication.

About this task

Remember: Replication of data in a file space occurs only if the following

conditions are true:
v Replication for data types in file spaces must be enabled. For example, if
replication for archive data in a file space is disabled, archive data in the file
space is not replicated.
v The controlling rule for the data type in the file space cannot be NONE. For
example, suppose the replication rule for archive data in a file space is
DEFAULT. If the file-space and client-node rules for archive data are both
DEFAULT and the server rule for archive data is NONE, archive data in the file
space is not replicated.

974 IBM Tivoli Storage Manager for Windows: Administrator's Guide

To replicate data by priority, complete one of the following steps:

Procedure
v To replicate only high-priority data, issue the REPLICATE NODE command and
specify PRIORITY=HIGH. For example, to replicate the high-priority data that
belongs to NODE1, issue the following command:
replicate node node1 priority=high
v To replicate only normal-priority data, issue the REPLICATE NODE command and
specify PRIORITY=NORMAL. For example, to replicate normal-priority data that
belongs to NODE1, issue the following command:
replicate node node1 priority=normal

Example

If you do not specify a priority on the REPLICATE NODE command, normal-priority

and high-priority data is replicated.

What to do next

Tip: Using the REPLICATE NODE command, you can also replicate data by file space
and by data type. To achieve greater control over replication processing, you can
combine replication by priority, file space, and data type.
Related concepts:
“Node replication processing” on page 923

Scenario: Combining replication by file space, data type, and

priority
You can combine replication by file space, data type, and priority to accomplish
unique replication tasks. In this scenario, you replicate normal-priority archive and
space-managed data in a file space that is common to two nodes.

About this task

The name of the file space is /a. It is common to NODE1 and NODE2.

Procedure

To replicate the data in the file space, issue the following command:
replicate node node1,node2 /a priority=normal datatype=archive,spacemanaged

Issuing this command replicates archive and space-managed data that is assigned
the replication rule ALL_DATA.
Related concepts:
“Node replication processing” on page 923

Chapter 30. Replicating client node data 975

 Controlling throughput for node replication
Bandwidth can affect the performance of replication processing workload,
processor capacity, the time available for replication, and the amount of network.

Managing the number of replication sessions

You can manage the maximum number of sessions that are used for sending data
to the target replication server to improve replication performance.

About this task

Use the MAXSESSIONS parameter to specify the maximum number of sessions to use.
When you calculate the value for the MAXSESSIONS parameter, consider the available
network bandwidth and the processor capacity of source and target replication
servers.

Consider the number of logical and physical drives that can be dedicated to the
replication process. You must ensure that there are enough drives available for
replication processing because other server processes or client sessions might also
be using drives. The number of mount points and drives available for replication
operations depends on the following factors:
v Tivoli Storage Manager server activity that is not related to replication
v System activity
v The mount limits of the device classes for the sequential-access storage pools
that are involved
v The availability of a physical drive on the source and target replication servers,
if the device type is not FILE
v The available network bandwidth and the processor capacity of source and
target replication servers

Procedure

Issue the REPLICATE NODE command and specify the MAXSESSIONS parameter to
determine the number of data sessions. For example, to set the maximum number
of replication sessions to 6 for NODE_GROUP1, issue the following command:
replicate node node_group1 maxsessions=6

Restriction: The following values apply to the MAXSESSIONS parameter:

v The value that is specified by the MAXSESSIONS parameter applies only to data
sessions. Data sessions are sessions during which data is sent to a target
replication server. If you issue a QUERY SESSION command, the total number of
sessions might exceed the number of data sessions. The difference can occur if
you use short control sessions that you use to query and set up replication
operations.
v The value of the MAXSESSIONS parameter represents the maximum allowable
number of sessions. The number of sessions that are used for replication
depends on the amount of data to be replicated. If you are replicating only a
small amount of data, there is no advantage to increasing the number of
sessions.

976 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Tuning replication processing
You can determine how many bytes per hour are replicated over your network at a
specific time by tuning the settings that affect replication and by testing replication
processing. Based on the replication results and the total amount of node data that
you replicate, you can choose which method to use for the initial replication.

About this task

Do not use a storage pool that is enabled for data deduplication to test replication.
By using storage pools that are not enabled for data deduplication to test
replication processing, you avoid processing extents that can increase the amount
of preprocessing time of the replication process. By determining the data transfer
and network capability of your replication operation without extent processing,
you get a better representation of the capability of your system. Test replication
processing with storage pools that are enabled for data deduplication if you want
to determine the effect of data deduplication on replication performance alone.

You must calculate the bytes-per-hour value for each source replication server
individually. You can determine which method is the most suitable for the server,
based on its bytes-per-hour value.

Complete the following steps to determine how much data you can replicate
during a specified timeframe so that you can tune replication processing for a
server. Repeat these steps to obtain bytes-per-hour value for each server that you
want to use for replication processing.

Procedure
1. Complete the following steps to select the appropriate data:
a. Select one or more nodes and file spaces that have approximately 500 GB to
1 TB of total data.
b. Select data that is typical of the data that you replicate on a routine basis.
c. Select nodes that are configured for replication.
2. To display the amount of data in a file space, issue the QUERY OCCUPANCY
command.
3. Select a timeframe during which replication is running normally.
4. If you plan to use Secure Sockets Layer (SSL) as the communication protocol
for replication processing, ensure that SSL is enabled.

Restriction: SSL can slow replication processing.

5. Start the replication process by issuing the REPLICATE NODE command.
6. After the replication is completed, review the summary message that is
displayed. Use the Amount transferred and Elapsed time values that display in
the summary message to determine the number of bytes per hour that can be
replicated.

Results

When you determine the bytes-per-hour value for each server, you can determine a
method to use for the initial replication.
Related tasks:
“Selecting a method for the initial replication” on page 943

Chapter 30. Replicating client node data 977

 Managing more workload during replication
By testing replication processing, you can determine whether your network can
manage more workload.

Procedure

To see how your network manages more workload during replication, complete
the following tasks:
1. Increase the value of the MAXSESSIONS parameter by 10 on the REPLICATE NODE
command and run the test again.
2. Increase the number of replication sessions by 10 to transfer more data
concurrently during replication. Alternatively, if you determine that 10
replication sessions (the default MAXSESSIONS value) cause your network to
degrade below acceptable levels, decrease the value of the MAXSESSIONS
parameter.
3. Repeat the process, and adjust the value of the MAXSESSIONS parameter to
determine optimal data transfer capability.

Tuning mount points for data deduplication and replication

You can specify the maximum allowable number of data sessions that are used to
send data to a target replication server. When you increase the number of data
sessions to improve node replication throughput. Use storage pools that are
enabled for data deduplication to improve replication performance.

Procedure
1. Determine the number of logical and physical drives that can be dedicated to
the replication process. For example, if a library has 10 tape drives and 4 of the
drives are used for another task, 6 drives are available for node replication.
2. Specify the number of mount points you require and ensure that there are
drives available to complete node replication. For example, to set the number of
point points to 6, issue the following UPDATE DEVCLASS command and specify
the MOUNTLIMIT parameter:
update devclass device_class_name mountlimit=6
3. Specify the number of data sessions that you want by using the MAXSESSIONS
parameter on the REPLICATE NODE command. For example, issue the following
REPLICATE NODE command and specify the MAXSESSIONS parameter:
replicate node prodnode maxsessions=6

Disabling and enabling node replication

With Tivoli Storage Manager, you can temporarily disable node replication. You
can enable node replication at any time.

Disabling and enabling replication of data types in a file space

Disabling replication of a data type in a file space can be useful if, for example, a
file space has large amounts of backup data and archive data. You can enable the
replication of backup data and disable replication of archive data. After the backup
data is replicated, you can enable replication of archive data.

About this task

To determine the replication state of a data type in a file space, issue the QUERY
FILESPACE command with the FORMAT parameter set to DETAILED.

978 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Restriction: You cannot disable or enable replication for an entire file space. You
can only disable and enable replication of a data type in a file space.

To disable and enable replication of data types in a file space, complete one or
more of the following actions:

Procedure
v Backup data:
– To disable replication, issue the following command:
update filespace node_name file_space_name datatype=backup
replstate=disabled
– To enable replication, issue the following command:
update filespace node_name file_space_name datatype=backup
replstate=enabled
v Archive data:
– To disable replication, issue the following command:
update filespace node_name file_space_name datatype=archive
replstate=disabled
– To enable replication, issue the following command:
update filespace node_name file_space_name datatype=archive
replstate=enabled
v Space-managed data:
– To disable replication, issue the following command:
update filespace node_name file_space_name datatype=spacemanaged
replstate=disabled
– To enable replication, issue the following command:
update filespace node_name file_space_name datatype=spacemanaged
replstate=enabled
Related concepts:
“Replication state” on page 927

Disabling and enabling replication for individual client nodes

Disabling replication for a client node causes the server to skip the node during
replication processing. Disabling replication for a client node can be useful if, for
example, data that belongs to the node is corrupted or contaminated. After the
problem is fixed, you can enable replication.

About this task

To determine the replication state of a node, issue the QUERY NODE command.

Procedure
v To disable replication for a node, issue the UPDATE NODE command and specify
REPLSTATE=DISABLED. For example, to disable replication for NODE1, issue the
following command:
update node node1 replstate=disabled
v To enable replication for a node, issue the UPDATE NODE command and specify
REPLSTATE=ENABLED. For example, to enable replication for NODE1, issue the
following command:
update node node1 replstate=enabled

Chapter 30. Replicating client node data 979

 What to do next

Remember: If you disable replication for a client node while data that belongs to
the node is being replicated, the replication process is not affected. Replication of
the data continues until all the data that belongs to the client node is replicated.
However, replication for the client node will be skipped the next time that
replication runs.
Related concepts:
“Replication state” on page 927

Disabling and enabling replication for all client nodes

You can disable and enable outbound and inbound sessions, including replication
sessions, between servers. If sessions are disabled, you must enable them to restore
function.

Disabling and enabling outbound or inbound sessions:

Disabling outbound or inbound sessions can be useful if, for example, you plan a
network outage that affects communication between source and target replication
servers. Disabling and enabling sessions affects not only node replication
operations but also certain other types of operations.

About this task

To display the status and direction of sessions for a particular server, issue the
QUERY STATUS command.

Remember:
v When you disable sessions for a particular server, you disable the following
types of sessions in addition to replication:
– Server-to-server event logging
– Enterprise management
– Server registration
– LAN-free sessions between storage agents and the Tivoli Storage Manager
server
– Data storage that uses virtual volumes
v If you disable only outbound sessions on a source replication server, data that
belongs to client nodes that store data on the source replication server do not
have their data replicated. However, inbound sessions to the target replication
server can occur.
If a server is the target for multiple source replication servers and you disable
outbound sessions on a single source replication server, the target replication
server continues to receive replicated data from the other source replication
servers.

To disable and enable replication sessions:

Procedure
v Between a single source replication server and a single target replication server:
– To disable sessions, issue one of the following commands:
- On the source replication server:
disable sessions server target_server_name direction=outbound

980 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 - On the target replication server:
disable sessions server source_server_name direction=inbound
– To enable sessions:
- If outbound sessions to the target replication server are disabled on the
source replication server, issue the following command on the source
replication server:
enable sessions server target_server_name direction=outbound
- If inbound sessions from the source replication server are disabled on the
target replication server, issue the following command on the target
replication server:
enable sessions server source_server_name direction=inbound
v Between multiple source replication servers and a single target replication
server:
– To disable sessions, issue one of the following commands:
- On the source replication servers:
disable sessions server * direction=outbound
- On the target replication server:
disable sessions server * direction=inbound
– To enable sessions:
- If outbound sessions to the target replication server are disabled on the
source replication servers, issue the following command on the source
replication servers:
enable sessions server * direction=outbound
- If inbound sessions to the target replication server are disabled on the
source replication servers, issue the following command on the target
replication server:
enable sessions server * direction=inbound

Disabling and enabling outbound node replication processing:

When you disable outbound node replication processing, you prevent new
replication processes from starting on a source replication server. Enabling
outbound node replication processing is required after a database restore.

About this task

Restriction: When you restore the Tivoli Storage Manager database, replication is
automatically disabled. Disabling replication prevents the server from deleting
copies of data on the target replication server that are not referenced by the
restored database. After a database restore, you must re-enable replication.

To display the status of replication processing for a particular server, issue the
QUERY STATUS command.

Issue the following commands on the source replication server to disable and
enable replication processing:

Procedure
v To disable replication, issue the DISABLE REPLICATION command.
v To enable replication, issue the ENABLE REPLICATION command.

Chapter 30. Replicating client node data 981

 Disabling and enabling replication rules
Disabling a replication rule suspends replication of all data to which the rule
applies. For example, you can disable replication of active backup data in all client
nodes and file spaces that have the replication rule ACTIVE_DATA.

About this task

Disabling a replication rule can be useful if, for example, you replicate groups of
normal-priority and high-priority client nodes on different schedules. For example,
suppose that the data that belongs to some client nodes is assigned the
ALL_DATA_HIGH_PRIORITY replication rule. The data that belongs to other
client nodes is assigned the ALL_DATA replication rule. The client nodes are
separated into groups, in which some of the nodes in each group have
high-priority data and other nodes in the group have normal-priority data.

You schedule replication for each group to take place at different times. However, a
problem occurs, and replication processes take longer than expected to complete.
As a result, the high-priority data that belongs to client nodes in groups that are
scheduled late in the replication cycle is not being replicated.

To replicate the high-priority data as soon as possible, you can disable the
ALL_DATA rule and rerun replication. When you rerun replication, only the client
node data that is assigned the ALL_DATA_HIGH_PRIORITY rule is replicated.
After high-priority data is replicated and the problem that caused the delay is
fixed, you can enable the ALL_DATA replication rule and resume scheduled
replication processing.

You can disable or enable the following replication rules:

ALL_DATA
Replicates all backup, archive, and space-managed data. The data is
replicated with a normal priority.
ACTIVE_DATA
Replicates only the active backup data. The data is replicated with a
normal priority.
ALL_DATA_HIGH_PRIORITY
Replicates backup, archive, and space-managed data. The data is replicated
before data that has a normal priority.
ACTIVE_DATA_HIGH_PRIORITY
Replicates only active backup data. The data is replicated before data that
has a normal priority.

To disable and enable replication rules, complete one of the following steps:

Procedure
v To disable a replication rule, issue the UPDATE REPLRULE command and specify
STATE=DISABLED. For example, to disable the replication rule
ACTIVE_DATA_HIGH_PRIORITY, issue the following command:
update replrule active_data_high_priority state=disabled
v To enable a replication rule, issue the UPDATE REPLRULE command and specify
STATE=ENABLED. For example, to enable the replication rule
ACTIVE_DATA_HIGH_PRIORITY, issue the following command:
update replrule active_data_high_priority state=enabled
Related concepts:

982 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 “Replication state” on page 927

Purging replicated data in a file space

You can purge the backup, archive, or space-managed data from a file space on the
target replication server and prevent future replication of the data type.

Procedure

To prevent replication of a data type and purge the data from the file space on the
target replication server, issue the UPDATE FILESPACE command and specify
REPLSTATE=PURGEDATA. For example, to prevent replication of backup data in file
space /a on NODE1 and delete the backup data in file space /a on the target
replication server, issue the following command:
update filespace node1 /a datatype=backup replstate=purgedata

Data is purged the next time that replication runs for the file space. After the data
is purged, the replication rule for the specified data type is set to DEFAULT.
Replication for the data type is disabled.

Replicating client node data after a database restore

When you restore the Tivoli Storage Manager database on a source replication
server, replication is automatically disabled. Before re-enabling replication, you can
take steps to preserve the client node data that is on the target replication server.

About this task

Disabling replication prevents the Tivoli Storage Manager server from deleting
copies of data on the target replication server that are not referenced by the
restored database. Before re-enabling replication, determine whether copies of data
that are on the target replication server are needed. If they are, complete the steps
that are described in the following example. In the example, the name of the
source replication server is PRODSRV. DRSRV is the name of the target replication
server. NODE1 is a client node with replicated data on PRODSRV and DRSRV.

Restriction: You cannot use Secure Sockets Layer (SSL) for database restore
operations.

Procedure
1. Remove NODE1 from replication on PRODSRV and DRSRV by issuing the
REMOVE REPLNODE command:
remove replnode node1
2. Update NODE1 definitions PRODSRV and DRSRV. When replication occurs,
DRSRV sends the data to PRODSRV that was lost because of the database
restore.
a. On DRSRV, issue the UPDATE NODE command and specify the replication
mode SYNCSEND:
update node node1 replstate=enabled replmode=syncsend
b. On PRODSRV, issue the UPDATE NODE command and specify the replication
mode SYNCRECEIVE:
update node node1 replstate=enabled replmode=syncreceive
3. On DRSRV, set the replication rules to match the rules on PRODSRV. For
example, if only archive data was being replicated from PRODSRV to DRSRV,

Chapter 30. Replicating client node data 983

 set the rules on DRSRV to replicate only archive data from DRSRV to
PRODSRV. Backup and space-managed data are not replicated to PRODSRV.
To set rules, you can issue the following commands:
v UPDATE FILESPACE
v UPDATE NODE
v SET ARREPLRULEDEFAULT
v SET BKREPLRULEDEFAULT
v SET SPREPLRULE
4. On DRSRV, issue the SET REPLSERVER command to set PRODSRV as the target
replication server:
set replserver prodsrv
5. On DRSRV, issue the REPLICATE NODE command to replicate data that is stored
on NODE1:
replicate node node1

Replication processing changes the replication state of NODE1 to SEND on

DRSRV and to RECEIVE on PRODSRV.
6. Remove NODE1 from replication on PRODSRV and DRSRV by issuing the
REMOVE REPLNODE command:
remove replnode node1
7. Update NODE1 definitions:
a. On DRSRV, issue the UPDATE NODE command and specify the replication
mode SYNCRECEIVE:
update node node1 replstate=enabled replmode=syncreceive
b. On PRODSRV, issue the UPDATE NODE command and specify the replication
mode SYNCSEND:
update node node1 replstate=enabled replmode=syncsend
8. On PRODSRV, enable replication by issuing the ENABLE REPLICATION command:
enable replication
9. On PRODSRV, issue the REPLICATE NODE command to replicate data that is
stored on NODE1:
replicate node node1

Replication processing changes the replication state of NODE1 to SEND on

PRODSRV and to RECEIVE on DRSRV.

Results

The original replication configuration is restored. PRODSRV has all the data that
was lost because of the database restore.

What to do next

Remember: In step 4, you set the PRODSRV as the target replication server for
DRSRV. If in your original configuration you were replicating data from DRSRV to
another server, you must reset the target replication server on DRSRV. For example,
if you were replicating data from DRSRV to BKUPDRSRV, issue the following
command on DRSRV:
set replserver bkupdrsrv

984 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 Canceling replication processes
To cancel all running node-replication processes, issue the CANCEL REPLICATION
command.

Monitoring node replication processing and verifying results

Tivoli Storage Manager provides various methods for monitoring node replication
processes and verifying results.

Displaying information about node replication settings

You can display information about replication settings for file spaces and client
nodes. You can also display information about replication rules.

Displaying information about node replication settings for file

spaces
To display the replication rules and the replication state of data types for file
spaces, issue the QUERY FILESPACE command.

Displaying information about node replication settings for client

nodes
To display information about the replication state, mode, and rules for client nodes,
issue the QUERY NODE command.

Displaying information about node replication rules

To display information about replication rules, including the state, issue the QUERY
REPLRULE command.

Displaying information about node replication processes

You can display information about running and ended replication processes, such
as the estimated time for completion of a replication process.

Displaying information about running and completed node

replication processes
A replication record is created when a replication process is started. The record is
retained in the database of the source replication server.

Procedure

To display records about running and completed node-replication processes for a

node or a file space that belongs to a node, issue the QUERY REPLICATION command.

Important: You cannot display information about running replication processes for
client nodes that are being converted from import and export operations to
replication operations. The data synchronization process might run for a long time,
but it occurs only once for a client node that is being converted.
The default record-retention period for completed processes is 30 days. To display
the retention period, issue the QUERY STATUS command and check the value in the
Replication Record Retention Period field.
The record for a running process is updated only after a group of files is processed
and committed. A file group consists of 2,000 files or 2 GB of data, whichever is
smaller. For example, if a single file is 450 GB, the record is not updated for a
relatively long time. If you notice that the number of files that are not yet
replicated for a running process is not decreasing fast enough, network bandwidth
or time might be insufficient to replicate the amount of data. Take one of the
following actions:

Chapter 30. Replicating client node data 985

 v Provide more time for replication.
v Decrease the amount of data to replicate.
v Create more parallel data-transmission sessions between the source and target
replication servers by increasing the value of the MAXSESSIONS parameter.
Increase the value of the MAXSESSIONS parameter only if network bandwidth and
processor resources for the source and target replication servers are sufficient.

Tip: To display summary of information about a running replication process,

including file spaces and bytes replicated, issue the QUERY PROCESS command.

Displaying the activity log of a replication process

To display messages that are associated with a node replication process, review the
source replication server activity log.

Procedure

The server activity log contains messages with the following information:
v The nodes that were enabled or disabled for replication
v The number of files that were eligible to be replicated compared to the number
of those files that were already stored on the target replication server
v The number of files that were successfully replicated and the number of files
that were missed
v The number of files on the target replication server that were deleted

Displaying summary records of replication processes

Summary records are created for each node that is processed by a REPLICATE NODE
command. Summary records also document the entire replication operation at the
process level.

Procedure

To view summary records, issue the following command:

select * from summary

Measuring the effectiveness of a replication configuration

A replication configuration is optimal if the number of replicated files that are
stored on a target replication server is equal to the number of files that are stored
on the source replication server. If there are fewer files on the target server, start
replication processing.

Procedure

To display the number of files that are stored on source and target replication
servers, issue the QUERY REPLNODE command. You can issue the command on a
source or a target replication server.

Results

The information in the output for QUERY REPLNODE includes files that are stored at
the time the command is issued. If a replication process is running, the information
does not include files that are waiting to be transferred. Information is reported by
data type. For example, you can determine the number of backup files that belong
to a client node that are stored on the source and the target replication servers.

986 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 The number of replication-eligible files on the source replication server might be
much larger than the number of replicated files on the target replication server, or
getting larger over time. If so, one or more of the following problems might be the
reason:
v Replication processes are being canceled prematurely.
v The number of sessions for transferring data is insufficient.
v The target replication server might be busy with operations not related to
replication, and replication processing is not being given sufficient processor
cycles.

To help diagnose problems with replication processing, issue the QUERY

REPLICATION command.

Measuring the effects of data deduplication on node

replication processing
The destination storage pool on the target replication server might be enabled for
data deduplication. If it is, the source replication server sends only extents that are
not in the storage pool.

Procedure

To display information about the effects of data deduplication on replication

processing, issue the QUERY REPLICATION command and specify FORMAT=DETAILED.

Results
In the output, check the values in the fields that represent bytes replicated and
bytes transferred for each data type:
v Replicated bytes are bytes that were replicated to the target replication server. If
a file was stored in a deduplicated storage pool, the number of bytes in the
stored file might be less than the number of bytes in the original file. This value
in this field represents the number of physical bytes in the original file.
v Transferred bytes represent the number of bytes that were sent to the target
replication server. For files stored in a deduplicated storage pool, the value in
this field includes the number of bytes in the original file before duplicate
extents were removed. If duplicate extents were already on the target replication
server, the number of bytes in the original file is more than the number of bytes
transferred.
Related concepts:
“Replication of deduplicated data” on page 931
“Active log mirror” on page 648
Related tasks:
Part 6, “Protecting the server,” on page 829

Chapter 30. Replicating client node data 987

 Retaining replication records
To maintain adequate information about replication processes, you can adjust the
length of time that the source replication server retains replication records in its
database.

About this task

A replication record is created when a replication process is started. By default,

replication records are retained for 30 calendar days. A calendar day comprises
24-hours, from midnight to midnight. For example, suppose that the retention
period is two calendar days. If a replication process completes at 11:00 p.m. on day
n, a record of that process is retained for 25 hours until midnight on day n+1.

To display the retention period for replication records, issue the QUERY STATUS
command on the source replication server.

Procedure

To set the retention period for replication records, issue the SET REPLRETENTION
command.

Example

Replication records that exceed the retention period are deleted from the database
by Tivoli Storage Manager during automatic inventory-expiration processing. As a
result, the amount of time that retention records are retained can exceed the
specified retention period

If a replication process runs longer than the retention period, the record of the
process is not deleted until the process ends, the retention period passes, and
expiration runs.

What to do next
To display the contents of replication records, issue the QUERY REPLICATION
command and specify FORMAT=DETAILED.

Recovering and storing client data after a disaster

If a source replication server is unavailable, client nodes can restore, retrieve, and
recall data from the target replication server. If a source replication server is
unavailable and client store operations are necessary, the nodes can back up,
archive, and migrate data to the server.

Restoring, retrieving, and recalling data manually from a

target replication server
To manually restore, retrieve, or recall data from a target replication server, update
the client options file to point to the target replication server. Changes to node
replication settings are not required.

988 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 Converting client nodes for store operations on a target
replication server
If a source replication server is unavailable, you can convert client nodes to be
non-replicating nodes. Non-replicating client nodes can back up, archive, or
migrate data to a target replication server.

About this task

Important: Before you convert client nodes to be non-replicating nodes, carefully

consider the consequences. For example, suppose that you remove multiple client
nodes from replication on source replication server SRCSRV and back up, archive,
and migrate the node data to target replication server TGTSRV. Later, you decide
to resume backup, archive, and migration operations to SRCSRV and to replicate to
TGTSRV. The client node data on the two servers must be synchronized. The
synchronization process can be lengthy and resource-intensive.

To convert a client node for storage operations on a target replication server,

complete the following steps:

Procedure
1. On the target replication server, issue the REMOVE REPLNODE command. For
example, to modify the definition of NODE1 so that it is a non-replicating
node, issue the following command:
remove replnode node1
2. To back up, archive, or migrate data, update the client options file to point to
the target replication server.

What to do next

If any schedules were defined on the source replication server, you can redefine
them on the target replication server. Client node data on the target replication
server is now managed by policies on the target replication server. For example,
file expiration and deletion are managed by the target replication server.

Removing a node replication configuration

If you do not want to replicate data, you can remove a node replication
configuration.

Before you begin

Delete any administrative schedules on source replication servers that issue the
REPLICATE NODE command for the client nodes that are included in the
configuration.

Procedure
1. Remove all nodes that are configured for replication by issuing the REMOVE
REPLNODE command on all source and target replication servers that have the
nodes that are configured for replication. For example, to remove
NODE_GROUP1 from replication, issue the following command:
remove replnode node_group1

Chapter 30. Replicating client node data 989

 2. Remove the target replication server by issuing the SET REPLSERVER command
on the source replication server. Do not specify the name of the target
replication server. For example, to remove target replication server TGTSRV,
issue the following command:
set replserver

What to do next

To verify that the target replication server was removed, issue the QUERY STATUS
command on the source replication server. If the target replication server was
removed, the field Target Replication Server is blank.

Tip: If you do not want to keep replicated node data on the target replication
server, you can delete the replicated data.

990 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Chapter 31. Configuring clustered environments
You can configure a Tivoli Storage Manager server for clustering on AIX, Linux, or
Windows systems.

You can use a clustered environment for the following operating systems:
v IBM PowerHA® SystemMirror for AIX
v IBM Tivoli System Automation for Multiplatforms for AIX and Linux
v Microsoft Failover Cluster for Windows

You can use other cluster products with Tivoli Storage Manager, however,
documentation is not available and support is limited. For the latest information
about support for clustered environments, see [Link]
[Link]?uid=swg21609772.

Before you use another cluster product, verify that DB2 supports the required file
systems. For more information about the level of DB2 that you are using, refer to
the DB2 documentation at: [Link]
Search on Recommended file systems.

For more information about upgrading the server in a clustered environment, see
the Installation Guide.

Clustered environment overview

Clusters consist of many components such as Tivoli Storage Manager servers,
hardware, and software. You can use clustering to join two or more servers or
nodes by using a shared disk system.

This configuration provides the nodes with the ability to share data, which allows
higher server availability and minimized downtime. For example:
v You can configure, monitor, and control applications and hardware components
that are deployed on a cluster.
v You can use an administrative cluster interface and Tivoli Storage Manager to
designate cluster arrangements and define a failover pattern. The server is part
of the cluster that provides an extra level of security by ensuring that no
transactions are missed because a server failed. The failover pattern that you
establish prevents future failures.
v You can apply clustering to the node replication process. In this way, server
availability is higher than it would be if node replication is used as a process on
its own because a client is less likely to fail over to another server in a clustered
environment. If you replicate data from several source replication servers to one
target replication server, there is a high dependency on the target replication
server. A clustered environment eases the dependency on the target replication
server.

Components in a server cluster are known as cluster objects. Cluster objects are
associated with a set of properties that have data values that describe the identity
and behavior of an object in the cluster. Cluster objects can include the following
components:
v Nodes

© Copyright IBM Corp. 1993, 2013 991

 v Storage
v Services and applications
v Networks
You manage cluster objects by manipulating their properties, typically through a
cluster management application.
Related concepts:
Chapter 30, “Replicating client node data,” on page 919

Cluster nodes
Nodes in a cluster all have similar characteristics, which allows them to work
together.

Cluster nodes include the following similarities:

v Every node has access to all cluster configuration data.
v Every node communicates with the other nodes in the cluster through one or
more physically independent networks (sometimes referred to as interconnects).
Network adapters, referred to in server clusters as network interfaces, attach
nodes to networks.
v Every node in the cluster is aware when another system joins or leaves the
cluster.
v Every node in the cluster is aware of the resources that are running locally as
well as the resources that are running the other cluster nodes.
v All nodes in the cluster are grouped under the cluster name, which is used to
access and manage the cluster.

Configuring a Windows clustered environment

You can configure a Tivoli Storage Manager server for Windows in a Microsoft
failover cluster environment. On Windows, cluster environments consist of
components such as Tivoli Storage Manager servers, hardware, and software.
When these components are connected to the same disk system, downtime is
minimized.

Microsoft software helps configure, monitor, and control applications and hardware
components that are deployed on a Windows cluster. The administrator uses the
Microsoft Cluster Administrator interface and Tivoli Storage Manager to designate
cluster arrangements and define the failover pattern.

Tivoli Storage Manager supports tape failover for a cluster environment by using a
Fibre or SCSI connection. Although Microsoft failover clusters do not support the
failover of tape devices, the failover configuration can be monitored through the
Microsoft Cluster Administrator interface after it is set up through Tivoli Storage
Manager.

992 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Microsoft Failover Cluster environment overview
With a Microsoft Failover Cluster Manager, you can place Tivoli Storage Manager
server cluster resources into a cluster group. The Tivoli Storage Manager cluster
group has a network name, an IP address, one or more physical disks, a DB2
server, and a Tivoli Storage Manager server service.

The Tivoli Storage Manager instance network name is independent of the name of
the physical node on which the Tivoli Storage Manager cluster group runs and
migrates from node to node. Clients connect to a Tivoli Storage Manager server by
using the Tivoli Storage Manager instance network name, rather than the Windows
node name. The Tivoli Storage Manager instance network name maps to a primary
or backup node. The mapping is dependent on which node owns the Tivoli
Storage Manager cluster group. Any client that uses Windows Internet Name
Service (WINS) or directory services to locate servers can automatically track the
Tivoli Storage Manager clustered server as it moves between nodes. You can
automatically track the Tivoli Storage Manager clustered server without modifying
or reconfiguring the client.

Each Tivoli Storage Manager cluster group has its own disk as part of a cluster
resource group. Tivoli Storage Manager cluster groups cannot share data between
the cluster groups. Each Tivoli Storage Manager server that is configured in a
Tivoli Storage Manager cluster group has its database, active logs, recovery logs,
and set of storage pool volumes on a separate disk owned by that Tivoli Storage
Manager cluster group.

Note: Microsoft Failover Cluster Manager only supports an IP address as a

resource. This means that any Tivoli Storage Manager server that runs on a cluster
must limit its supported communication method to just TCP/IP. Any client that
does not use TCP/IP as a communication method is not able to reach the Tivoli
Storage Manager cluster group if it fails over to the other cluster node.

The following example demonstrates the way that a Microsoft Failover Cluster
Manager for a Tivoli Storage Manager cluster server works.

Assume that a clustered Tivoli Storage Manager server that is named JUPITER is
running on Node Z and a clustered Tivoli Storage Manager server that is named
SATURN is running on Node X. Clients connect to the Tivoli Storage Manager
server JUPITER and the Tivoli Storage Manager server SATURN without knowing
which node hosts their server.

Chapter 31. Configuring clustered environments (AIX, Linux, Windows) 993

 Tivoli Storage Manager
cluster group
JUPITER
Shared SCSI
disk array
Local
disk

Node Z
SATURN

Node X
Client Local
disk

Client

Client
Figure 115. Clustering with JUPITER as Node Z and SATURN as Node X

When one of the software or hardware resources fails, failover occurs. Resources
(for example: applications, disks, and an IP address) migrate from the failed node
to the remaining node. The remaining node:
v Takes over the Tivoli Storage Manager cluster group
v Brings the disk resources, the network resources, and the DB2 resource online
v Restarts the Tivoli Storage Manager service
v Provides access to administrators and clients

If Node X fails, Node Z assumes the role of running SATURN. To a client, it is

exactly as if Node X were turned off and immediately turned back on again.
Clients experience the loss of all connections to SATURN and all active
transactions are rolled back to the client. Clients must reconnect to SATURN after
the connection is lost. The location of SATURN is not apparent to the client.

Tape failover for nodes in a cluster

Groups in a cluster can be transferred to other nodes when the node that is hosting
the groups fails.

A node can host physical or logical units, referred to as resources. Administrators

organize these cluster resources into functional units called groups and assign
these groups to individual nodes. If a node fails, the server cluster transfers the
groups that were being hosted by the node to other nodes in the cluster. This
transfer process is called failover. The reverse process, failback, occurs when the
failed node becomes active again and the groups that were failed over to the other
nodes are transferred back to the original node.

994 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Fiber tape failover
Tivoli Storage Manager can manage the failover of Fibre Channel direct-attached
tape and library devices on a Microsoft Windows system in a clustered
environment without additional hardware.

The following table describes the hardware and software that you can use with
Tivoli Storage Manager fiber-tape failover.
Table 78. Hardware and software supported for fiber-tape failover
Operating system Fibre Channel adapter Fibre Channel tape library
and drives
Microsoft Windows Server QLogic QLE2462 with a IBM or other vendor Fibre
2008, Microsoft Windows Storport driver. Channel directly attached
Server 2008 R2 (64-bit), or tape and library devices. For
Microsoft Windows 2012. the latest list of supported
devices, see the IBM Support
Portal.

Ensure that the same level of Windows (Windows 2008, Windows 2008 R2, or
Windows 2012) is installed on all computers in the cluster.

To meet the minimum requirements for failover, ensure that one of the following
components is installed on the Windows system:
v Directly attached Fibre Channel tape or library device from IBM
v Directly attached tape or library device from another vendor

For best results, install both a Fibre Channel adapter and a Fibre Channel tape
library or drive on the Windows system. If you install a Fibre Channel adapter, it
must use a Storport driver.

Restriction: You cannot use a miniport driver.

The LUN reset function that is used to clear the tape device reservation can only
be used with a Storport driver. However, with Tivoli Storage Manager V6.3.1 and
later, if you use persistent reservation, you can install a Fibre Channel adapter
without the Storport driver because the LUN reset function is not required. Do not
use the LUN reset function to break the reservation if the drive is reserved by
persistent reservation. Use the LUN reset function only to release the drive if drive
is reserved by the previous reserve commands.

If you use persistent reservation, ensure that you select a tape drive that you can
use for persistent reservation for all clusters within the system. For IBM tape
drives, see the IBM Tape Device Drivers Installation and User's Guide. For more
information about persistent reservation, see the section that describes persistent
reserve for tape drives.

Tip: If there is a failover in a clustered environment, the server tries to open an

available drive. If the drive is locked, the server tries to break the persistent
reservation.

Chapter 31. Configuring clustered environments (AIX, Linux, Windows) 995

 Planning for a clustered environment
Configuration in a clustered environment takes planning to ensure the optimal
performance of your system. Whether you configure your system to include
clusters depends on your business needs.

Plan for a cluster configuration that accommodates your environment. In addition

to assuring the right type of hardware and the applicable software, you must set
up a failover pattern.

When a node fails or needs to be taken offline, which node or nodes in the cluster
picks up the transaction processing? In a two-node cluster, there is little planning
necessary. In a more complex arrangement, you want to consider how your
transaction processing is best handled. A form of load balancing among your
nodes needs to be accounted for so that you maintain peak performance. Another
consideration is to ensure that your customers do not see any lag and little drop in
productivity.

Microsoft Cluster Servers and Microsoft Failover Clusters require each Tivoli
Storage Manager server instance to have a private set of disk resources. Although
nodes can share disk resources, only one node can actively control a disk at a time.

Attention: Ensure that the same level of Windows (Windows 2008, Windows 2008
R2, or Windows 2012) is installed on all computers in the cluster.

Is one configuration better than the other? To determine your best installation, you
need to look at the differences in performance and cost. Assume that you have a
Tivoli Storage Manager server-dedicated cluster whose nodes have comparable
power. During failover, the performance of a configuration might degrade because
one node must manage both Tivoli Storage Manager Cluster Instances. If each
node handles 100 clients in a normal operation, one node must handle 200 clients
during a failure.

Cluster configuration worksheet

Record your answers to the following planning questions before you set up the
cluster configuration.
1. What type of cluster solution best fits my business needs?
2. What type of failover pattern do I need?
The use of tape failover support also affects the pattern.
3. Will tape failover support be needed?
Consider how tape devices are used by the Tivoli Storage Manager Cluster
instances. Remember that this limits the number of nodes in the failover
pattern to two.
4. What are the resources to be dedicated to Tivoli Storage Manager?

Resource type Resource name

Cluster Resource Group
Physical Disk Resources
IP address
Subnet Mask
Network
Network Name (server name)
Nodes

996 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Resource type Resource name
Tape Failover (optional): device name - both
nodes

Planning for cluster hardware and software configuration

Cluster hardware and software configuration is determined during the planning
stage and before the actual installation.

Procedure

The following guidelines help determine what resources are needed for a
successful Tivoli Storage Manager cluster:
1. Decide which cluster configuration you must use with servers that use disk
devices. Each Tivoli Storage Manager Cluster Instance needs a separate set of
disk resources on the shared disk subsystem. You might have problems if you
configure the I/O subsystem as one large array when you configure a two
server cluster and later decide to expand to a four server cluster.
2. Identify the disk resources to be dedicated to Tivoli Storage Manager. Do not
divide a shared disk into multiple partitions with each partition assigned to a
different application and thus a different cluster group.
For example, Application A, a stable application, might be forced to fail over
because of a software problem with Application B if both applications use
partitions that are part of the same physical disk. This failover might happen,
for example, when a software problem with Application B occurs. This problem
causes the Cluster Services to fail over Application B and its corequisite disk
resource. Because the partitions exist on the same physical drive, Application A
is also forced to fail over. Therefore, as a best practice, when you install and
configure a Tivoli Storage Manager application, dedicate a shared disk as a
resource that can fail if necessary.
3. Ensure that you have an IP address and network name for each Tivoli Storage
Manager server instance that you configure. For a cluster that involves two
Tivoli Storage Manager cluster instances, two network names are required.
4. Create a cluster resource group and move disk resources to it. Each Tivoli
Storage Manager server instance requires a cluster resource group. Initially, the
group must contain only disk resources. You might choose just to rename an
existing resource group that contains only disk resources.
5. Tivoli Storage Manager is installed to a local disk on each node in the cluster.
Determine the disk to be used on each node. It is a best practice to use the
same drive letter on each system. When Tivoli Storage Manager server is
installed in a cluster environment, the SANDISCOVERY option must be set to ON.
By default, this option is set to OFF.
6. If you choose not to use Tivoli Storage Manager tape failover support, you can
attach tape devices in either of the following configurations:

Chapter 31. Configuring clustered environments (AIX, Linux, Windows) 997

 What to do
Advantages and Disk space that How to enable when a failover
Configuration disadvantages is required migration occurs
Attach to the This Define enough Set up a storage Manually
node on which configuration volume space for pool hierarchy disconnect the
the Tivoli allows which data is so that data is tape device and
Storage Manager high- disk-based to migrated reattach it to the
server instance is performance keep more than efficiently to the node on which
active. backup and two days worth tape device. the server was
restore of average data. activated.
operations.
However, it is
not entirely
automated
because operator
intervention is
required to
service a failover
when repair
delays occur.
Attach to a This Define enough Use the virtual No action is
third, configuration volume space for volumes to required; the
non-clustered might not be which data is migrate the data activated server
system on which feasible in disk-based to from the local continues to use
an extra instance installations with keep more than disk volumes to the virtual
of the Tivoli low-bandwidth two days worth the tape device. volumes.
Storage Manager communications of average data.
server is active. between the
servers in the
cluster and the
tape-device
controller server.

Configuring Tivoli Storage Manager in Microsoft Failover Cluster

The Tivoli Storage Manager cluster configuration procedure must be performed on
the set of nodes that hosts a Tivoli Storage Manager cluster group.

Steps for the procedure vary depending upon which node you are currently
configuring. When configuring the primary node in the set, the Tivoli Storage
Manager server instance is created and configured. When configuring the
remaining nodes in the set, each node is updated in such a way that permits it to
host the Tivoli Storage Manager server instance created on the primary node. A
Tivoli Storage Manager server must be installed and configured on the first node
in the set before configuring the remaining nodes in the set. Violating this
requirement causes the configuration to fail.

Ensure that you completely configure one Tivoli Storage Manager cluster group
before moving on to the next when configuring multiple Tivoli Storage Manager
cluster groups. Because you are dealing with separate IP addresses and network
names for each Tivoli Storage Manager cluster group, you lessen the possibility of
mistakes by configuring each Tivoli Storage Manager cluster group separately.

998 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Setting up Tivoli Storage Manager in a Microsoft Failover
Cluster
You must ensure that your cluster is properly installed and configured before you
install Tivoli Storage Manager.

Procedure

To configure the Tivoli Storage Manager in a Microsoft Failover Cluster, complete

the following steps:
1. Ensure that the Windows operating system is installed on all computers that
are part of the cluster. Failover clustering is supported by Microsoft Windows
Server 2008 and Windows Server 2008 R2. See Microsoft Failover Cluster
documentation for detailed support information.
2. Log on with domain user ID. The domain user must be in the same domain as
the Tivoli Storage Manager server.
3. Ensure that the failover cluster is installed and configured for all the computers
in the cluster. To install Tivoli Storage Manager in a Microsoft Failover Cluster,
the cluster must be installed and fully functional on your servers.
4. Verify that each node and shared disk in the cluster is operational.
5. Ensure that the shared tape devices are operational if Tivoli Storage Manager
tape failover support is being used.

Preparing a Microsoft Failover Cluster group for a basic virtual

server
Each Tivoli Storage Manager server instance requires a cluster resource group.

Before you begin

Use the Failover Cluster Manager program on the computer that owns the shared
disk or tape resource to prepare your resource group. Initially, the group must
contain only disk resources. You can create a group and move disk resources to it.
You can also choose to rename an existing resource group that contains only disk
resources.

As you construct your resource groups, consider the following items:

v Ensure that each resource group has a distinctive name. Do not change the
names after the group is created because it can cause a corrupted configuration.
v Ensure that all nodes in the cluster are online.
v Ensure that the group is online and owned by the node where the initial server
instance is installed.

Procedure

To prepare a resource group for cluster configuration, complete the following steps:
1. Open the Failover Cluster Manager program. Right-click on Services and
Applications and then choose More Actions > Create Empty Service or
Application.
2. Right-click on New Service or Application, select Change the name and
choose a new name for the resource group, for example, TSMGROUP.
3. Right-click on the resource group TSMGROUP and select Add storage.

Chapter 31. Configuring clustered environments (AIX, Linux, Windows) 999

 4. On the Add storage area panel, select the shared volume or volumes for Tivoli
Storage Manager and click OK. The resource group TSMGROUP, which
contains the disk volumes you just added, is displayed.

Installing Tivoli Storage Manager in a Microsoft Failover Cluster

Install the Tivoli Storage Manager server on every node in the cluster that hosts a
Tivoli Storage Manager clustered server.

Procedure

Complete the following steps for each node in your cluster to install the Tivoli
Storage Manager server:
1. Log in with the domain user ID. The domain user must be a member of the
Domain Administrators group.
2. Install the Tivoli Storage Manager server to a local disk on each node. Use the
same local disk drive letter for each node.
3. Restart the system after the server installation completes.

Initializing the Tivoli Storage Manager server for a Microsoft

Failover Cluster on the primary node
After you install Tivoli Storage Manager on the nodes in the cluster, you must
initialize the server on the primary node.

Procedure
1. Ensure that all systems are restarted after the installation. Verify that all
systems are running correctly.
2. Log on with the domain user ID. The domain user must be in the same domain
as the Tivoli Storage Manager server.
3. Open the Failover Cluster Manager program and verify that the resources are
online and available to the primary node.
4. Begin the initialization procedure on the primary node in your cluster. In the
Failover Cluster Manager program, verify that the owner of the resource group
is the primary node in your cluster.
5. From the Start menu, click All programs > Tivoli Storage Manager >
Configuration wizard.
6. Follow the wizard directions, clicking Next to step through the wizard. When
you are prompted to enter the user ID, enter the name of the domain account
to associate with the cluster.
7. If the initialization completed, click Done.

Verifying the configuration of Tivoli Storage Manager in a

Microsoft Failover Cluster
When you finish configuring Tivoli Storage Manager in a Microsoft Failover
Cluster, you can review the Failover Cluster Manager summary window. Verify
that clustering is completed successfully and the Tivoli Storage Manager server is
started.

Procedure

To verify that the Tivoli Storage Manager server instance in a Microsoft Failover
Cluster is created and configured correctly, complete the following steps:
1. From the Failover Cluster Manager, select the server instance. The network
name that you configured is displayed in the Server Name pane.

1000 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 2. In the Other Resources pane, confirm that the server instance and the IBM DB2
server resource are displayed.
3. Right-click the Tivoli Storage Manager server instance and select Bring this
resource online.

Completing a failover test for your cluster

After completing the cluster configuration, run a failover test to ensure that the
nodes are working properly.

Procedure
1. Open Failover Cluster Manager. Under Other Resources, right-click the Tivoli
Storage Manager Instance(x) resource. Select Bring this resource online.
2. To test the failover, right-click the Tivoli Storage Manager cluster resource
group and select Move this service or application to another node.
3. Verify that the failover from the second node to the first node completes
successfully.

Maintaining the clustered environment

After you have set up your initial cluster or clusters, maintenance needs are
minimal.

Check your Windows Event log on a regular, if not daily, basis to monitor the
activity of the nodes in the cluster. By checking the log, you are informed that a
node has failed and needs maintenance.

The following list of topics describes situations that might affect the configuration
or format of your cluster after it is operational.

Migrating an existing Tivoli Storage Manager server into a

cluster
The reason for migrating an existing server into a cluster is similar to the reason
for adding a server to a cluster. You want to increase the availability and reliability
of data to all your users. By having the server as part of the cluster, you provide
an extra level of security by ensuring that no transactions are missed due to a
failed server. The failover pattern you establish prevents future failures.

About this task

To migrate an existing Tivoli Storage Manager server into a cluster, you can either
move the clients or perform a backup and restore procedure. The choice depends
primarily on the availability and capacity of other Tivoli Storage Manager server
computers in your site and your familiarity with the backup and restore procedure.

Moving the clients:

If you move clients from a non-clustered Tivoli Storage Manager server computer
to a clustered one, you can have the time to gradually migrate your users to the
new system and not interrupt services. However, you must have the hardware
needed to run two Tivoli Storage Manager servers simultaneously.

About this task

You can move clients in either of the following ways:

Chapter 31. Configuring clustered environments (AIX, Linux, Windows) 1001

 v You can register clients with the new clustered Tivoli Storage Manager server.
Have them perform a full backup of their data to the clustered server. When you
are sure that the client data is safely on the clustered server, that client can be
deleted from the old server. When all clients have been moved to the clustered
server, the old server can be eliminated.
v You can export the clients from the old server and import those clients into the
clustered server. After you have verified that a client's data is on the clustered
server, you can delete that client from the old server. After you have imported
all clients to the clustered server, you can delete the old server.

Adding a Tivoli Storage Manager server with backup and restore

If you have limited hardware resources, you can add an existing Tivoli Storage
Manager server to a cluster by using a backup and restore procedure.

About this task

For example, suppose that you have no hardware other than the two servers to be
clustered and you plan to use the computer that is currently running the Tivoli
Storage Manager server as a node. Follow this procedure to remove Tivoli Storage
Manager from the computer and reinstall it in the cluster:

Procedure
1. Back up all disk storage pools to a copy storage pool.
2. Back up the database of the existing Tivoli Storage Manager server.
3. Perform the installation and configuration of the cluster.
4. Restore the database to the clustered Tivoli Storage Manager server.
5. Restore the disk storage pool volumes from the copy storage pool.
6. After you verify that all of your data is on the clustered server, delete the old
server.

Managing a virtual Tivoli Storage Manager server on a cluster

For most tasks, you can administer a virtual Tivoli Storage Manager server as you
would a non-clustered server. To complete some important tasks such as starting
and stopping the server or moving a resource group to another node to complete
system maintenance, you must use the Microsoft Cluster Administrator interface.
The virtual server configuration includes details such as the physical Windows
servers that are part of the cluster and their resources, network connections, and
status.

About this task

The Microsoft Cluster Administrator interface is available through the

Administrative Tools program group. The interface is a detailed view of a virtual
server configuration. View the components of a virtual server configuration and
start, stop, or fail back a virtual server by using this interface. Manage a virtual
Tivoli Storage Manager server by using the Microsoft Cluster Administrator
interface to avoid server failures and error messages. For example, if you use the
Windows Service Control Manager to shut down the server, you might receive
messages that the server failed.

You might want to move a virtual Tivoli Storage Manager server when the
Windows server acts as the primary node and this server requires hardware or
system maintenance. Use the Microsoft Cluster Administrator interface to move the
management of the virtual Tivoli Storage Manager server to the secondary node

1002 IBM Tivoli Storage Manager for Windows: Administrator's Guide

until the maintenance is completed.

Managing tape failover in a cluster

As part of your regular routine, check the event log to ensure that the
configuration is operating properly. If a server fails, the error is logged. The log
provides you with information to understand why the failure took place.

About this task

Sometimes a node must rejoin the cluster, for example:

v When a node failed
v When a new Host Bust Adapter fiber card is added (equipment changes)

Procedure

Complete the following tasks in any order to ensure that a node can successfully
join the cluster:
v Update, if necessary, the drive and library that use the Tivoli Storage Manager
cluster tool.
v Take the Tivoli Storage Manager server offline until the failed node rejoins the
cluster. This action helps ensure that the Tivoli Storage Manager server that is
running on the other node is not affected.

Troubleshooting with Tivoli Storage Manager cluster log

The Tivoli Storage Manager Cluster Resource DLL reports events and errors to the
cluster log. The cluster log is a useful troubleshooting tool. When this log is
enabled, it records the actions of each component of the Cluster service as the
result of each action.

The cluster log is a complete record of cluster activity compared to the Microsoft
Windows Event Log. The cluster log records the cluster service activity that is
recorded in the event log. Although the event log can point you to a problem, the
cluster log helps you resolve the problem.

The cluster log is enabled by default in Windows. Its output is printed as a log file
in %SystemRoot%\Cluster. For more information, see the Windows online help
documentation.

Chapter 31. Configuring clustered environments (AIX, Linux, Windows) 1003

1004 IBM Tivoli Storage Manager for Windows: Administrator's Guide
Chapter 32. Disaster recovery manager
You can use the disaster recovery manager (DRM) function to prepare a plan that
can help you to recover your applications if a disaster occurs.

About this task

You can recover data at an alternative site, on replacement computer hardware,

and with people who are not familiar with the applications. You can also manage
your offsite recovery media, store client recovery information, and use the disaster
recovery plan for audits to certify the recoverability of the server.

To recover from a disaster, you must know the location of your offsite recovery
media. DRM helps you to determine which volumes to move offsite and back
onsite and track the location of the volumes.

You can use complementary technologies to protect the Tivoli Storage Manager
server and to provide an alternative to disaster recovery. For example, you can use
DB2 HADR to replicate the Tivoli Storage Manager database or device-to-device
replication.

To ensure high server availability, you can use node replication with clustering to
minimize a disaster risk. In a clustered environment, a client is less likely to fail
over to another server. If you replicate data from several source replication servers
to one target replication server, there is a high dependency on the target replication
server. A clustered environment eases the dependency on the target replication
server.

The following list identifies key tasks:

v “Querying defaults for the disaster recovery plan file” on page 1006
v “Specifying recovery instructions for your site” on page 1011
v “Specifying information about your server and client node machines” on page
1013
v “Specifying recovery media for client machines” on page 1017
v “Creating and storing the disaster recovery plan” on page 1018
v “Managing disaster recovery plan files stored on target servers” on page 1022
v “Moving backup media” on page 1024
v “Managing the Disaster Recovery Manager tasks” on page 1029
v “Preparing for disaster recovery” on page 1032
v “Recovering from a disaster” on page 1034

Before you use DRM, familiarize yourself with Chapter 29, “Protecting and
recovering the server infrastructure and client data,” on page 869.

Note: Unless otherwise noted, you need system privilege class to perform DRM
tasks.
Related concepts:
“Clustered environment overview” on page 991
Related reference:
“Disaster recovery manager checklist” on page 1043

© Copyright IBM Corp. 1993, 2013 1005

 “The disaster recovery plan file” on page 1048

Querying defaults for the disaster recovery plan file

DRM provides default settings for the preparation of the recovery plan file and for
the management of offsite recovery media.

About this task

To query the settings, issue the following command:

query drmstatus

The output will be similar to the following:

Recovery Plan Prefix: c:\Program Files\Tivoli\TSM\server2\recplans\
Plan Instructions Prefix: c:\Program Files\Tivoli\TSM\server2\recinstr\
Replacement Volume Postfix: @
Primary Storage Pools: PRIM1 PRIM2
Copy Storage Pools: COPY*
Active-data Storage Pools: ACTIVEPOOL
Not Mountable Location Name: Local
Courier Name: Joe’s Courier Service
Vault Site Name: Ironvault, D. Lastname, 1-000-000-0000
DB Backup Series Expiration Days: 30 Day(s)
Recovery Plan File Expiration Days: 60 Day(s)
Check Label?: Yes
Process FILE Device Type?: No
Command File Name: c:\drm\orm\[Link]

Specifying defaults for the disaster recovery plan file

You can override the default settings in the recovery plan file.

About this task

The following table describes how to set defaults for the disaster recovery plan file.
Table 79. Defaults for the disaster recovery plan file
Process Default
Primary storage pools to be When the recovery plan file is generated, you can limit processing to specified pools.
processed The recovery plan file will not include recovery information and commands for
storage pools with a data format of NETAPPDUMP.

The default at installation: All primary storage pools.

To change the default: SET DRMPRIMSTGPOOL

For example, to specify that only the primary storage pools named PRIM1 and PRIM2
are to be processed, enter:
set drmprimstgpool prim1,prim2

Note: To remove all previously specified primary storage pool names and thus select
all primary storage pools for processing, specify a null string ("") in SET
DRMPRIMSTGPOOL.

To override the default: Specify primary storage pool names in the PREPARE
command

1006 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Table 79. Defaults for the disaster recovery plan file (continued)
Process Default
Copy storage pools to be When the recovery plan file is generated, you can limit processing to specified pools.
processed
The default at installation: All copy storage pools.

To change the default: SET DRMCOPYSTGPOOL

For example, to specify that only the copy storage pools named COPY1 and COPY2
are to be processed, enter:
set drmcopystgpool copy1,copy2

To remove any specified copy storage pool names, and thus select all copy storage
pools, specify a null string ("") in SET DRMCOPYSTGPOOL. If you specify both
primary storage pools (using the SET DRMPRIMSTGPOOL command) and copy
storage pools (using the SET DRMCOPYSTGPOOL command), the specified copy
storage pools should be those used to back up the specified primary storage pools.

To override the default: Specify copy storage pool names in the PREPARE command
Active-data pools to be When the recovery plan file is generated, you can limit processing to specified pools.
processed
The default at installation: None

To specify the default: SET DRMACTIVEDATASTGPOOL

For example, to specify that only the active-data pools named ACTIVEPOOL1 and
ACTIVEPOOL2 are to be processed, enter:
set drmactivedatastgpool activepool1,activepool2

To remove any specified active-data pool names, specify a null string ("") in SET
DRMACTIVEDATASTGPOOL.

Active-data pool volumes in MOUNTABLE state are processed only if you specify the
active-data pools using the SET DRMACTIVEDATASTGPOOL command or the
ACTIVEDATASTGPOOL parameter on the MOVE DRMEDIA, QUERY DRMEDIA,
and PREPARE commands. Processing of active-data pool volumes in MOUNTABLE
state is different than the processing of copy storage pool volumes in MOUNTABLE
state. All MOUNTABLE copy storage pool volumes are processed regardless whether
you specify copy storage pools with either the SET DRMCOPYSTGPOOL command
or the COPYSTGPOOL parameter.

If you do not issue the SET DRMACTIVEDATASTGPOOL command or if you use

this command to remove the names of all active-data storage pools, the Tivoli Storage
Manager server processes active-data pool volumes specified using the
ACTIVEDATASTGPOOL parameter:
v MOVE DRMEDIA and QUERY DRMEDIA: The server processes all active-data
pool volumes except those in MOUNTABLE state.
v PREPARE: The server processes only the active-data pool volumes that are marked
onsite at the time the PREPARE command is run. These volumes are marked
UNAVAILABLE.

To override the default: Specify active-data pool names using the MOVE DRMEDIA,
QUERY DRMEDIA, or PREPARE command.

Chapter 32. Disaster recovery manager 1007

Table 79. Defaults for the disaster recovery plan file (continued)
Process Default
Identifier for replacement To restore a primary storage pool volume, mark the original volume destroyed and
volume names create a replacement volume having a unique name. You can specify a character to be
appended to the name of the original volume in order to create a name for the
replacement volume. This character can help you find the replacement volume names
in the disaster recovery plan.

The default identifier at installation: @

To change the default: SET DRMPLANVPOSTFIX

For example, to use the character r, enter:

set drmplanvpostfix r
Recovery instructions You can specify a prefix for the names of the recovery instructions source files in the
prefix recovery plan file.

The default at installation: For a description of how DRM determines the default
prefix, see the INSTRPREFIX parameter of the PREPARE command section in the
Administrator's Reference or enter HELP PREPARE from administrative client
command line.

To set a default: SET DRMINSTRPREFIX

For example, to specify the prefix, enter:

set drminstrprefix c:\Program Files\Tivoli\
TSM\server2\recinstr\rpp

The disaster recovery plan file will include, for example, the following file:
c:\Program Files\Tivoli\TSM\server2\recinstr\
[Link]

To override the default: The INSTRPREFIX parameter with the PREPARE command
Prefix for the recovery plan You can specify a prefix to the path name of the recovery plan file. DRM uses this
file prefix to identify the location of the recovery plan file and to generate the macros and
script file names included in the [Link]
and [Link] stanzas.

The default at installation: For a description of how DRM determines the default
prefix, see the PLANPREFIX parameter of the PREPARE command section in the
Administrator's Reference or enter HELP PREPARE from administrative client
command line.

To change the default: SET DRMPLANPREFIX

For example, to specify the prefix, enter the following command:

set drmplanprefix c:\Program Files\Tivoli\TSM\
server2\recplans\

The disaster recovery plan file name created by PREPARE processing will be in the
following format:
c:\Program Files\Tivoli\TSM\server2\recplans\20000603.013030

To override the default: The PLANPREFIX parameter with the PREPARE command

1008 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Table 79. Defaults for the disaster recovery plan file (continued)
Process Default
The disaster recovery plan You can set the numbers of days after creation that a disaster recovery plan file stored
expiration period on a target server expires. After the number of days has elapsed, all recovery plan
files that meet both of the following conditions are eligible for expiration:
v The last recovery plan associated with the database series is older than the set
number of days.
v The recovery plan file is not associated with the most recent backup series.

The default at installation: 60 days

To change the default: SET DRMRPFEXPIREDAYS

For example, to change the time to 90 days, enter:

set drmrpfexpiredays 90

Specifying defaults for offsite recovery media management

You can set defaults for offsite recovery media management.

About this task

Table 80. Defaults for offsite recovery media management
Process Default
Copy storage pool volumes MOVE DRMEDIA and QUERY DRMEDIA can process copy storage pool volumes in
to be processed the MOUNTABLE state. You can limit processing to specified copy storage pools.

The default at installation: All copy storage pool volumes in the MOUNTABLE state

To change the default: SET DRMCOPYSTGPOOL

To override the default: COPYSTGPOOL parameter on MOVE DRMEDIA or QUERY

DRMEDIA
Active-data pool volumes MOVE DRMEDIA and QUERY DRMEDIA can process active-data pool volumes
to be processed except those in the MOUNTABLE state. You can limit processing to specified
active-data pools.

The default at installation: None

To specify the default: SET DRMACTIVEDATASTGPOOL

To override the default: ACTIVEDATASTGPOOL parameter on MOVE DRMEDIA,

QUERY DRMEDIA, or PREPARE
Executable commands file You can use MOVE DRMEDIA or QUERY DRMEDIA to generate executable
name commands and store them in a file.

The default file name at installation: None

To set a default: SET DRMCMDFILENAME. For example, enter:

set drmcmdfilename c:\drm\orm\[Link]

To override the default: CMDFILENAME parameter on MOVE DRMEDIA or QUERY

DRMEDIA

Chapter 32. Disaster recovery manager 1009

Table 80. Defaults for offsite recovery media management (continued)
Process Default
Location name for volumes MOVE DRMEDIA generates a location name for volumes that move to the
that move to the NOTMOUNTABLE state.
NOTMOUNTABLE state
The default at installation: NOTMOUNTABLE

To change the default: SET DRMNOTMOUNTABLENAME

For example, to specify a location named LOCAL, enter:

set drmnotmountablename local
Location name for volumes MOVE DRMEDIA generates a location name for volumes that are changing from
that move to the COURIER NOTMOUNTABLE to COURIER or from VAULTRETRIEVE to COURIERRETRIEVE.
or COURIERRETRIEVE
state The default at installation: COURIER

To change the default: SET DRMCOURIERNAME

For example, to specify a courier named Joe's Courier Service, enter:

set drmcouriername "Joe’s Courier Service"
Reading labels of checked To determine whether DRM reads the sequential media labels of volumes that are
out volumes checked out with MOVE DRMEDIA.

The default at installation: DRM reads the volume labels.

To change the default: SET DRMCHECKLABEL

For example, to specify that DRM should not read the volume labels, enter:
set drmchecklabel no
Expiration period of a A database backup series (full plus incremental and snapshot) is eligible for
database backup series expiration if all of these conditions are true:
v The volume state is VAULT or the volume is associated with a device type of
SERVER (for virtual volumes).
v It is not the most recent database backup series.
v The last volume of the series exceeds the expiration value, number of days since
the last backup in the series.

The default at installation: 60 days

To change the default: SET DRMDBBACKUPEXPIREDAYS

For example, to set the expiration value to 30 days, enter:

set drmdbbackupexpiredays 30
Whether to process copy At installation, MOVE DRMEDIA and QUERY DRMEDIA will not process copy
storage pool and active-data storage pool or active-data pool volumes that are associated with a device type of
pool volumes of the FILE FILE.
device type
The default at installation: Copy storage pool and active-data pool volumes of the
FILE device type are not processed

To change the default: SET DRMFILEPROCESS

To allow processing, enter:

set drmfileprocess yes

1010 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Table 80. Defaults for offsite recovery media management (continued)
Process Default
Vault Name MOVE DRMEDIA uses the vault name to set the location of volumes that are moving
from the COURIER state to the VAULT state

The default at installation: The vault name is set to VAULT.

To change the default: SET DRMVAULTNAME

For example, to specify the vault name as IRONVAULT, the contact name as J.
SMITH, and the telephone number as 1-555-000-0000, enter:
set drmvaultname "Ironvault, J. Smith, 1-555-000-0000"

Specifying recovery instructions for your site

The plan file that DRM generates is a template that contains information, including
commands for recovering the database, that might not apply to your replacement
systems or to your particular recovery scenario. To modify the plan or to store
additional instructions that you will need during recovery from an actual disaster,
use the [Link] stanzas.

About this task

Tip: Enter your site-specific information in the stanzas when you first create the
plan file or after you test it.

Enter your instructions in flat files that have the following names:
v [Link]
v [Link]
v [Link]
v [Link]
v [Link]

Note: The files created for the recovery instructions must be physical sequential
files.
[Link]
Include information such as administrator names, telephone numbers, and
location of passwords. For example:
Recovery Instructions for Tivoli Storage Manager Server ACMESRV on system ZEUS
Joe Smith (wk 002-000-1111 hm 002-003-0000): primary system programmer
Sally Doe (wk 002-000-1112 hm 002-005-0000): primary recovery administrator
Jane Smith (wk 002-000-1113 hm 002-004-0000): responsible manager

Security Considerations:
Joe Smith has the password for the Admin ID ACMEADM. If Joe is unavailable,
you need to either issue SET AUTHENTICATION OFF or define a new
administrative user ID at the replacement Tivoli Storage Manager server console.

[Link]
Include information such as the offsite vault location, courier name, and
telephone numbers. For example:

Chapter 32. Disaster recovery manager 1011

 Our offsite vault location is Ironvault, Safetown, AZ.
The phone number is 1-800-000-0008. You need to contact them directly
to authorize release of the tapes to the courier.
The name of the courier is Fred Harvey. You can contact him at 1-800-444-0000.
Since our vault is so far away, be sure to give the courier a list
of both the database backup, copy storage pool volumes, and active-data
storage pool volumes required. Fred is committed to returning these volumes
to us in less than 12 hours.

[Link]
Include the following installation information:
Server installation and the location of installation volumes and the
license number.
For example:
You will need to reinstall the Tivoli Storage Manager server and administrative
client after installing the Windows operating system.
The install volume for the Tivoli Storage Manager server is INS001. If that is
lost, you will need to contact Copy4You Software, at 1-800-000-0000, and obtain
a new copy. Another possibility is the local IBM Branch office at 555-7777.

A sample file, [Link], is shipped with DRM. You may want to copy
[Link] into your [Link] file to
supplement your installation-specific instructions.
- Obtain a workstation with at least the following:

o 16MB of memory (32MB recommended)

o CD-ROM device

- Install Microsoft Windows NT 4.0 or later

- Install the Tivoli Storage Manager Server and Tivoli Storage Manager
Administrative Client

o It is recommended that you have at least 100MB of disk storage

o The installation directory to specify on the Setup Type screen
can be found in the [Link] stanza within the
disaster recovery plan file
o It is recommended that the Tivoli Storage Manager Server be brought
up to the most current maintenance level. If this is not possible,
then the minimum acceptable maintenance level is the level used
to create the disaster recovery plan file. The level used to
create the disaster recovery plan file can be found in the
[Link] stanza within the disaster recovery plan
file.

[Link]
Include information about how to recover the database and about how
much hardware space requirements. For example:
You will need to find replacement disk space for the server database. We
have an agreement with Joe Replace that in the event of a disaster, he
will provide us with disk space.

[Link]
Include information on primary storage pool recovery instructions. For
example:

1012 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 Do not worry about the archive storage pools during this disaster recovery.
Focus on migration and backup storage pools.
The most important storage pool is XYZZZZ.

Specifying information about your server and client node machines

You need information about your server machine to rebuild its replacement. You
also need information about client node machines to rebuild or restore them.

About this task

Tip: The plan file that DRM generates is a template that contains information,
including commands for recovering the database, that might not apply to your
replacement systems or to your particular recovery scenario. To modify the plan or
to store additional instructions that you will need during recovery from an actual
disaster, use the [Link] stanzas. Enter your site-specific
information in the stanzas when you first create the plan file or after you test it.

Use the following procedure to specify information about server and client
machines and to store it in the server database:

Procedure
1. Specify server machine information by issuing the DEFINE MACHINE
command with ADSMSERVER=YES. For example, to define machine MACH22
in building 021, 2nd floor, in room 2929, with a priority of 1, enter the
following command:
define machine tsm1 adsmserver=yes priority=1
2. Specify the client node location and business priority by issuing the DEFINE
MACHINE command. For example, to define machine MACH22 in building
021, 2nd floor, in room 2929, with a priority of 1, enter:
define machine mach22 building=021 floor=2 room=2929 priority=1
3. Associate one or more client nodes with a machine by issuing the DEFINE
MACHNODEASSOCIATION command. Use this association information to
identify client nodes on machines that were destroyed. You should restore the
file spaces associated with these nodes. For example, to associate node
CAMPBELL with machine MACH22, enter:
define machnodeassociation mach22 campbell
4. To query machine definitions, issue the QUERY MACHINE command. See the
example, in “Client recovery scenario” on page 1038.
5. To add machine characteristics and recovery instructions to the database, issue
the INSERT MACHINE command. You must first query the operating system
to identify the characteristics for your client machine.
You can add the information manually or use a Microsoft VBScript command
procedure. A sample program is shipped with DRM.
v Add information manually:
The following partial output is from a query on an AIX client machine.
--1 Host Name: mach22 with 256 MB Memory Card
--- 256 MB Memory Card
---
--4 Operating System: AIX Version 4 Release 3
---
--- Hardware Address: [Link]x:a8:6a:46

Chapter 32. Disaster recovery manager 1013

 Specify characteristics and recovery instructions one line at a time with
separate INSERT MACHINE commands:
– To save the first line (Host Name: mach22 with 256 MB Memory Card) as
line 1 and to save the fourth line (Operating System: AIX Version 4
Release 3) as line 2 for machine MACH22, issue the following commands:
insert machine mach22 1 characteristics="Host Name: mach22 with
256 MB Memory Card"

insert machine mach22 2 characteristics="Operating System:

AIX Version 4 Release 3"
– To specify recovery instructions for your client machine, issue the
following command:
insert machine mach22 1 -
recoveryinstructions="Recover this machine for accounts
receivable dept."
v Add Information Using a Microsoft VBScript Command Procedure
To help automate the adding of client machine information, a sample
VBScript command procedure named [Link] is shipped with DRM. The
following example shows how to use a local program to add machine
characteristics or recovery instructions:
a. The file, [Link], from the AIX client machine is transferred to a
Windows machine that has an administrative command-line client
installed. The file is then processed by the a sample VBScript command
procedure, which builds a macro of INSERT commands (one command
for each line in the file). The following screens show an example of a
procedure named [Link] that is used to insert machine
characteristics.

1014 IBM Tivoli Storage Manager for Windows: Administrator's Guide

’***************************************************************************
’ Tivoli Disaster Recovery Manager for Windows NT/2000 Sample Script
’
’ Read machine characteristics from an input file and build an output file
’ that is a TSM macro. The TSM macro contains statements which are
’ TSM commands to insert client machine information into the ADSM server
’ database. The TSM macro is used with the TSM administrative client.
’
’ Invoke with:
’ cscript [Link] machinename inputmachinefilename outputmacrofilename
’ where:
’ machinename is the name of a machine that has previously
’ been defined to the TSM server with the
’ DEFINE MACHINE command
’ inputmachinefilename is the name of the input file which contains
’ the client machine characteristics. This file
’ would typically be built on the client machine
’ then the file would be transferred to the
’ Windows machine where the TSM Administrative
’ client is installed.
’ outputmacrofilename is the name of the output file in an existing
’ directory which will be the TSM macro. The
’ TSM macro will consist of a series of commands
’ to insert machine characteristics into the TSM
’ server database. For example:
’
’ INSERT MACHINE mch1 n characteristics=’xxx...’
’
’ where:
’ n represents the sequence number
’ this line will have in the
’ TSM server database
’ ’xxx...’ represents a single line from
’ the input file
’
’ NOTE: The maximum length of a line of machine
’ characteristics is 1024
’
’ Example usage:
’ cscript [Link] mch1 c:\client1\[Link] c:\client1\[Link]
’***************************************************************************
Dim args
Dim MACHINENAME, INFILE, OUTFILE
dim fso
dim fo, fi
dim SEQUENCE
Dim CRLF

Chapter 32. Disaster recovery manager 1015

 CRLF = Chr(13) & Chr(10)
Const ForReading = 1, ForWriting = 2

’***************************************************************************
’ Get input arguments: MACHINENAME =machinename
’ INFILE =inputmachinefilename
’ OUTFILE =outputmacrofilename
’***************************************************************************

set args = [Link]

If [Link] < 3 Then

[Link] _
"usage: cscript [Link] machinename inputmachinefilename
outputmacrofilename" & CRLF & _
"example: cscript [Link] mch1 c:\client1\[Link]
c:\client1\[Link]"
[Link](1)
Else
MACHINENAME = [Link](0)
INFILE = [Link](1)
OUTFILE = [Link](2)
End if

Set fso = CreateObject("[Link]")

’***************************************************************************
’ Create the TSM macro file.
’***************************************************************************

Set fo = [Link](OUTFILE, ForWriting, True)

[Link] "Creating TSM macro file: " & OUTFILE

’***************************************************************************
’ Place a TSM command in the TSM macro to delete any existing machine
’ characteristics for this machine from the TSM server database.
’***************************************************************************

[Link] "delete machine " & MACHINENAME & " type=characteristics"

’***************************************************************************
’ Read a line from the input machine characteristics file, add the TSM
’ command to insert the line of machine characteristics into the TSM server
’ database, and write the result to the output TSM macro.
’***************************************************************************

SEQUENCE = 1

Set fi = [Link](INFILE, ForReading, False)

Do While [Link] <> True

INLINE = [Link]
[Link] "insert machine " & MACHINENAME & " " & SEQUENCE & "
char=’" & INLINE &"’"
SEQUENCE = SEQUENCE + 1

Loop

’***************************************************************************
’ Close the files.
’***************************************************************************

[Link]
[Link]

The [Link] VBScript command procedure is run:

cscript [Link] acctsrecv [Link] [Link]

1016 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 Then the macro is run to load the data into the database.
b. To add the machine characteristics, start the administrative client and run
the macro. For example:
> dsmadmc -id=xxx -pw=xxx macro [Link]
You can view your machine characteristics by issuing the QUERY
MACHINE command with FORMAT=CHARACTERISTICS parameter.
c. To specify recovery instructions for your client machine, you can use a
similar VBScript command procedure but with the
RECOVERYINSTRUCTIONS parameter.

Specifying recovery media for client machines

Follow these steps to specify the bootable media needed to reinitialize or reinstall
an operating system on a client machine and to associate machines with media.
You can also associate non-executable media such as application user guides with
client machines.

Procedure
1. Define the bootable media. For example, define the media named
TELLERWRKSTNIMAGE which is for AIX Version 4.3, contains the required
volumes named AIX001, AIX002, and AIX003, and is located in Building 21.
define recoverymedia tellerwrkstnimage type=boot
volumenames=aix001,aix002,aix003 product="AIX 4.3"
location="Building 21"
You should define the recovery media after a client machine configuration
changes. For example, after you have installed a new level of AIX on a client
machine and created a bootable image using mksysb, issue the DEFINE
RECOVERYMEDIA command to define the new mksysb volumes.
To query your recovery media definitions, issue the QUERY
RECOVERYMEDIA command with the FORMAT=DETAILED parameter.
2. Associate one or more machines with recovery media. Use the association
information to identify the boot media to use in the replacement machines. For
example, to associate machine MACH255 with recovery media
TELLERWRKSTNIMAGE, issue the following command:
define recmedmachassociation tellerwrkstnimage mach255
3. When the boot media is moved offsite, update its location. For example, to
update the location of boot media TELLERWRKSTNIMAGE to the offsite
location IRONVAULT, issue the following command:
update recoverymedia tellerwrkstnimage location=ironvault

Results

You can define media that contain softcopy manuals that you would need during
recovery. For example, to define a CD-ROM containing the AIX 5.1 manuals that
are on volume CD0001, enter:
define recoverymedia aix51manuals type=other volumes=cd0001
description="AIX 5.1 Bookshelf"

Chapter 32. Disaster recovery manager 1017

Creating and storing the disaster recovery plan
You can create a disaster recovery plan file and store the file locally or on another
server.

About this task

The recovery plan contains the following information:

v The recovery procedure
v A list of required database volumes, copy storage pool volumes, and active-data
pool volumes, devices to read those volumes, and database and recovery log
space requirements
v Copies of the server options file, device configuration file, and volume history
information file
v Commands for performing database recovery and primary storage pool recovery
v Commands for registering licenses
v Instructions that you define
v Machine and recovery media information that you define

For details about the recovery plan file, see “The disaster recovery plan file” on
page 1048.

DRM creates one copy of the disaster recovery plan file each time you issue the
PREPARE command. You should create multiple copies of the plan for safekeeping.
For example, keep copies in print, on CD, on disk space that is located offsite, or
on a remote server.

Before creating a disaster recovery plan, back up your storage pools then backup
the database. See “Backing up primary storage pools” on page 883 and “Backing
up the server database” on page 870 for details about these procedures.

If you manually send backup media offsite, see “Moving copy storage pool and
active-data pool volumes offsite” on page 1026. If you use virtual volumes, see
“Using virtual volumes to store data on another server” on page 732.

When your backups are both offsite and marked offsite, you can create a disaster
recovery plan.

You can use the Tivoli Storage Manager scheduler to periodically run the
PREPARE command (see Chapter 19, “Automating server operations,” on page
619).

Tips:
v The plan file that DRM generates is a template that contains information,
including commands for recovering the database, that might not apply to your
replacement systems or to your particular recovery scenario. To modify the plan
or to store additional instructions that you will need during recovery from an
actual disaster, use the [Link] stanzas. Enter your
site-specific information in the stanzas when you first create the plan file or after
you test it.
v DRM creates a plan that assumes that the latest database full plus incremental
series would be used to restore the database. However, you may want to use
DBSNAPSHOT backups for disaster recovery and retain your full plus

1018 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 incremental backup series on site to recover from possible availability problems.
In this case, you must specify the use of DBSNAPSHOT backups in the
PREPARE command. For example:
prepare source=dbsnapshot

Backup of Centera storage pools is not supported so commands for Centera

primary storage pools should not be included in the recovery plan file. To work
around this, do one of the following:
v Use the SET DRMPRIMSTGPOOL command or the PRIMSTGPOOL parameter
with the PREPARE command to specify the names of the primary storage pools
that will be eligible for DRM PREPARE processing and do not include Centera
storage pools. If a specified pool name contains a wildcard, ensure that the
expanded name will not match the name of any Centera storage pool defined to
the Tivoli Storage Manager server.
v Or, if a Centera storage pool is included in the primary pool list when a
PREPARE command is executed, update the recovery plan file that it generated.
Delete or comment out the commands for the Centera storage pool that may be
included in the following stanzas:
– [Link]
– [Link]
– [Link]

Storing the disaster recovery plan locally

When you create a recovery plan file but do not specify a device class, the file is
stored locally in a file system. If you store the file locally, you can specify a storage
location.

About this task

For example, to store the recovery plan file locally in the c:\Program
Files\Tivoli\TSM\server2\recplans\ directory, enter:
prepare planprefix=c:\Program Files\Tivoli\TSM\server2\recplans\

What to do next

Recovery plan files that are stored locally are not automatically expired. You
should periodically delete down-level recovery plan files manually. DRM appends
to the file name the date and time ([Link]). For example:
c:\Program Files\Tivoli\TSM\server2\recplans\20000925.120532

Storing the disaster recovery plan on a target server

When you create a recovery plan file and specify a device class, the file is stored
on a target server.

About this task

Storing recovery plan files on a target server provides the following:

v A central repository on a target server for recovery plan files
v Automatic expiration of plan files
v Query capabilities that display information about recovery plan files and the
ability to display the contents of a recovery plan file located on a target server
v Recovery plan file retrieval from a target server

Chapter 32. Disaster recovery manager 1019

 Set up the source and target servers and define a device class a device type of
SERVER (see “Setting up source and target servers for virtual volumes” on page
734 for details). For example, assume a device class named TARGETCLASS is
defined on the source server where you create the recovery plan file. Then to
create the plan file, enter:
prepare devclass=targetclass

The recovery plan file is written as an object on the target server, and a volume
history record is created on the source server. For more about recovery plan files
that are stored on target servers, see “Displaying information about recovery plan
files” on page 1022.

Disaster recovery plan environmental considerations

The environment (for example, directory structure and location of the server
executable and enrollment certificates) for a recovery using the disaster recovery
plan file is assumed to be the same as the original environment. If the recovery
environment is not the same, then you must edit the plan file to account for the
changes in the environment.

If Tivoli Storage Manager is installed in the c:\Program Files\Tivoli\TSM directory,

the files are stored as follows:
v Server executable, message and help files, and enrollment certificates:
c:\Program Files\Tivoli\TSM\server\[Link]
c:\Program Files\Tivoli\TSM\server\[Link]
c:\Program Files\Tivoli\TSM\server\[Link]
c:\Program Files\Tivoli\TSM\server\*.lic
v Administrative command-line client:
c:\Program Files\Tivoli\TSM\saclient\[Link]

Each instance of the server has a unique set of files. For example, you might see
the following in this instance-specific directory:
v Server options file, disk file, database and log paths, and storage pool volumes:
c:\Program Files\Tivoli\TSM\server2\[Link]
c:\Program Files\Tivoli\TSM\server2\[Link]
c:\\Program Files\Tivoli\TSM\server2\db1
c:\ \Program Files\Tivoli\TSM\server2\activlog
c:\ \Program Files\Tivoli\TSM\server2\archlog
c:\Program Files\Tivoli\TSM\server2\[Link]
The database and log paths, and storage pool volumes could also be in a different
directory. For example, you might see:
c:\c:\Program Files\Tivoli\TSM\server2\stg\db1
c:\ c:\Program Files\Tivoli\TSM\server2\stg\activelog
c:\Program Files\Tivoli\TSM\server2\stg\[Link]

Files that typically reside in an instance-specific directory (that is, [Link],

[Link]) and storage pool volumes may instead reside in the same directory in
which [Link], resides (c:\Program Files\Tivoli\TSM\server). In this case, the
directory containing [Link] would also be referred to as the instance-specific
directory.

1020 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 When the disaster recovery plan is created, information about the server
environment is used in the stanzas within the plan file. This environmental
information includes the location of [Link], the location of the disk formatting
utility, the instance-specific directory, the directories storage pool volumes, and so
on. During a recovery, it is assumed that the same server environment exists.

Additionally, the plan file itself will reside in a directory that you may have
specified or it may reside in the default directory (which is the instance–specific
directory). For example, if you specified the disaster recovery plan file prefix
c:\Program Files\Tivoli\TSM\server2\prepare\, you might see the following:
c:\Program Files\Tivoli\TSM\server2\prepare\20000925.120532

The disaster recovery plan file prefix specified (or the instance–specific directory if
no disaster recovery plan file prefix was specified) is also used in the stanzas
within the plan file. During a recovery, when the plan file has been split into
individual files, it is assumed that these individual files will reside in this same
directory.

To summarize, the environment for a recovery using the disaster recovery plan file
is assumed to be the same as the original environment which includes:
v The directory structure and location of the server executable and enrollment
certificates (for licensing)
v The directory structure and location of the administrative command line client
v The directory structure for server instance-specific files
v The directory structure for the database path, active log, and archive log
directories and storage pool volumes.
v The directory structure and the files created when the plan file was split into
multiple files, such as the following based on the earlier plan file example (the
following is not the entire output):
c:\Program Files\Tivoli\TSM\server2\prepare\[Link]
c:\Program Files\Tivoli\TSM\server2\prepare\[Link]
c:\Program Files\Tivoli\TSM\server2\prepare\[Link]
c:\Program Files\Tivoli\TSM\server2\prepare\[Link]

If the recovery environment is not the same, then you must edit the plan file to
account for the changes in the environment.

To help understand where these various directories and expected locations for
executables are used within the plan file, see “Example disaster recovery plan file”
on page 1054 and you will see the following usage:

Usage Directory
Server executable c:\Program Files\Tivoli\TSM\server
Enrollment certificates c:\Program Files\Tivoli\TSM\server
(licensing)
Administrative command line c:\Program Files\Tivoli\TSM\saclient
client
Disk formatting utility c:\Program Files\Tivoli\TSM\console
Instance-specific files c:\Program Files\Tivoli\TSM\server2
Storage pool volumes c:\Program Files\Tivoli\TSM\server2\stg
Plan file location c:\Program Files\Tivoli\TSM\server2\prepare

Chapter 32. Disaster recovery manager 1021

 Usage Directory
Individual files split out from c:\Program Files\Tivoli\TSM\server2\prepare
plan

Managing disaster recovery plan files stored on target servers

The following sections describe how to view information about disaster recovery
plans stored on a target server and view their contents. It also describes how to
direct the contents of a disaster recovery plan file to another file and how to delete
volume history records of the recovery plan files.

Displaying information about recovery plan files

You can display information about recovery plan files from the server that created
the files (the source server) or from the server on which the files are stored (the
target server).

Procedure
1. From the source server: Issue QUERY RPFILE the command with the
DEVCLASS parameter that was used on the PREPARE command. Specify the
type of database backups that were assumed when the plan was created (either
full plus incremental or snapshot). For example, to display a list of all recovery
plan files that have been saved for the source server on any target servers and
created assuming snapshot database backups, enter:
query rpfile devclass=* source=dbsnapshot
2. You can also issue the QUERY VOLHISTORY command to display a list of
recovery plan files for the source server. Specify recovery plan files that were
created assuming either full plus incremental database backups (TYPE=RPFILE)
or database snapshot backups (TYPE=RPFSNAPSHOT). For example:
query volhistory type=rpfile
3. From the target server: Issue a QUERY RPFILE command that specifies the
node name associated with the server or servers that prepared the plan. For
example, to display a list of all recovery plan files that have been saved in the
target server, enter:
query rpfile nodename=*

Displaying the contents of a recovery plan file

From the server that created the recovery plan file (the source server) or from the
server on which the file is stored (the target server), you can display the contents
of that file that was saved as on object on the target server.

About this task

For an example of the contents of a recovery plan file, see “The disaster recovery
plan file” on page 1048. You cannot issue the commands shown below from a
server console. An output delay can occur if the plan file is located on tape.

Procedure
v From the source server: Issue the following command for a recovery plan file
created on September 1, 2000 at 4:39 a.m. with the device class TARGETCLASS:
query rpfcontent marketing.20000901.043900 devclass=targetclass

1022 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 v From the target server: Issue the following command for a recovery plan file
created on August 31,2000 at 4:50 a.m. on a source server named MARKETING
whose node name is BRANCH8:
query rpfcontent marketing.20000831.045000 nodename=branch8

Restoring a recovery plan file

To restore a recovery plan file, use the QUERY RPFCONTENT command and
direct the output to a file. You can issue the command from the server that created
the files (the source server) or from the server on which the files are stored (the
target server). To see a list of recovery plan file names, issue the QUERY RPFILE
command.

About this task

For example, a recovery plan file named marketing.20000831.045000 was created

using the device class of TARGETCLASS and on a source server whose node name
at the target server is BRANCH8. You want to restore the file and direct the output
to [Link]:

Procedure
v From the source server: Issue the following command:
query rpfcontent marketing.20000831.045000
devclass=targetclass > [Link]
v From the target server: Issue the following command:
query rpfcontent marketing.20000831.045000
nodename=branch8 > [Link]

What to do next

To display a list of recovery plan files, use the QUERY RPFILE command. See
“Displaying information about recovery plan files” on page 1022 for more
information.

Expiring recovery plan files automatically

You can set DRM to expire recovery plan files stored on a target server a certain
number of days after they are created.

About this task

All recovery plan files that meet the criteria are eligible for expiration if both of the
following conditions exist:
v The last recovery plan file of the series is over 90 days old.
v The recovery plan file is not associated with the most recent backup series. A
backup series consists of a full database backup and all incremental backups that
apply to that full backup. Another series begins with the next full backup of the
database.

Expiration applies to plan files based on both full plus incremental and snapshot
database backups. Note, however, that expiration does not apply to plan files
stored locally. See “Storing the disaster recovery plan locally” on page 1019.

Chapter 32. Disaster recovery manager 1023

 Procedure

To set up expiration, issue the SET DRMRPFEXPIREDAYS command. The default

value is 60 days. For example, to change the time to 90 days, enter:
set drmrpfexpiredays 90

Deleting recovery plan files manually

You can delete volume history records containing information about recovery plan
file objects.

About this task

When the records are deleted from the source server and the grace period is
reached, the objects are deleted from the target server The record for the latest
recovery plan file is not deleted.

Procedure

To delete recovery plan files, issue the DELETE VOLHISTORY command For
example, to delete records for recovery plan files that were created on or before
08/30/2000 and assuming full plus incremental database backup series, enter the
following command:
delete volhistory type=rpfile todate=08/30/2000

To limit the operation to recovery plan files that were created assuming database
snapshot backups, specify TYPE=RPFSNAPSHOT.

Moving backup media

To recover from a disaster you need database backup volumes, copy storage pool
volumes, and, optionally, active-data pool volumes. To stay prepared for a disaster,
you need to perform a number of daily tasks.

Before you begin

Task Required Privilege Class
Send copy storage pool and active-data pool Unrestricted storage or operator
volumes offsite and back onsite

Procedure
1. Move new backup media offsite and update the database with their locations.
See “Moving copy storage pool and active-data pool volumes offsite” on page
1026 for details.
2. Return expired or reclaimed backup media onsite and update the database with
their locations. See “Moving copy storage pool and active-data pool volumes
on-site” on page 1028 for details.
3. Offsite recovery media management does not process virtual volumes. To
display all virtual copy storage pool, active-data pool, and database backup
volumes that have their backup objects on the remote target server, issue the
QUERY DRMEDIA command. For example, enter the following command.
query drmedia * wherestate=remote

1024 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 Results

Offsite recovery media management does not move or display any two-sided
volumes that have a REMOVABLEFILE device type.

The disaster recovery plan includes the location of copy storage pool volumes and
active-data pool volumes. The plan can provide a list of offsite volumes required to
restore a server.

The following diagram shows the typical life cycle of the recovery media:

Offsite
Onsite
In Transit Vault

Storage Hierarchy

COURIER

Backup Active-
storage data
pool pool

NOTMOUNTABLE

MOUNTABLE

KUPEXPIREDAYS
VAULT
EDELAY

Backup
database
US

Private Scratch
RE

BAC

r/w
DB
M
DR

Scratch
VAULTRETRIEVE

ONSITERETRIEVE

COURIERRETRIEVE

Figure 116. Recovery media life cycle

DRM assigns the following states to volumes. The location of a volume is known
at each state.
MOUNTABLE
The volume contains valid data, and Tivoli Storage Manager can access it.
NOTMOUNTABLE
The volume contains valid data and is onsite, but Tivoli Storage Manager
cannot access it.
COURIER
The volume contains valid data and is in transit to the vault.

Chapter 32. Disaster recovery manager 1025

 VAULT
The volume contains valid data and is at the vault.
VAULTRETRIEVE
The volume, which is located at the offsite vault, no longer contains valid
data and is to be returned to the site. For more information about
reclamation of offsite copy storage pool volumes and active-data pool
volumes, see “Reclamation of offsite volumes” on page 375. For
information on expiration of database backup volumes, see step 1 on page
1028.
COURIERRETRIEVE
The volume no longer contains valid data and is in the process of being
returned by the courier.
ONSITERETRIEVE
The volume no longer contains valid data and has been moved back to the
onsite location. The volume records of database backup, scratch copy
storage pool volumes, and scratch active-data pool volumes are deleted
from the database. For private copy storage pool volumes and active-data
pool volumes, the access mode is updated to READWRITE.

Moving copy storage pool and active-data pool volumes

offsite
After you created the backup copies of your primary storage pools and database,
you can send your backup media offsite. To send media offsite, mark the volumes
as unavailable to Tivoli Storage Manager and give them to the courier.

Procedure

Complete the following steps to identify the database backup, copy storage pool,
and active-data pool volumes and move them offsite:
1. Identify the copy storage pool, active-data pool, and database backup volumes
to be moved offsite For example, issue the following command:
query drmedia * wherestate=mountable
DRM displays information similar to the following output:
Volume Name State Last Update Automated
Date/Time LibName
--------------- ---------------- ------------------- -----------------
TPBK05 Mountable 01/01/2000 [Link] LIBRARY
TPBK99 Mountable 01/01/2000 [Link] LIBRARY
TPBK06 Mountable 01/01/2000 [Link] LIBRARY

2. Indicate the movement of volumes whose current state is MOUNTABLE. For

example, issue the following command:
move drmedia * wherestate=mountable

Restriction: Do not run the MOVE DRMEDIA and BACKUP STGPOOL commands
concurrently. Ensure that the storage pool backup processes are complete before
you issue the MOVE DRMEDIA command.
For all volumes in the MOUNTABLE state, DRM does the following:
v Updates the volume state to NOTMOUNTABLE and the volume location
according to the SET DRMNOTMOUNTABLENAME. If this command is not
issued, the default location is NOTMOUNTABLE.

1026 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 v For a copy storage pool volume or active-data pool volume, updates the
access mode to unavailable.
v For a volume in an automated library, checks the volume out of the library.
a. During checkout processing, SCSI libraries request operator intervention. To
bypass these requests and eject the cartridges from the library, first issue the
following command:
move drmedia * wherestate=mountable remove=no
b. Access a list of the volumes by issuing the following command:
query drmedia wherestate=notmountable

From this list identify and remove the cartridges (volumes) from the library.
c. For the 349X library type, if the number of cartridges to be checked out of
the library is greater than the number of slots in the I/O station, you can
define a high capacity area in your library. Then, use the following
command to eject the cartridges to the high capacity area, rather than to the
I/O station:
move drmedia * wherestate=mountable remove=bulk
3. Send the volumes to the offsite vault. Issue the following command to have
DRM select volumes in the NOTMOUNTABLE state:
move drmedia * wherestate=notmountable

For all volumes in the NOTMOUNTABLE state, DRM updates the volume state
to COURIER and the volume location according to the SET
DRMCOURIERNAME. If the SET command is not yet issued, the default
location is COURIER. For more information, see “Specifying defaults for offsite
recovery media management” on page 1009
4. When the vault location confirms receipt of the volumes, issue the MOVE
DRMEDIA command in the COURIER state. For example:
move drmedia * wherestate=courier
For all volumes in the COURIER state, DRM updates the volume state to
VAULT and the volume location according to the SET DRMVAULTNAME command.
If the SET command is not yet issued, the default location is VAULT. For more
information, see “Specifying defaults for offsite recovery media management”
on page 1009.
5. Display a list of volumes that contain valid data at the vault. Issue the
following command:
query drmedia wherestate=vault

DRM displays information similar to the following output:

Volume Name State Last Update Automated
Date/Time LibName
----------------- -------------- ------------------- -----------------
TAPE0P Vault 01/05/2000 [Link]
TAPE1P Vault 01/05/2000 [Link]
DBT02 Vault 01/05/2000 [Link]
TAPE3S Vault 01/05/2000 [Link]

6. If you do not want to step through all the states, you can use the TOSTATE
parameter on the MOVE DRMEDIA command to specify the destination state. For
example, to change the volumes from NOTMOUNTABLE state to VAULT state,
issue the following command:
move drmedia * wherestate=notmountable tostate=vault

Chapter 32. Disaster recovery manager 1027

 For all volumes in the NOTMOUNTABLE state, DRM updates the volume state
to VAULT and the volume location according to the SET DRMVAULTNAME
command. If the SET command is not yet issued, the default location is VAULT.

Results

See “Preparing for disaster recovery” on page 1032 for an example that
demonstrates sending server backup volumes offsite using MOVE DRMEDIA and QUERY
DRMEDIA commands.

Moving copy storage pool and active-data pool volumes

on-site
Use the following procedure to expire the non-virtual database backup volumes
and return the volumes back on-site for reuse or disposal.

Procedure
1. To specify the number of days before a database backup series is expired, issue
the SET DRMDBBACKUPEXPIREDAYS command The following example sets
the number of days to 30.
set drmdbbackupexpiredays 30

To ensure that the database can be returned to an earlier level and database
references to files in the copy storage pool or active-data pool are still valid,
specify the same value for the REUSEDELAY parameter in your copy storage
pool and active-data pool definitions. If copy storage pools or active-data pools
managed by DRM have different REUSEDELAY values, set the
DRMDBBACKUPEXPIREDAYS value to the highest REUSEDELAY value.
A database backup volume is considered eligible for expiration if all of the
following conditions are true:
v The age of the last volume of the series has exceeded the expiration value.
This value is the number of days since the last backup in the series. At
installation, the expiration value is 60 days. To override this value, issue the
SET DRMDBBACKUPEXPIREDAYS command.
v For volumes that are not virtual volumes, all volumes in the series are in the
VAULT state.
v The volume is not part of the most recent database backup series.
Database backup volumes that are virtual volumes are removed during
expiration processing. This processing is started manually by issuing the
EXPIRE INVENTORY command or automatically through the EXPINTERVAL
option setting specified in the server options file.
2. Move a copy storage pool volume or an active-data pool volume on-site for
reuse or disposal. A copy storage pool volume or an active-data pool volume
can be moved on-site if it has been EMPTY for at least the number of days
specified with the REUSEDELAY parameter on the DEFINE STGPOOL
command. A database backup volume can be moved on-site if the database
backup series is EXPIRED according to the rules outlined in step 1. To
determine which volumes to retrieve, issue the following command:
query drmedia * wherestate=vaultretrieve

The server dynamically determines which volumes can be moved back on-site.
When you issue QUERY DRMEDIA WHERESTATE=VAULTRETRIEVE, the
field Last Update Date/Time in the output will contain the data and time that
the state of the volume was moved to VAULT, not VAULTRETRIEVE. Because

1028 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 the server makes the VAULTRETRIEVE determination dynamically, issue
QUERY DRMEDIA WHERESTATE=VAULTRETRIEVE without the
BEGINDATE, ENDDATE, BEGINTIME or ENDTIME parameters. Doing so will
ensure that you identify all volumes that are in the VAULTRETRIEVE state.
3. After the vault location acknowledges that the volumes have been given to the
courier, issue the MOVE DRMEDIA command.
move drmedia * wherestate=vaultretrieve
The server does the following for all volumes in the VAULTRETRIEVE state:
v Change the volume state to COURIERRETRIEVE.
v Update the location of the volume according to what is specified in the SET
DRMCOURIERNAME command. For more information, see “Specifying
defaults for offsite recovery media management” on page 1009.
4. When the courier delivers the volumes, acknowledge that the courier has
returned the volumes on-site. Issue the following command:, by issuing:
move drmedia * wherestate=courierretrieve
The server does the following for all volumes in the COURIERRETRIEVE state:
v Moves the volumes on-site where they can be can be reused or disposed of.
v Deletes the database backup volumes from the volume history table.
v For scratch copy storage pool volumes or active-data pool volumes, deletes
the record in the database. For private copy storage pool volumes or
active-data pool volumes, updates the access to read/write.
5. If you do not want to step through all the states, you can use the TOSTATE
parameter on the MOVE DRMEDIA command to specify the destination state.
For example, to move the volumes from VAULTRETRIEVE state to
ONSITERETRIEVE state, issue the following command:
move drmedia * wherestate=vaultretrieve tostate=onsiteretrieve

The server does the following for all volumes with in the VAULTRETRIEVE
state:
v Moves the volumes on-site where they can be can be reused or disposed of.
v Deletes the database backup volumes from the volume history table.
v For scratch copy storage pool volumes or active-data pool volumes, deletes
the record in the database. For private copy storage pool volumes or
active-data pool volumes, updates the access to read/write.

Managing the Disaster Recovery Manager tasks

To protect data, you can use the Disaster Recovery Manager (DRM) during routine
operations and during disaster recovery.

Before you begin

If IBM Tivoli Storage Manager is set up to use Secure Sockets Layer (SSL) for
client/server authentication, a digital certificate file, [Link], is created as part of
the process. This file includes the server's public key, which allows the client to
encrypt data. The digital certificate file cannot be stored in the server database
because the Global Security Kit (GSKit) requires a separate file in a certain format.
1. Keep backup copies of the [Link] and [Link] files.
2. Regenerate a new certificate file, if both the original files and any copies are
lost or corrupted. For details about this procedure, see “Troubleshooting the
certificate key database” on page 837.

Chapter 32. Disaster recovery manager 1029

 About this task

Ensure that you set up the DRM and perform the daily operations to protect the
database, data, and storage pools.
Setup
1. License DRM by issuing the REGISTER LICENSE command.
2. Ensure that the device configuration and volume history files exist.
3. Back up the storage pools by issuing the BACKUP STGPOOL command.
4. Copy active data to active-data pools by using the COPY ACTIVEDATA
command.
5. Do a full backup the database (for example, a database snapshot
backup) by using the BACKUP DB command.
6. Define site-specific server recovery instructions in the DRM plan file.
7. Describe priority client machines by using the DEFINE MACHINE
command.
8. Generate the disaster recovery plan.
Daily Preparation Operations
Day 1
1. Back up client files.
2. Back up the primary storage pools to copy storage pools.
3. Copy active data from primary storage pools to active-data pools.
4. Back up the database (for example, a database snapshot backup).

Restriction: Ensure that the primary storage pool backup process is

complete before you back up the database.
5. Mark the backup volumes as unavailable to Tivoli Storage Manager.

Restriction: Ensure that the BACKUP STGPOOL command and the BACKUP
DB command are complete before you issue the MOVE DRMEDIA
command.
6. Send the backup volumes and disaster recovery plan file to the vault.
7. Generate the disaster recovery plan.
Day 2
1. Back up client files
2. Back up active and inactive data that is in the primary storage pools to
copy storage pools. Copy the active data that is in primary storage
pools to active-data pools.
3. Back up the database (for example, a database snapshot backup).

Restriction: Ensure that the primary storage pool backup process is

complete before you back up the database.
4. Mark the backup volumes as unavailable to Tivoli Storage Manager.

Restriction: Ensure that the BACKUP STGPOOL command and the BACKUP
DB command are complete before you issue the MOVE DRMEDIA
command.
5. Send the backup volumes and disaster recovery plan file to the vault.
6. Generate the disaster recovery plan.
Day 3

1030 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 1. Automatic storage pool reclamation processing occurs.
2. Back up client files.
3. Back up the active and inactive data that is in primary storage pools to
copy storage pools. Copy the active data that is in primary storage
pools to active-data pools.
4. Back up the database (for example, a database snapshot backup).

Restriction: Ensure that the primary storage pool backup process is

complete before you back up the database.
5. Send the backup volumes and a list of expired volumes to be reclaimed
to the vault.
6. The vault acknowledges receipt of the volumes that were sent on the
previous day.
7. Generate the disaster recovery plan.
Disaster and Recovery
Day 4 The server and the client machines are destroyed.
1. Restore the server by using the latest recovery plan.
2. Identify the top priority client nodes at the disaster site.
3. Restore urgently needed client-system files from active-data pools.
Restore other, less urgently needed client-system files from copy storage
pools.
4. Restore the primary storage pools from copy storage pools.
Attention: Restoring a primary storage pool from an active-data pool
might cause some or all inactive files to be deleted from the database.
Some or all of the inactive files might be deleted if the server
determines that an inactive file must be replaced but cannot find it in
the active-data pool.
5. Move database backup, copy storage pool, and active-data pool
volumes to the vault.
Daily Operations
Day 5
1. Back up client files.
2. Back up the active and inactive data that is in primary storage pools to
copy storage pools. Copy the active data that is in primary storage
pools to active-data pools.
3. Back up the database (for example, a database snapshot backup).
4. Send the backup volumes and a list of expired volumes to be reclaimed
to the vault.
5. Generate the disaster recovery plan.

Chapter 32. Disaster recovery manager 1031

Preparing for disaster recovery
Stay prepared for a disaster by establishing a procedure for disaster recovery by
using the Disaster Recovery Manager (DRM). The steps are performed by the
onsite administrator unless otherwise indicated.

Procedure

Complete the following steps to prepare for disaster recovery:

1. Record the following information in the [Link] stanza
source files:
v Software license numbers
v Sources of replacement hardware
v Any recovery steps specific to your installation
2. Store the following information in the database:
v Server and client node machine information (DEFINE MACHINE, DEFINE
MACHINENODE ASSOCIATION, and INSERT MACHINE commands)
v The location of the boot recovery media (DEFINE RECOVERYMEDIA command)
3. Schedule automatic nightly backups to occur in the following order:
v Primary Storage Pools
– Backup active and inactive data to copy storage pools.
– Copy active data to active-data pools.
v Database
– Back up the database

Restriction: Ensure that the primary storage pool backup process is

complete before you back up the database.
4. Daily, create a list of the previous night's database, copy storage pool, and
active-data pool volumes to be sent offsite:
query drmedia * wherestate=mountable
a. Check the volumes out of the library:
move drmedia * wherestate=mountable

Restriction: Ensure that the BACKUP STGPOOL command and the BACKUP DB
command are complete before you issue the MOVE DRMEDIA command.
b. Send the volumes offsite and record that the volumes were given to the
courier:
move drmedia * wherestate=notmountable
5. Create a recovery plan:
prepare
6. Give a copy the recovery plan file to the courier.
7. Create a list of tapes that contain data that is no longer valid and that should
be returned to the site:
query drmedia * wherestate=vaultretrieve
8. Give the courier the database backup tapes, storage pool backup tapes,
active-data pool tapes, the recovery plan file, and the list of volumes to be
returned from the vault.
9. The courier gives you any tapes that were on the previous day's return from
the vault list.
Update the state of these tapes and check them into the library:
1032 IBM Tivoli Storage Manager for Windows: Administrator's Guide
 move drmedia * wherestate=courierretrieve cmdf=c:\drm\[Link]
cmd="checkin libvol libauto &vol status=scratch"
The volume records for the tapes that were in the COURIERRETRIEVE state
are deleted from the database. The MOVE DRMEDIA command also generates
the CHECKIN LIBVOL command for each tape that is processed in the file
c:\drm\[Link]. For example:
checkin libvol libauto tape01 status=scratch
checkin libvol libauto tape02 status=scratch
...

Note: An administrator can run the MACRO command by specifying

c:\drm\[Link] as the macro name to process the CHECKIN LIBVOL

commands.
> dsmadmc -id=xxxxx -pa=yyyyyy MACRO c:\drm\[Link]
10. The courier takes the database backup tapes, storage pool backup tapes,
active-data pool tapes, the recovery plan, and the list of volumes to return
from the vault.
11. Call the vault and verify that the backup tapes arrived and are secure, and
that the tapes to be returned to the site have been given to the courier.
12. Set the location of the volumes sent to the vault:
move drmedia * wherestate=courier
13. Set the location of the volumes that are given to the courier by the vault:
move drmedia * wherestate=vaultretrieve

Example

The following example of a custom maintenance script shows the recommended

sequence of steps for a specific Tivoli Storage Manager server configuration:
PARALLEL
BACKUP STGPOOL PRIMDISK DRM_COPYTAPE WAIT=YES
BACKUP STGPOOL PRIMTAPE DRM_COPYTYPE WAIT=YES
SERIAL
PARALLEL
COPY ACTIVEDATA PRIMDISK DRM_ACTTAPE WAIT=YES
COPY ACTIVEDATA PRIMTAPE DRM_ACTTAPE WAIT=YES
SERIAL
BACKUP DB DEVCLASS=TS7650G_N34 TYPE=FULL WAIT=YES
MOVE DRMEDIA * WHERESTATE=MOUNTABLE TOSTATE=VAULT SOURCE=DBBACKUP WAIT=YES
PREPARE SOURCE=DBBACKUP WAIT=YES
MIGRATE STGPOOL PRIMDISK LOWMIG=0 WAIT=YES
EXPIRE INVENTORY SKIPDIRS=NO RESOURCE=4 WAIT=YES
PARALLEL
RECLAIM STGPOOL PRIMDISK THRESHOLD=50 WAIT=YES
RECLAIM STGPOOL PRIMTAPE THRESHOLD=50 WAIT=YES

Restriction: Ensure that the BACKUP STGPOOL command and the BACKUP DB
command complete before you issue other commands, for example, the MOVE
DRMEDIA command.
Related tasks:
“Creating a custom maintenance script” on page 628

Chapter 32. Disaster recovery manager 1033

Recovering from a disaster
This section provides an overview of the tasks involved in recovering the server
and clients. It also presents scenarios of both procedures.

Procedure

Here are guidelines for recovering your server:

1. Obtain the latest disaster recovery plan file.
2. Break out the file to view, update, print, or run as macros or scripts (for
example, batch programs or batch files).
3. Obtain the copy storage pool volumes and active-data pool volumes from the
vault.
4. Locate a suitable replacement machine.
5. Restore the Windows operating system and Tivoli Storage Manager to your
replacement machine. When using the Tivoli Storage Manager device driver
(ADSMSCSI), you will also need to start ADSMSCSI.
6. Review the [Link]
[Link] scripts because they are important for
restoring the server to a point where clients can be recovered (see “Disaster
recovery mode stanza” on page 1060).

Results

Recovering the Clients: To recover clients, do the following:

1. Get the following information by querying the recovered database:
v Client machines that have been defined to Tivoli Storage Manager, along
with their location and restore priority value
v The location of the boot recovery media
v Specific recovery instructions for the machine
v Hardware requirements for the machine
2. With this information restore the client machines. As a first priority, check into
the library the volumes that belong to active-data pools. When a client initiates
a restore attempt, storage volumes within an active-data pool have a higher
restore priority than standard primarysequential storage pool volumes or copy
storage pool volumes.

Server recovery scenario

Here is the procedure for a complete recovery of the server after a disaster has
destroyed it. In this example virtual volumes are not used. The steps are
performed by the onsite administrator unless otherwise indicated.

Procedure

Complete the following steps.

Restriction: When you run the disaster recovery script or the commands that the
script contains, the determination must be made whether to run as root or as the
DB2 instance user ID.
1. Review the recovery steps that are described in the
[Link] stanza of the plan.
2. Request the server backup tapes from the offsite vault.

1034 IBM Tivoli Storage Manager for Windows: Administrator's Guide

3. Break out the recovery plan file stanzas into multiple files (see “Breaking out a
disaster recovery plan file” on page 1048.) These files can be viewed, updated,
printed, or run as Tivoli Storage Manager macros or scripts.
4. Print the [Link] file. Give the printout to the
courier to retrieve the copy storage pool volumes and active-data pool
volumes.
5. Find a replacement server. The [Link] stanza
specifies the device type that is needed to read the backups. The
[Link] stanza specifies the disk space required.

Note: When you use the Tivoli Storage Manager device driver (ADSMSCSI),
you must start ADSMSCSI.
6. The recovery media names and their locations are specified in the
[Link] stanza and the
[Link] stanza. Ensure that the
environment is the same as when the disaster recovery plan file was created.
The environment includes the following directory structures:
v The directory structure of the Tivoli Storage Manager server executable and
disk formatting utility
v The directory structure for Tivoli Storage Manager server configuration files
(disk log, volume history file, device configuration file, and server options
file)
v The directory structure and the files that are created when the disaster
recovery plan file was split into multiple files
7. Restore the operating system, the Tivoli Storage Manager server software, the
Tivoli Storage Manager licenses, and the administrative client on the
replacement hardware.
a. Build a new replacement server instead of restoring the environment from
a backup:
1) Install the Tivoli Storage Manager server software.
2) Create the database directories, the active directories, and the archive
directories as in the original.
3) Run the dsmicfgx utility to configure the replacement instance. This
step configures the API for the DSMSERV RESTORE DB utility.
v Specify the instance userid and password.
v Specify the database directories, the active directories, and the
archive directories.
4) Remove the database instance that was created by the dsmicfgx utility.
For example,
dsmserv removedb TSMDB1
5) Restore the original [Link], volume history, and device
configuration files to the instance directory (as done by the recovery
script in plan file)
6) Run the DSMSERV RESTORE DB (as done by the recovery script in the
plan file).
8. Review the Tivoli Storage Manager macros that are contained in the recovery
plan:
v If, at the time of the disaster, the courier had not picked up the previous
night's database and storage pool incremental backup volumes but they
were not destroyed, remove the entry for the storage pool backup volumes
from the [Link] file.

Chapter 32. Disaster recovery manager 1035

 v If, at the time of the disaster, the courier had not picked up the previous
night's database and active-data pool volumes but they were not destroyed,
remove the entry for the active-data pool volumes from the
[Link] file.
9. If some required storage pool backup volumes could not be retrieved from the
vault, remove the volume entries from the
[Link] file.
If some required active-data pool volumes could not be retrieved from the
vault, remove the volume entries from the
[Link] file.
10. If all primary volumes were destroyed, no changes are required to the
[Link] script and Tivoli Storage Manager macro files.
11. Review the device configuration file to ensure that the hardware configuration
at the recovery site is the same as the original site. Any differences must be
updated in the device configuration file. Examples of configuration changes
that require updates to the configuration information are:
v Different device names
v Use of a manual library instead of an automated library
v For automated libraries, the requirement of manually placing the database
backup volumes in the automated library and updating the configuration
information to identify the element within the library. This allows the server
to locate the required database backup volumes.
For information about updating the device configuration file, see “Updating
the device configuration file” on page 906.
12. To restore the database to a point where clients can be recovered, ensure that
the Tivoli Storage Manager server is halted and then invoke the
[Link] script file. Enter the script
file name at the command prompt. As an alternative, you can use the recovery
script as a guide and manually issue the steps.
The following steps are an example recovery script:
a. Copy the Tivoli Storage Manager server options file from the [Link]
file to its original location.
b. Copy the volume history file that is required by database restore
processing from the [Link] file to its original location.
Use this copy of the volume history file unless you have a more recent
copy (after the disaster occurred).
c. Copy the device configuration file that is required by database restore
processing from the [Link] file to its original
location.
d. Issue the DSMSERV RESTORE DB command.
e. Start the server.
f. Register Tivoli Storage Manager server licenses.
g. Mark copy storage pool volumes and active-data pool volumes retrieved
from the vault as available.
h. Mark copy storage pool volumes and active-data pool volumes that cannot
be obtained as unavailable.
i. Mark primary storage pool volumes as destroyed.
Due to changes in hardware configuration during recovery, you might have
to update the device configuration file in the restored Tivoli Storage
Manager database (see “Updating the device configuration file” on page
906

1036 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 You can mount copy storage pool volumes and active-data pool volumes
upon request, check in the volumes in advance, or manually place the
volumes in the library and ensure consistency by issuing the AUDIT LIBRARY
command. Use the AUDIT LIBRARY command to ensure that the restored
Tivoli Storage Manager database is consistent with the automated library
volumes.
13. If client workstations are not damaged, invoke the
[Link] script file to restore the server primary
storage pools. If client workstations are damaged, you may want to delay this
action until after all clients are recovered.
This action is optional because Tivoli Storage Manager can access the copy
storage pool volumes and active-data pool volumes directly to restore client
data. Using this feature, you can minimize client recovery time because server
primary storage pools do not have to be restored first. However, in this
scenario, the client workstations were not damaged, so the focus of the
administrator is to restore full Tivoli Storage Manager server operation.
As an alternative, you can use the recovery script as a guide and manually
run each step. The steps run in this script are:
a. Create replacement primary volumes.
b. Define the replacement primary volumes to Tivoli Storage Manager.
c. Restore the primary storage pools from copy storage pools.
Attention: Restoring a primary storage pool from an active-data pool
might cause some or all inactive files to be deleted from the database if the
server determines that an inactive file must be replaced but cannot find it
in the active-data pool.
14. Collect the database backup, copy storage pool volumes, and active-data pool
volumes used in the recovery for return to the vault. For these backup
volumes to be returned to the vault using the routine MOVE DRMEDIA
process, issue the following commands:
update volhist TPBK50 devcl=lib8mm ormstate=mountable
update volhist TPBK51 devcl=lib8mm ormstate=mountable

The copy storage pool volumes and active-data pool volumes used in the
recovery already have the correct ORMSTATE.
15. Issue the BACKUP DB command to back up the newly restored database.
16. Issue the following command to check the volumes out of the library:
move drmedia * wherestate=mountable
17. Create a list of the volumes to be given to the courier:
query drmedia * wherestate=notmountable
18. Give the volumes to the courier and issue the following command:
move drmedia * wherestate=notmountable
19. Issue the PREPARE command.

Chapter 32. Disaster recovery manager 1037

 Client recovery scenario
The following scenario demonstrates the recovery of clients.

Procedure
1. To view a list of client machines that were lost in building 21 and their restore
priority, issue the following command:
query machine building=021 format=detailed
DRM displays information similar to the following:
Machine Name: POLARIS
Machine Priority: 1
Building: 21
Floor: 2
Room: 1
Server?: No
Description: Payroll
Node Name: POLARIS
Recovery Media Name: MKSYSB1
Characteristics?: Yes
Recovery Instructions?: Yes

Identify which client machines have the highest priority so that restores can
begin using active-data pool volumes.
2. For each machine, issue the following commands:
a. Determine the location of the boot media. For example:
query recoverymedia mksysb1
The server displays the following information:
Recovery Media Name Volume Names Location Machine Name
-------------------- ----------- ---------- ----------------
MKSYSB1 vol1 vol2 IRONVAULT POLARIS
vol3

b. Determine the machine-specific recovery instructions. For example:

query machine polaris format=recoveryinstructions
The server displays the following:
Recovery Instructions for Polaris.
Primary Contact:
Jane Smith (wk 520-000-0000 hm 520-001-0001)
Secondary Contact:
John Adams (wk 520-000-0001 hm 520-002-0002)

c. Determine the machine hardware requirements.

query machine polaris format=characteristics
The server displays information similar to the following:

1038 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 devices
aio0 Defined Asynchronous I/O
bus0 Available 00-00 Microchannel Bus
fd0 Available 00-00-0D-00 Diskette Drive
fda0 Available 00-00-0D Standard I/O Diskette Adapter
fpa0 Available 00-00 Floating Point Processor
gda0 Available 00-04 Color Graphics Display Adapter
hd1 Defined Logical volume
hd2 Defined Logical volume
hd3 Defined Logical volume
hdisk0 Available 00-01-00-00 400 MB SCSI Disk Drive
hdisk1 Available 00-01-00-40 Other SCSI Disk Drive
hft0 Available High Function Terminal Subsystem
inet0 Available Internet Network Extension
ioplanar0 Available 00-00 I/O Planar
kbd0 Defined 00-00-0K-00 United States keyboard
lb0 Available 00-02-00-20 TIVSM Library
lo0 Available Loopback Network Interface
loglv00 Defined Logical volume
lp0 Available 00-00-0P-00 IBM 4201 Model 3 Proprinter III
lv03 Defined Logical volume
lv04 Defined Logical volume
lvdd Available N/A
mem0 Available 00-0B 8 MB Memory Card
mem1 Available 00-0C 16 MB Memory Card
mous0 Defined 00-00-0M-00 3 button mouse
mt0 Available 00-02-00-40 TIVSM Tape Drive
ppa0 Available 00-00-0P Standard I/O Parallel Port Adapter
pty0 Available Asynchronous Pseudo-Terminal
rootvg Defined Volume group
sa0 Available 00-00-S1 Standard I/O Serial Port 1
sa1 Available 00-00-S2 Standard I/O Serial Port 2
scsi0 Available 00-01 SCSI I/O Controller
scsi1 Available 00-02 SCSI I/O Controller
sio0 Available 00-00 Standard I/O Planar
siokb0 Available 00-00-0K Keyboard Adapter
sioms0 Available 00-00-0M Mouse Adapter
siotb0 Available 00-00-0T Tablet Adapter
sys0 Available 00-00 System Object
sysplanar0 Available 00-00 CPU Planar
sysunit0 Available 00-00 System Unit
tok0 Available 00-03 Token-Ring High-Performance Adapter
tr0 Available Token Ring Network Interface
tty0 Available 00-00-S1-00 Asynchronous Terminal
tty1 Available 00-00-S2-00 Asynchronous Terminal
usrvice Defined Logical volume
veggie2 Defined Volume group
logical volumes by volume group
veggie2:
LV NAME TYPE LPs PPs PVs LV STATE MOUNT POINT
hd2 jfs 103 103 1 open/syncd /usr
hd1 jfs 1 1 1 open/syncd /home
hd3 jfs 3 3 1 open/syncd /tmp
hd9var jfs 1 1 1 open/syncd /var
file systems
Filesystem Total KB free %used iused %iused Mounted on
/dev/hd4 8192 420 94% 909 44% /
/dev/hd9var 4096 2972 27% 87 8% /var
/dev/hd2 421888 10964 97% 17435 16% /usr
/dev/hd3 12288 11588 5% 49 1% /tmp
/dev/hd1 4096 3896 4% 26 2% /home

3. With the information obtained, restore each client machine.

Once the high-priority clients have begun restoring their data from active-data
pools, the other, lower-priority clients can begin restoring directly from copy
storage pools. Restoration from copy storage pools can run concurrently with
the restoration from active-data pools. High-priority clients do not attempt to
access the copy storage pool volumes because active-data pools have a higher
restore priority than copy storage pools.

Chapter 32. Disaster recovery manager 1039

Recovering with different hardware at the recovery site
You may have to recover your system using hardware that is different from that
used when you backed up your database and created disaster recovery plan file.
Before restoring the database, update the device configuration file included in the
recovery plan file. After restoring the database, update the device configuration on
the database.

About this task

This section describes a number of such situations in detail. If the hardware

environment is different at the recovery site, you must update the device
configuration file. Tivoli Storage Manager uses the device configuration file to
access the devices that are needed to read the database backup volumes. The
[Link] stanza in the plan file identifies the volumes
that are needed to restore the database.

Automated SCSI library at the original and recovery sites

Manually place the database backup volumes in the automated library and note
the element numbers where you place them. Then update the comments in the
device configuration file to identify the locations of those volumes.

About this task

Note: You may also need to audit the library after the database is restored in order
to update the server inventory of the library volumes.

Here is an example of an original device configuration file, which describes an

automated tape library:
/* Device Configuration */

define devclass auto8mm_class devtype=8mm format=drive

mountlimit=2 mountwait=60 mountretention=60
prefix=tsm library=auto8mmlib

define library auto8mmlib libtype=scsi

define drive auto8mmlib 8mm_tape0 element=82 online=yes

define drive auto8mmlib 8mm_tape1 element=83 online=yes

define path server1 auto8mmlib srctype=server desttype=library

device=lb4.0.0.1

define path server1 8mm_tape0 srctype=server desttype=drive

library=auto8mmlib device=mt1.0.0.1

define path server1 8mm_tape1 srctype=server desttype=drive

library=auto8mmlib device=mt2.0.0.1

/* LIBRARYINVENTORY SCSI AUTO8MMLIB KEV004 1 101*/

/* LIBRARYINVENTORY SCSI AUTO8MMLIB KEV005 3 101*/

Here is an example of the updated device configuration file when an automated

library is used at the recovery site to read a database volume DBBK01:
/* Device Configuration */

define devclass auto8mm_class devtype=8mm format=drive

1040 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 mountlimit=2 mountwait=60 mountretention=60
prefix=tsm library=auto8mmlib

define library auto8mmlib libtype=scsi

define drive auto8mmlib 8mm_tape0 element=82 online=yes

define drive auto8mmlib 8mm_tape1 element=83 online=yes

define path server1 auto8mmlib srctype=server desttype=library

device=lb4.0.0.1

define path server1 8mm_tape0 srctype=server desttype=drive

library=auto8mmlib device=mt1.0.0.1

define path server1 8mm_tape1 srctype=server desttype=drive

library=auto8mmlib device=mt2.0.0.1

/* LIBRARYINVENTORY SCSI AUTO8MMLIB DBBK01 1 101*/

In this example, database backup volume DBBK01 was placed in element 1 of the
automated library. Then a comment is added to the device configuration file to
identify the location of the volume. Tivoli Storage Manager needs this information
to restore the database restore. Comments that no longer apply at the recovery site
are removed.

Automated SCSI library at the original site and a manual scsi

library at the recovery site
Ensure that the DEFINE DRIVE and DEFINE LIBRARY commands in the device
configuration file are valid for the new hardware configuration.

About this task

For example, if an automated tape library was used originally and cannot be used
at the recovery site, update the device configuration file. Include the DEFINE
LIBRARY and DEFINE DRIVE commands that are needed to define the manual
drive to be used. In this case, you must manually mount the backup volumes.

Note: If you are using an automated library, you may also need to update the
device configuration file to specify the location of the database backup volume.

Here is an example of an original device configuration file, which describes an

automated tape library:
/* Device Configuration */

define devclass auto8mm_class devtype=8mm format=drive

mountlimit=2 mountwait=60 mountretention=60
prefix=tsm library=auto8mmlib

define library auto8mmlib libtype=scsi

define drive auto8mmlib 8mm_tape0 element=82 online=yes

define drive auto8mmlib 8mm_tape1 element=83 online=yes

define path server1 auto8mmlib srctype=server desttype=library

device=lb4.0.0.1

define path server1 8mm_tape0 srctype=server desttype=drive

library=auto8mmlib device=mt1.0.0.1

Chapter 32. Disaster recovery manager 1041

 define path server1 8mm_tape1 srctype=server desttype=drive
library=auto8mmlib device=mt2.0.0.1

/* LIBRARYINVENTORY SCSI AUTO8MMLIB KEV004 1 101*/

/* LIBRARYINVENTORY SCSI AUTO8MMLIB KEV005 3 101*/

Here is an example of the updated device configuration file when a manual library
is used at the recovery site:
/* Device Configuration */

define devclass auto8mm_class devtype=8mm format=drive

mountlimit=1 mountwait=60 mountretention=60 prefix=tsm
library=manual8mm

define library manual8mm libtype=manual

define drive manual8mm 8mm_tape0

define path server1 8mm_tape0 srctype=server desttype=drive

library=manual8mm device=mt1.0.0.1

The following changes were made:

v In the device class definition, the library name was changed from
AUTO8MMLIB to MANUAL8MM. The device class name remains the same
because it is associated with the database backup volumes in the volume history
file.
v The manual library, MANUAL8MM, was defined.
v A new drive, 8MM_TAPE0, was defined for the manual library.
v The comments that named the location of volumes in the automated library
were removed.

After you restore the database, modify the device configuration file in the
database. After starting the server, define, update, and delete your library and
drive definitions to match your new configuration.

Note: If you are using an automated library, you may need to use the AUDIT
LIBRARY command to update the server inventory of the library volumes.

Managing copy storage pool volumes and active-data pool

volumes at the recovery site
The [Link] stanza in the recovery plan file identifies
the required copy storage pool volumes and active-data pool volumes. All volumes
must be available to the restored server.

About this task

The restored server uses copy storage pool volumes to satisfy requests (for
example, from backup/archive clients) and to restore primary storage pool
volumes that were destroyed. If they are available, the server uses active-data
pools to restore critical client data.

Attention: Restoring a primary storage pool from an active-data pool might

cause some or all inactive files to be deleted from the database if the server
determines that an inactive file needs to be replaced but cannot find it in the
active-data pool.

1042 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 After the database is restored, you can handle copy storage pool volumes and
active-data pool volumes at the recovery site in three ways:
v Mount each volume as requested by Tivoli Storage Manager. If an automated
library is used at the recovery site, check the volumes into the library.
v Check the volumes into an automated library before Tivoli Storage Manager
requests them.
v Manually place the volumes in an automated library and audit the library to
update the server inventory.

If you are using an automated library, you may also need to audit the library after
the database is restored in order to update the Tivoli Storage Manager inventory of
the volumes in the library.

Disaster recovery manager checklist

The following checklist can help you set up disaster recovery manager.
Table 81. Checklist
Activity Start Date End Date Status Responsible Backup
Person Person
Plan for DRM
Evaluate your disaster recovery
requirements
v What are the business priorities for
recovering your clients?
v Where is the recovery site?
v Is the recovery site hot, warm, or cold?
v Do the clients have connectivity to
recovery server?
v Who are the system and Tivoli Storage
Manager administrators?
v Will you need to return to the original
site?
v Where are the offsite backups stored?
v How does the vault handle the backup
media?
v How are the backups packaged or
processed?
v Who provides the courier service?
Evaluate the current storage pool backup
implementation
v What primary storage pools are being
backed up?
v When are the backups performed?
v Will the backups remain onsite or be sent
offsite?
v Naming conventions for replacement
volumes for primary storage pools

Chapter 32. Disaster recovery manager 1043

Table 81. Checklist (continued)
Activity Start Date End Date Status Responsible Backup
Person Person
Evaluate the current database backup
implementation
v When are the backups performed?
v Backup purpose: offsite or onsite
v Will you use snapshot database backups
or full plus incremental database
backups?
v How long do you want to keep backup
series? Verify that the values for copy
storage pool and active-data pool
REUSEDELAY and
DRMDBBACKUPEXPIREDAYS are the
same. If copy storage pools or active-data
pools managed by DRM have different
REUSEDELAY values, set the
DRMDBBACKUPEXPIREDAYS value to
the highest REUSEDELAY value.
Determine which primary storage pools
are to be managed by DRM
Determine which copy storage pools are to
be managed by DRM
v Offsite copy storage pools
Determine which active-data pools are to
be managed by DRM
v Offsite active-data pools
Where to Save the Recovery Plan File

Locally:
v What is the recovery plan file pathname
prefix?
v How will recovery plan files be made
available at the recovery site?
– Print and store offsite
– Copy stored offsite
– Copy sent/NFS to recovery site

On Another Server:
v What server is to be used as the target
server?
v What is the name of the target server's
device class?
v How long do you want to keep recovery
plan files?
Determine where you want to create the
user-specified recovery instructions

What is the prefix of the instructions

pathname?

1044 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Table 81. Checklist (continued)
Activity Start Date End Date Status Responsible Backup
Person Person
Analyze the sequence of steps related to
the PREPARE command backup
movement

Document the flow of activities and timings

v Sending of volumes offsite
v Return of empty volumes
v PREPARE timing
Installation
Receive and Install the Tivoli Storage
Manager code
License DRM
v REGISTER LICENSE or
v Update the server options

Chapter 32. Disaster recovery manager 1045

Table 81. Checklist (continued)
Activity Start Date End Date Status Responsible Backup
Person Person
Set DRM defaults

Issue:
v SET DRMDBBACKUPEXPIREDAYS to
define the database backup expiration
v SET DRMPRIMSTGPOOL to specify the
DRM-managed primary storage pools
v SET DRMCOPYSTGPOOL to specify the
DRM-managed copy storage pools
v SET DRMACTIVEDATASTGPOOL to
specify the DRM-managed active-data
pools
v SET DRMPLANVPOSTFIX to specify a
character to be appended to new storage
pools
v SET DRMPLANPREFIX to specify the
RPF prefix
v SET DRMINSTRPREFIX to specify the
user instruction file prefix
v SET DRMNOTMOUNTABLENAME to
specify the default location for media to
be sent offsite
v SET DRMCOURIERNAME to specify the
default courier
v SET DRMVAULTNAME to specify the
default vault
v SET DRMCMDFILENAME to specify the
default file name to contain the
commands specified with the CMD
parameter on MOVE and QUERY
DRMEDIA
v SET DRMCHECKLABEL to specify
whether volume labels are verified when
checked out by the MOVE DRMEDIA
command
v SET DRMRPFEXPIREDAYS to specify a
value for the frequency of RPF expiration
(when plan files are stored on another
server)

1046 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Table 81. Checklist (continued)
Activity Start Date End Date Status Responsible Backup
Person Person
Define the site-specific recovery
instructions

Identify:
v Target disaster recovery server location
v Target server software requirements
v Target server hardware requirements
(storage devices)
v Tivoli Storage Manager administrator
contact
v Courier name and telephone number
v Vault location and contact person

Create:
v Enter the site-specific recovery
instructions data into files created in the
same path/HLQ as specified by SET
DRMINSTRPREFIX
Test disaster recovery manager
Test the installation and customization
v QUERY DRMSTATUS to display the
DRM setup
v Back up the active and inactive data that
is in primary storage pools to copy
storage pools. Copy the active data that
is in primary storage pools to active-data
pools.
v Back up the Tivoli Storage Manager
database
v QUERY DRMEDIA to list the copy
storage pool and active-data pool
volumes
v MOVE DRMEDIA to move offsite
v PREPARE to create the recovery plan file
Examine the recovery plan file created
Test the recovery plan file break out
v VBScript procedure [Link]
v Locally written procedure
Set up the schedules for automated
functions

Chapter 32. Disaster recovery manager 1047

The disaster recovery plan file
The disaster recovery plan file contains the information required to recover a Tivoli
Storage Manager server to the point in time represented by the last database
backup operation that is completed before the plan is created. The plan is
organized into stanzas, which you can break out into multiple files.

Tip: The plan file that DRM generates is a template that contains information,
including commands for recovering the database, that might not apply to your
replacement systems or to your particular recovery scenario. To modify the plan or
to store additional instructions that you will need during recovery from an actual
disaster, use the [Link] stanzas. Enter your site-specific
information in the stanzas when you first create the plan file or after you test it.

Breaking out a disaster recovery plan file

You can break out the stanzas of the disaster recovery plan file into individual
files.

You can use a Microsoft VBScript command procedure or an editor to break out
the stanzas in the disaster recovery plan file into individual files. A sample
procedure, [Link], is shipped with DRM. You can modify the procedure for
your installation. Store a copy of the procedure offsite for recovery.

1048 IBM Tivoli Storage Manager for Windows: Administrator's Guide

’****************************************************************************
’ IBM TSM Disaster Recovery Manager for Windows Server 2008 Sample Script
’
’ Explode a recovery plan file into separate files (batch programs,
’ TSM macros, TSM server options file etc.)
’
’ Invoke with:
’ cscript [Link] recoveryplanfilename
’ where:
’ recoveryplanfilename is the name of the recovery plan file created
’ by the DRM PREPARE command
’
’ Example usage:
’ cscript [Link] c:\adsmsrv\recplans\20001115.051421
’*****************************************************************************

Dim args
Dim PLANFILE, OUTDIR, OUTFILE
Dim STANZAS
Dim VOLNAMES(100),NbrV,LOGDBVOLS
Dim fso, fi, fo
Dim WORDS
Dim CRLF
Dim RESULTS, RESULTS2
CRLF = Chr(13) & Chr(10)
LOGDBVOLS = False : NbrV = 0
OUTDIR = "" : OUTFILE = ""
RESULTS = "" : RESULTS2 = ""

’*****************************************************************************
’* Get input arguments: PLANFILE=recoveryplanfilename
’*****************************************************************************

set args = [Link]

If [Link] < 1 Then

[Link] _
"usage: cscript [Link] recoveryplanfilename" & CRLF & _
"example: cscript [Link] c:\adsmsrv\recplans\20001115.051421"
[Link](1)
Else
PLANFILE = [Link](0)
End If

RESULTS = RESULTS & "Planfile: " & PLANFILE & CRLF

’****************************************************************************
’ For each recovery plan file stanza name determine the extension (if any)
’ to be added to the file name created by using the stanza name and extension
’****************************************************************************

Chapter 32. Disaster recovery manager 1049

 Set STANZAS = CreateObject("[Link]")
[Link] "[Link]", ".CMD"
[Link] "[Link]" , ".CMD"
[Link] "[Link]" , "none"
[Link] "[Link]" , "none"
[Link] "[Link]" , "none"
[Link] "[Link]" , ".MAC"
[Link] "[Link]" , ".MAC"
[Link] "[Link]" , ".MAC"
[Link] "[Link]" , ".MAC"
[Link] "[Link]" , ".MAC"
[Link] "[Link]" , ".MAC"
[Link] "[Link]" , ".MAC"
[Link] "[Link]" , ".MAC"
[Link] "[Link]" , "none"
[Link] "[Link]" , "none"
[Link] "[Link]" , "none"
[Link] "[Link]" , "none"
[Link] "[Link]" , "none"
[Link] "[Link]" , "none"
[Link] "[Link]" , "none"
[Link] "[Link]" , "none"
[Link] "[Link]" , "none"
[Link] "[Link]" , "none"
[Link] "[Link]" , "none"
[Link] "[Link]" , "none"
[Link] "[Link]" , "none"

Set fso = CreateObject("[Link]")

Set fi = [Link](PLANFILE, 1, False)

Do While [Link] <> True

’****************************************************************************
’ Read a line from the input recovery plan file
’****************************************************************************

ALINE = [Link]

’****************************************************************************
’ Get the first 2 words. We’re looking for ’begin’/’end’ and a stanza name
’****************************************************************************

WORD1 = "" : WORD2 = "" : THEREST = ""

If Not ALINE = "" then
WORDS = Split(ALINE, " ", -1, 1)
WORD1 = WORDS(0)
If Ubound(WORDS) > 0 Then WORD2 = WORDS(1)
if Ubound(WORDS) > 1 Then THEREST = WORDS(2)
End If

1050 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 ’****************************************************************************
’ If the first word is ’begin’ and this is a stanza that we’ll create a file
’ for then build the output file name using the output directory. Add an
’ extension if needed. Erase the previous version of the file and then
’ indicate that the new file is being created.
’****************************************************************************

If WORD1 = "begin" And [Link](WORD2) Then

OUTFILE = OUTDIR & WORD2

If Not [Link](WORD2) = "none" Then

OUTFILE = OUTFILE & [Link](WORD2)
End If

Set fo = [Link](OUTFILE, 2, True)

RESULTS = RESULTS & "Creating file " & OUTFILE & CRLF
’****************************************************************************
’ If the first word is ’end’ and this was a stanza that we created a file
’ for then close the output file.
’****************************************************************************

Elseif WORD1 = "end" And [Link](WORD2) Then

[Link]
OUTFILE = ""
End If

’****************************************************************************
’ This is the line within the plan file that identifies the plan file prefix.
’****************************************************************************

Elseif OUTDIR = "" And WORD1 = "DRM" And WORD2 = "PLANPREFIX" Then
OUTDIR = THEREST
If Not Right(OUTDIR,1) = "\" Then
OUTDIR = OUTDIR & "."
End If
RESULTS = RESULTS & "set planprefix to " & OUTDIR & CRLF

End If ’/* select on first word of input line from the recovery plan file */

Loop ’/* do while more lines in input recovery plan file */

[Link]

Structure of the disaster recovery plan file

The disaster recovery plan is divided into the following types of stanzas:

Tip: The plan file that DRM generates is a template that contains information,
including commands for recovering the database, that might not apply to your
replacement systems or to your particular recovery scenario. To modify the plan or
to store additional instructions that you will need during recovery from an actual
disaster, use the [Link] stanzas. Enter your site-specific
information in the stanzas when you first create the plan file or after you test it.
Command stanzas
Consist of scripts (for example, batch programs or batch files) and Tivoli
Storage Manager macros. You can view, print, and update these stanzas,
and run them during recovery.

Note: The [Link] and

[Link] stanzas contain the
commands that invoke the scripts and macros contained in the other
stanzas.

Chapter 32. Disaster recovery manager 1051

 Instruction stanzas
Consist of recovery instructions specific to your site. You can view, print,
and update these stanzas, and use them during recovery.
Server requirements stanzas
Include the database and recovery log requirements, device and volume
requirements, and license information. You can view and print these
stanzas, and use them during recovery.
Configuration file stanzas
Consist of the volume history, device configuration, and server options
files.
Machine and recovery media stanzas
Consist of machine recovery instructions and information about machine
hardware, software, and recovery media. You can print and update these
stanzas, and use them during server recovery.

Table 82 lists the recovery plan file stanzas, and indicates what type of
administrative action is required during set up or periodic updates, routine
processing, and disaster recovery. The table also indicates whether the stanza
contains a macro, a script, or a configuration file.

Note: For tasks identified as During setup or periodic updates, DRM

automatically collects this information for the plan.
Table 82. Administrative tasks associated with the disaster recovery plan file
Stanza Name Tasks
[Link] None
[Link] None
[Link] None
[Link] During setup or periodic updates: Edit the
source file associated with the stanza
(optional)
[Link] During setup or periodic updates: Edit the
source file associated with the stanza
(optional)
[Link] During setup or periodic updates: Edit the
source file associated with the stanza
(optional)
[Link] During setup or periodic updates: Edit the
source file associated with the stanza
(optional)
[Link] During setup or periodic updates: Edit the
source file associated with the stanza
(optional)
[Link] During routine processing: MOVE
DRMEDIA
[Link] None
[Link]. [Link] script During disaster recovery: Edit and run
(optional)
[Link]. [Link] script During disaster recovery: Edit and run
(optional)
[Link] During disaster recovery: Edit (optional)

1052 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Table 82. Administrative tasks associated with the disaster recovery plan file (continued)
Stanza Name Tasks
[Link] macro During disaster recovery: Edit and run
(optional)
[Link] During routine processing: MOVE
DRMEDIA

During disaster recovery: Edit and run

(optional)
[Link] During routine processing: MOVE
DRMEDIA

During disaster recovery: Edit and run

(optional)
[Link] macro During routine processing: MOVE
DRMEDIA

During disaster recovery: Edit and run

(optional)
[Link] macro During routine processing: MOVE
DRMEDIA

During disaster recovery: Edit and run

(optional)
[Link] macro During disaster recovery: Edit and run
(optional)
[Link] macro During disaster recovery: Edit and run
(optional)
[Link] macro During disaster recovery: Edit and run
(optional)
[Link] configuration file During disaster recovery: Copy (optional)
[Link] configuration file During disaster recovery: Edit and copy
(optional)
[Link] configuration file During disaster recovery: Edit and copy
(optional)
[Link] None
[Link] During setup or periodic updates: Issue
DEFINE MACHINE ADSMSERVER=YES
(optional)
[Link] During setup or periodic updates: Issue
INSERT MACHINE
RECOVERYINSTRUCTIONS (optional)
[Link] During setup or periodic updates: Issue
INSERT MACHINE CHARACTERISTICS
(optional)
[Link] During setup or periodic updates: Issue
DEFINE RECOVERYMEDIA and DEFINE
RECMEDMACHASSOCIATION (optional)

Chapter 32. Disaster recovery manager 1053

 Example disaster recovery plan file
This section contains an example of a disaster recovery plan file and information
about each stanza. The disaster recovery plan file has been divided into separate
figures that correlate to the descriptions of specific stanzas within each figure.

Tip: The plan file that DRM generates is a template that contains information,
including commands for recovering the database, that might not apply to your
replacement systems or to your particular recovery scenario. To modify the plan or
to store additional instructions that you will need during recovery from an actual
disaster, use the [Link] stanzas. Enter your site-specific
information in the stanzas when you first create the plan file or after you test it.

Description and table of contents stanzas

These stanzas identify the server for the recovery plan and the date and time the
plan is created, and also list all the stanzas in the plan.

[Link]
begin [Link]

Recovery Plan for Server COUPE

Created by DRM PREPARE on 09/26/2008 [Link]
DRM PLANPREFIX D:\TSM\SERVER1\PLANPRE
Storage Management Server for Windows - Version 6, Release 1, Level 0.0

end [Link]

[Link]
begin [Link]

[Link]
[Link]

Server Recovery Stanzas:

[Link]
[Link]
[Link]
[Link] script
[Link] script
[Link]
[Link] macro
[Link] macro
[Link] macro
[Link] macro
[Link] macro
[Link] macro
[Link] macro
[Link] macro
[Link]
[Link]
[Link]
[Link]

end [Link]

1054 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Server requirements stanza
The [Link] stanza identifies the database and recovery log
storage requirements for the server.

The replacement server must have enough disk space to install the database and
recovery log.

This stanza also identifies the Tivoli Storage Manager installation directory. When
Tivoli Storage Manager is re-installed on the replacement server, specify this
directory on the Setup Type panel during installation. If you specify a different
directory, edit the plan file to account for this change.

Chapter 32. Disaster recovery manager 1055

 begin [Link]

Database Requirements Summary:

Database Name: tsm_serv

Total Space(MB): 285,985
Used Space(MB): 384
Free Space(MB): 285,527
Page Size(Bytes): 16,384
Total Pages: 26,627
Usable Pages: 26,491
Used Pages: 22,003
Free Pages: 4,488
Full Device Class Name: VTL

Location: E:\tsmdata\DBSpace
Total Space(MB): 285,985
Used Space(MB): 457
Free Space(MB): 285,527

Recovery Log Requirements Summary:

Assigned Capacity (MB): 8,192

Used Space(MB): 0
Free Space(MB): 8,159
Active Log Directory: H:\tsmdata\Alog
Mirror Log Directory:
Archive Failover Log Directory: c:\tsmdata\Flog
Archive Log Directory: H:\tsmdata\archive

Server Installation Directory: D:\tsm\

end [Link]

*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*

begin [Link]

Volumes required for data base restore

Location = dkvault
Device Class = VTL
Volume Name =
003902L4

Volumes required for storage pool restore

Location = dkvault
Copy Storage Pool = COPYPOOL
Device Class = VTL
Volume Name =
003900L4

Volumes required for active-data storage pool(s)

Location = dkvault
Active-data Storage Pool = ADP1
Device Class = VTL
Volume Name =
003901L4

end [Link]

1056 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Recovery instructions stanzas
The administrator enters recovery instructions into source files that the PREPARE
command includes in the plan files.

See “Specifying recovery instructions for your site” on page 1011 for details. In the
following descriptions, prefix represents the prefix portion of the file name. See
“Specifying defaults for the disaster recovery plan file” on page 1006 for details.

[Link]

Identifies site-specific instructions that the administrator has entered in the file
identified by prefix [Link]. The instructions
should include the recovery strategy, key contact names, an overview of key
applications backed up by this server, and other relevant recovery instructions.
begin [Link]

This server contains the backup and archive data for FileRight Company
accounts receivable system. It also is used by various end users in the
finance and materials distribution organizations.
The storage administrator in charge of this server is Jane Doe 004-001-0006.
If a disaster is declared, here is the outline of steps that must be completed.
1. Determine the recovery site. Our alternate recovery site vendor is IBM
BRS in Tampa, Fl, USA 213-000-0007.
2. Get the list of required recovery volumes from this recovery plan file
and contact our offsite vault so that they can start pulling the
volumes for transfer to the recovery site.
3. etc...

end [Link]

[Link]

Contains instructions that the administrator has entered in the file identified by
prefix [Link]. The instructions should include the
name and location of the offsite vault, and how to contact the vault (for example, a
name and phone number).
begin [Link]

Our offsite vaulting vendor is OffsiteVault Inc.

Their telephone number is 514-555-2341. Our account rep is Joe Smith.
Our account number is 1239992. Their address is ...
Here is a map to their warehouse ...
Our courier is ...

end [Link]

[Link]

Contains instructions that the administrator has entered in the file identified by
prefix [Link]. The instructions should include how
to rebuild the base server machine and the location of the system image backup
copies.

Chapter 32. Disaster recovery manager 1057

 begin [Link]

The base server system is Windows Server 2008: Standard running on

an IBM PC-350. The Windows Server 2008 operating system and product
installation media is stored at the vault. There is also a copy
in bldg 24 room 4 cabinet a. The system administrator responsible
for the Windows Server 2008 and server installation is Fred Myers.
Following are the instructions for installation of Windows Server
2008 and the server:

end [Link]

[Link]

Contains instructions that the administrator has entered in the file identified by
prefix [Link]. The instructions should include
how to prepare for the database recovery. For example, you may enter instructions
on how to initialize or load the backup volumes for an automated library. No
sample of this stanza is provided.

[Link]

Contains instructions that the administrator has entered in the file identified by
prefix [Link]. The instructions should include the
names of your software applications and the copy storage pool names containing
the backup of these applications. No sample of this stanza is provided.

Volume and device requirements stanzas

These stanzas provide a list of volumes required to recover the server and details
about the devices needed to read those volumes.

[Link]

Provides a list of the database backup, copy storage-pool volumes, and active-data
pool volumes required to recover the server. This list can include both virtual
volumes and nonvirtual volumes. A database backup volume is included if it is
part of the most recent database backup series. A copy storage pool volume or an
active-data pool volume is included if it is not empty and not marked destroyed.

If you are using a nonvirtual volume environment and issuing the MOVE
DRMEDIA command, a blank location field means that the volumes are onsite and
available to the server. This volume list can be used in periodic audits of the
volume inventory of the courier and vault. You can use the list to collect the
required volumes before recovering the server.

For virtual volumes, the location field contains the target server name.

1058 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 begin [Link]

Volumes required for data base restore

Location = dkvault
Device Class = VTL
Volume Name =
003902L4

Volumes required for storage pool restore

Location = dkvault
Copy Storage Pool = COPYPOOL
Device Class = VTL
Volume Name =
003900L4

Volumes required for active-data storage pool(s)

Location = dkvault
Active-data Storage Pool = ADP1
Device Class = VTL
Volume Name =
003901L4

end [Link]

[Link]

Provides details about the devices needed to read the backup volumes.
begin [Link]

Purpose: Description of the devices required to read the

volumes listed in the recovery volumes required stanza.

Device Class Name: VTL

Device Access Strategy: Sequential
Storage Pool Count: 2
Device Type: LTO
Format: DRIVE
Est/Max Capacity (MB):
Mount Limit: 2
Mount Wait (min): 5
Mount Retention (min): 1
Label Prefix: ADSM
Drive Letter:
Library: VTL
Directory:
Server Name:
Retry Period:
Retry Interval:
Twosided:
Shared:
WORM: No
Drive Encryption: Allow
Scaled Capacity:
Last Update by (administrator): ADMIN
Last Update Date/Time: 09/26/2008 [Link]

end [Link]

Chapter 32. Disaster recovery manager 1059

 Disaster recovery mode stanza
The [Link] stanza contains a script
with the commands needed to recover the server.

[Link]

You can use the script as a guide and run the commands from a command line. Or
you can copy it to a file, modify it and the files it refers to, and run the script.

Tip: The commands in the plan file that is generated by DRM might not work on
your replacement systems. If necessary, use the [Link] stanzas in the
plan file to store information about the particular commands to be used during
recovery from an actual disaster. Enter your site-specific information in the
[Link] stanzas when you first create the plan file or after you test it.
At the completion of these steps, client requests for file restores are satisfied
directly from copy storage pool volumes and active-data pool volumes.

The disaster recovery plan issues commands by using the administrative client.

The commands in the script do the following:

v Restore the server options, volume history, and device configuration information
files.
v Invoke the macros contained in the following stanzas:
– [Link]
– [Link]
– [Link]
– [Link]
– [Link]
– [Link].

Note: Because this script invokes the administrative command-line client, ensure
that the communications options in the administrative client options file are set to
communicate with the recovered server before running this script. To review the
communications options used in the recovered server, see the server options file in
the [Link] stanza.

To invoke this script, specify the following positional parameters:

v %1 (the administrator ID)
v %2 (the administrator password)

For example, to invoke this script by using an administrator ID of don and a

password of mox, enter the following command:
[Link] don mox

For more information, see the entry for the recovery plan prefix in Table 79 on
page 1006.

1060 IBM Tivoli Storage Manager for Windows: Administrator's Guide

begin [Link] script

@echo off

rem Purpose: This script contains the steps required to recover the server
rem to the point where client restore requests can be satisfied
rem directly from available copy storage pool volumes.
rem Note: This script assumes that all volumes necessary for the restore have
rem been retrieved from the vault and are available. This script assumes
rem the recovery environment is compatible (essentially the same) as the
rem original. Any deviations require modification to this script and the
rem macros and scripts it runs. Alternatively, you can use this script
rem as a guide, and manually execute each step.

if not %1.==. if not %2.==. goto start

echo Specify the following positional parameters:
echo administrative client ID and password.
echo Script stopped.
goto end
:start

rem Set the server working directory.

pushd "D:\tsm\server1\"

rem Restore server options, volume history, device configuration files.

copy "D:\TSM\SERVER1\[Link]" "D:\TSM\SERVER1\[Link]"
copy "D:\TSM\SERVER1\[Link]" "D:\TSM\SERVER1\[Link]"
copy "D:\TSM\SERVER1\[Link]" "D:\TSM\SERVER1\[Link]"

rem Make sure db storage paths exist.

mkdir "E:\tsmdata\DBSpace"

rem Restore the server database to latest version backed up per the
rem volume history file.
"D:\TSM\SERVER\DSMSERV" -k "Server1" restore db todate=09/26/2008 totime=[Link] +
source=dbb

rem "D:\TSM\SERVER\DSMSERV" -k "Server1" restore db todate=09/26/2008 totime=[Link] +

source=dbb on="D:\TSM\SERVER1\[Link]" activelogdirectory="H:\tsmdata\Alog"

rem Start the server.

start "Server1" "D:\TSM\SERVER\DSMSERV" -k "Server1"
echo Wait for the server to start. Ensure that the Administrative command
echo line client option file is set up to communicate with this server, then
echo press enter to continue recovery script execution.
pause

rem Set the administrative command line client directory.

pushd "D:\tsm\Server\tsmdiag"
set DSM_DIR=D:\tsm\Server\tsmdiag

rem Register the Server Licenses.

dsmadmc -id=%1 -pass=%2 -ITEMCOMMIT +
-OUTFILE="D:\TSM\SERVER1\[Link]" macro +
"D:\TSM\SERVER1\[Link]"

Chapter 32. Disaster recovery manager 1061

 rem Tell the server these active-data pool volumes are available for use.
rem Recovery Administrator: Remove from macro any volumes not obtained from vault.
dsmadmc -id=%1 -pass=%2 -ITEMCOMMIT +
-OUTFILE="D:\TSM\SERVER1\[Link]" +
macro "D:\TSM\SERVER1\[Link]"

rem Active-data pool volumes in this macro were not marked as ’offsite’ at the time
rem PREPARE ran. They were likely destroyed in the disaster.
rem Recovery Administrator: Remove from macro any volumes not destroyed.
dsmadmc -id=%1 -pass=%2 -ITEMCOMMIT +
-OUTFILE="D:\TSM\SERVER1\[Link]" +
macro "D:\TSM\SERVER1\[Link]"

rem Tell the server these copy storage pool volumes are available for use.
rem Recovery Administrator: Remove from macro any volumes not obtained from vault.
dsmadmc -id=%1 -pass=%2 -ITEMCOMMIT +
-OUTFILE="D:\TSM\SERVER1\[Link]" +
macro "D:\TSM\SERVER1\[Link]"

rem Copy storage pool volumes in this macro were not marked as ’offsite’ at the time
rem PREPARE ran. They were likely destroyed in the disaster.
rem Recovery Administrator: Remove from macro any volumes not destroyed.
dsmadmc -id=%1 -pass=%2 -ITEMCOMMIT +
-OUTFILE="D:\TSM\SERVER1\[Link]" +
macro "D:\TSM\SERVER1\[Link]"

rem Mark primary storage pool volumes as ACCESS=DESTROYED.

rem Recovery administrator: Remove from macro any volumes not destroyed.
dsmadmc -id=%1 -pass=%2 -ITEMCOMMIT +
-OUTFILE="D:\TSM\SERVER1\[Link]" +
macro "D:\TSM\SERVER1\[Link]"

rem Restore the previous working directory.

popd

rem Restore the previous working directory.

popd

:end
end [Link] script

Related tasks:
“Restoring to a point-in-time in a shared library environment” on page 915
“Scenario: Protecting the database and storage pools” on page 898
“Scenario: Recovering a lost or damaged storage pool volume” on page 910
“Example: Restoring a library manager database” on page 903
“Example: Restoring a library client database” on page 905
Related reference:
“Recovery instructions stanzas” on page 1057

Normal mode stanza

The [Link] stanza contains a script with the
commands needed to restore the server primary storage pools.

[Link]

You can use the script as a guide and run the commands from a command line. Or
you can copy it to a file, modify it and the files it refers to, and run the script. You
may need to modify the script because of differences between the original and the
replacement systems.

The disaster recovery plan issues commands using the administrative client.

1062 IBM Tivoli Storage Manager for Windows: Administrator's Guide

At the completion of these steps, client requests for file restores are satisfied from
primary storage pool volumes. Clients should also be able to resume file backup,
archive, and migration functions.

This script invokes the macros contained in the following stanzas:

[Link]
[Link]

Note: Because this script invokes the administrative client, you should ensure that
the communications options in the client options file are set to communicate with
the recovered server before running this script. To review the communications
options used in the recovered server, see the server options file in the
[Link] stanza.

To invoke this script, specify the following positional parameters:

v %1 (the administrator ID)
v %2 (the administrator password)
For example, to invoke this script using an administrator ID of don and a password
of mox, enter the following command:
[Link] don mox

For more information, see the entry for the recovery plan prefix in Table 79 on
page 1006.

The following stanza contains text strings that are too long to display in the
hardcopy or softcopy publications. The long text strings utilize a plus symbol (+)
to indicate string continuation on the next line.

Chapter 32. Disaster recovery manager 1063

 begin [Link] script

@echo off

rem Purpose: This script contains the steps required to recover the server
rem primary storage pools. This mode allows you to return the
rem copy storage pool volumes to the vault and to run the
rem server as normal.
rem Note: This script assumes that all volumes necessary for the restore
rem have been retrieved from the vault and are available. This script
rem assumes the recovery environment is compatible (essentially the
rem same) as the original. Any deviations require modification to this
rem this script and the macros and scripts it runs. Alternatively, you
rem can use this script as a guide, and manually execute each step.

if not %1.==. if not %2.==. goto start

echo Specify the following positional parameters:
echo administrative client ID and password.
echo Script stopped.
goto end
:start

rem Set the administrative command line client directory.

pushd "D:\tsm\Server\tsmdiag"
set DSM_DIR=D:\tsm\Server\tsmdiag

rem Define replacement volumes in the primary storage pools. Must

rem have different name than original.
rem Recovery administrator: Edit macro for your replacement volumes.
dsmadmc -id=%1 -pass=%2 -ITEMCOMMIT +
-OUTFILE="D:\TSM\SERVER1\[Link]" +
macro "D:\TSM\SERVER1\[Link]"

rem Restore the primary storage pools from the copy storage pools.
dsmadmc -id=%1 -pass=%2 -ITEMCOMMIT +
-OUTFILE="D:\TSM\SERVER1\[Link]" +
macro "D:\TSM\SERVER1\[Link]"

rem Restore the previous working directory.

popd

:end
end [Link] script

Related tasks:
“Restoring to a point-in-time in a shared library environment” on page 915
“Scenario: Protecting the database and storage pools” on page 898
“Scenario: Recovering a lost or damaged storage pool volume” on page 910
“Example: Restoring a library manager database” on page 903
“Example: Restoring a library client database” on page 905

Database directories stanza

The [Link] stanza identifies the directories for the Tivoli Storage
Manager database.

This stanza is referred to by the alternate DSMSERV RESTORE DB command in

the [Link] script. It is filled out with
the "locations" from QUERY DBSPACE . If you need to restore the database to an
alternate location, update this file with the new directories. You must also update
the [Link] script to invoke the
alternate DSMSERV RESTORE DB command instead of the default one.

1064 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 begin [Link]
E:\tsmdata\DBSpace
end [Link]

License registration stanza

The [Link] stanza contains a macro to register your server
licenses.

[Link]

This macro is invoked by the [Link]

script.
begin [Link] macro

/* Purpose: Register the server licenses by specifying the names */

/* of the enrollment certificate files necessary to re-create the */
/* licenses that existed in the server. */
/* Recovery Administrator: Review licenses and add or delete licenses */
/* as necessary. */

register license file([Link])

register license file([Link])
register license file([Link])

end [Link] macro

Copy storage pool volumes stanzas

The copy storage pool volumes stanzas contain macros to mark copy storage pool
volumes as available or unavailable.

[Link]

Contains a macro to mark copy storage pool volumes that were moved offsite and
then moved back onsite. This stanza does not include copy storage pool virtual
volumes. You can use the information as a guide and issue the administrative
commands, or you can copy it to a file, modify it, and run it. This macro is
invoked by the [Link] script.

After a disaster, compare the copy storage pool volumes listed in this stanza with
the volumes that were moved back onsite. You should remove entries from this
stanza for any missing volumes.
begin [Link] macro

/* Purpose: Mark copy storage pool volumes as available for use in recovery. */
/* Recovery Administrator: Remove any volumes that have not been obtained */
/* from the vault or are not available for any reason. */
/* Note: It is possible to use the mass update capability of the server */
/* UPDATE command instead of issuing an update for each volume. However, */
/* the ’update by volume’ technique used here allows you to select */
/* a subset of volumes to be processed. */

upd vol "003900L4" acc=READO wherestg=COPYPOOL

end [Link] macro

[Link]

Chapter 32. Disaster recovery manager 1065

 Contains a macro to mark copy storage pool volumes as unavailable if the volumes
were onsite at the time of the disaster. This stanza does not include copy storage
pool virtual volumes. These volumes are considered offsite and have not been
destroyed in a disaster. You can use the information as a guide and issue the
administrative commands from a command line, or you can copy it to a file,
modify it, and run it. This macro is invoked by the
[Link] script.

After a disaster, compare the copy storage pool volumes listed in this stanza with
the volumes that were left onsite. If you have any of the volumes and they are
usable, you should remove their entries from this stanza.
begin [Link] macro

/* Purpose: Mark destroyed copy storage pool volumes as unavailable. */

/* Volumes in this macro were not marked as ’offsite’ at the time the */
/* PREPARE ran. They were likely destroyed in the disaster. */
/* Recovery Administrator: Remove any volumes that were not destroyed. */

end [Link] macro

Active-data storage pool volumes stanzas

The active-data storage pool volumes stanzas contain macros to mark active-data
storage pool volumes as available or unavailable.

[Link]

Contains a macro to mark active-data pool volumes that were moved offsite and
then moved back onsite. This stanza does not include active-data pool virtual
volumes. You can use the information as a guide and issue the administrative
commands, or you can copy it to a file, modify it, and run it. This macro is
invoked by the [Link] script.

After a disaster, compare the active-data pool volumes listed in this stanza with the
volumes that were moved back onsite. You should remove entries from this stanza
for any missing volumes.
begin [Link] macro

/* Purpose: Mark active-data storage pool volumes as available for use in recovery. */
/* Recovery Administrator: Remove any volumes that have not been obtained */
/* from the vault or are not available for any reason. */
/* Note: It is possible to use the mass update capability of the server */
/* UPDATE command instead of issuing an update for each volume. However, */
/* the ’update by volume’ technique used here allows you to select */
/* a subset of volumes to be processed. */

upd vol "003901L4" acc=READO wherestg=ADP1

end [Link] macro

[Link]

Contains a macro to mark active-data pool volumes as unavailable if the volumes

were onsite at the time of the disaster. This stanza does not include active-data
pool virtual volumes. These volumes are considered offsite and have not been
destroyed in a disaster. You can use the information as a guide and issue the
administrative commands from a command line, or you can copy it to a file,

1066 IBM Tivoli Storage Manager for Windows: Administrator's Guide

modify it, and run it. This macro is invoked by the
[Link] script.

After a disaster, compare the active-data pool volumes listed in this stanza with the
volumes that were left onsite. If you have any of the volumes and they are usable,
you should remove their entries from this stanza.
begin [Link] macro

/* Purpose: Mark destroyed active-data storage pool volumes as unavailable. */

/* Volumes in this macro were not marked as ’offsite’ at the time the */
/* PREPARE ran. They were likely destroyed in the disaster. */
/* Recovery Administrator: Remove any volumes that were not destroyed. */

end [Link] macro

Primary storage pool volumes stanzas

These stanzas contain a macro to mark primary storage pool volumes as destroyed.

[Link]

Contains a macro to mark primary storage pool volumes as destroyed if the

volumes were onsite at the time of disaster. You can use the information as a guide
and run the administrative commands from a command line, or you can copy it to
a file, modify it, and run it. This macro is invoked by the
[Link] script.

During recovery, compare the primary storage pool volumes listed in this stanza
with the volumes that were onsite. If you have any of the volumes and they are
usable, remove their entries from the stanza.

This stanza does not include primary storage pool virtual volumes. These volumes
are considered offsite and have not been destroyed in a disaster.
begin [Link] macro

/* Purpose: Mark primary storage pool volumes as ACCESS=DESTROYED. */

/* Recovery administrator: Delete any volumes listed here */
/* that you do not want to recover. */
/* Note: It is possible to use the mass update capability of the server */
/* UPDATE command instead of issuing an update for each volume. However */
/* the ’update by volume’ technique used here allows you to select */
/* a subset of volumes to be marked as destroyed. */

vary offline "D:\[Link]" wait=yes

upd vol "D:\[Link]" acc=DESTROYED wherestg=PRIMPOOL1

end [Link] macro

[Link]

Contains a macro to define primary storage pool volumes to the server. You can
use the macro as a guide and run the administrative commands from a command
line, or you can copy it to a file, modify it, and execute it. This macro is invoked
by the [Link] script.

Primary storage pool volumes with entries in this stanza have at least one of the
following three characteristics:
v Original volume in a storage pool whose device class was DISK.

Chapter 32. Disaster recovery manager 1067

 v Original volume in a storage pool with MAXSCRATCH=0.
v Original volume in a storage pool and volume scratch attribute=no.

The SET DRMPLANVPOSTFIX command adds a character to the end of the names
of the original volumes listed in this stanza. This character does the following:
v Improves the retrievability of volume names that must be renamed in the
stanzas. Before using the volume names, change these names to new names that
are valid for the device class on the replacement system.
v Generates a new name that can be used by the replacement server. Your naming
convention must take into account the appended character.

Note:
1. Replacement primary volume names must be different from any other
original volume name or replacement name.
2. The RESTORE STGPOOL command restores storage pools on a logical basis.
There is no one-to-one relationship between an original volume and its
replacement.
3. There could be entries for the same volume in
[Link] if the volume has a device class of
DISK.

This stanza does not include primary storage pool virtual volumes. These volumes
are considered offsite and have not been destroyed in a disaster.

Primary storage volumes replacement stanza

begin [Link] macro

/* Purpose: Define replacement primary storage pool volumes for either: */

/* 1. Original volume in a storage pool whose device class was DISK. */
/* 2. Original volume in a storage pool with MAXSCRATCH=0. */
/* 3. Original volume in a storage pool and volume scratch=no. */
/* Recovery administrator: Edit this section for your replacement */
/* volume names. New name must be unique, i.e. different from any */
/* original or other new name. */

/* Replace D:\[Link] DISK 8,096.0M in PRIMPOOL1 */

def vol PRIMPOOL1 "D:\[Link]" acc=READW f=8,096 wait=yes

end [Link] macro

Storage pools restore stanza

This stanza contains a macro to restore the primary storage pools.

[Link]

You can use the stanza as a guide and execute the administrative commands from
a command line. You can also can copy it to a file, modify it, and execute it. This
macro is invoked by the [Link] script.

This stanza does not include primary storage pool virtual volumes. These volumes
are considered offsite and have not been destroyed in a disaster.

1068 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 egin [Link] macro

/* Purpose: Restore the primary storage pools from copy storage pool(s). */
/* Recovery Administrator: Delete entries for any primary storage pools */
/* that you do not want to restore. */

restore stgp PRIMPOOL1

end [Link] macro

Configuration stanzas
These stanzas contain copies of the following information: volume history, device
configuration, and server options.

[Link]

Contains a copy of the volume history information when the recovery plan was
created. The DSMSERV RESTORE DB command uses the volume history file to
determine what volumes are needed to restore the database. It is used by the
[Link] script.

The following rules determine where to place the volume history file at restore
time:
v If the server option file contains VOLUMEHISTORY options, the server uses the
fully qualified file name associated with the first entry. If the file name does not
begin with a directory specification, the server uses the prefix volhprefix.
v If the server option file does not contain VOLUMEHISTORY options, the server
uses the default name volhprefix followed by [Link]. The volhprefix is set to
the directory representing this instance of the server, which is typically the
directory that the server was originally installed from.

If a fully qualified file name was not specified in the server options file for the
VOLUMEHISTORY option, the server adds it to the [Link] stanza.

Chapter 32. Disaster recovery manager 1069

 ***********************************************************************
*
* Sequential Volume Usage History
* Updated 12/13/2008 [Link]
*
***********************************************************************
Operation Date/Time: 2008/11/13 [Link]

Volume Type: BACKUPFULL

* Location for volume C:\TEST318\DUMP\[Link] is: ’NOTMOUNTABLE’
Database Backup LLA: FULL_BACKUP.20081113144416.1
Database Backup HLA: \NODE0000\
Volume Name: "C:\TEST318\DUMP\[Link]"
Backup Series: 1
Backup Op: 0
Volume Seq: 1
Device Class Name: FILES
Database Backup ID: 0,1025
Database Backup Home Position: 0
Database Backup Total Data Bytes: 0,327385099
Database Backup Total Log Bytes: 0,2125835
Database Backup Log Block Number: 4294967295,4294967295
**************************************************
Operation Date/Time: 2008/11/13 [Link]

Volume Type: BACKUPFULL

* Location for volume C:\TEST318\DUMP\[Link] is: ’NOTMOUNTABLE’
Database Backup LLA: FULL_BACKUP.20081113144416.2
Database Backup HLA: \NODE0000\
Volume Name: "C:\TEST318\DUMP\[Link]"
Backup Series: 1
Backup Op: 0
Volume Seq: 2
Device Class Name: FILES
Database Backup ID: 0,1026
Database Backup Home Position: 0
Database Backup Total Data Bytes : 0,327385099
Database Backup Total Log Bytes: 0,2125835
Database Backup Log Block Number: 4294967295,4294967295
**************************************************

[Link]

Contains a copy of the server device configuration information when the recovery
plan was created. The DSMSERV RESTORE DB command uses the device
configuration file to read the database backup volumes. It is used by the
[Link] script.

At recovery time, you may need to modify this stanza. You must update the device
configuration information if the hardware configuration at the recovery site has
changed. Examples of changes requiring updates to the configuration information
are:
v Different device names
v Use of a manual library instead of an automated library
v For automated libraries, the requirement to manually place the database backup
volumes in the automated library and update the configuration information to
identify the element within the library. This allows the server to locate the
required database backup volumes.
For details, see “Updating the device configuration file” on page 906.

The following rules determine where the device configuration file is placed at
restore time:

1070 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 v If the server options file contains DEVCONFIG entries, the server uses the fully
qualified file name associated with the first entry. If the specified file name does
not begin with a directory specification, the server adds the prefix devcprefix.
v If the server options file does not contain DEVCONFIG entries, the server uses
the default name devcprefix followed by [Link].
For example, if devcprefix is c:\Program Files\Tivoli\TSM\server2\ the file name
used by PREPARE is c:\Program Files\Tivoli\TSM\server2\[Link].

Note: The devcprefix is set to the directory representing this instance of the server
which is typically the directory from which the server was originally installed.

If a fully qualified file name was not specified for the DEVCONFIG option in the
server options file, the server adds it to the stanza [Link].
begin [Link]

/* Tivoli Storage Manager Device Configuration */

DEFINE DEVCLASS COOL8MM DEVTYPE=8MM FORMAT=DRIVE MOUNTLIMIT=1 MOUNTWAIT=60
MOUNTRETENTION=60 PREFIX=TIVSM LIBRARY=ITSML
DEFINE DEVCLASS FILES DEVTYPE=FILE MAXCAPACITY=4096K MOUNTLIMIT=2 DIRECTORY=c:\Program Files\Tivoli\TSM\server2\stg
DEFINE DEVCLASS FILESSM DEVTYPE=FILE MAXCAPACITY=100K MOUNTLIMIT=2 DIRECTORY=c:\Program Files\Tivoli\TSM\server2\stg
DEFINE DEVCLASS LIB8MM DEVTYPE=8MM FORMAT=DRIVE MOUNTLIMIT=1 MOUNTWAIT=60 MOUNTRETENTION=60 PREFIX=TIVSM
LIBRARY=RLLIB
DEFINE LIBRARY ITSML LIBTYPE=MANUAL
DEFINE LIBRARY RLLIB LIBTYPE=MANUAL
SERVERBACKUPNODEID 1

end [Link]

[Link]

Contains a copy of the server options file. This stanza is used by the
[Link] script.

Note: The following figure contains text strings that are too long to display in
hardcopy or softcopy publications. The long text strings have a plus symbol (+) at
the end of the string to indicate that they continue on the next line.

The disaster recovery plan file adds the DISABLESCHEDS option to the server
options file and sets it to YES. This option disables administrative and client
schedules while the server is being recovered. After the server is recovered, you
can enable scheduling by deleting the option or setting it to NO and then
restarting the server.
begin [Link]

* Server options file located in c:\Program Files\Tivoli\TSM\server2\[Link]

TCPPort 1509
VOLUMEHISTORY c:\Program Files\Tivoli\TSM\server2\[Link]
DEVCONFIG c:\Program Files\Tivoli\TSM\server2\[Link]
* The following option was added by PREPARE.
DISABLESCHEDS YES

end [Link]

Chapter 32. Disaster recovery manager 1071

 License information stanza
The license information stanza contains a copy of the latest license audit results
and the server license terms.

[Link]
begin [Link]

Last License Audit: 09/26/2008 [Link]

Is IBM System Storage Archive Manager in use ?: No
Is IBM System Storage Archive Manager licensed ?: Yes
Is Tivoli Storage Manager Basic Edition in use: Yes
Is Tivoli Storage Manager Basic Edition licensed: Yes
Is Tivoli Storage Manager Extended Edition in use: No
Is Tivoli Storage Manager Extended Edition licensed: Yes
Server License Compliance: Valid

end [Link]

Machine files stanza

These stanzas contain information needed to rebuild the server machine.

[Link]

Provides information for the server machine (for example, machine location). This
stanza is included in the plan file if the machine information is saved in the
database using the DEFINE MACHINE with ADSMSERVER=YES.
begin [Link]
Purpose: General information for machine DSMSRV1.
This is the machine that contains DSM server DSM.
Machine Name: DSMSRV1
Machine Priority: 1
Building: 21
Floor: 2
Room: 2749
Description: DSM Server for Branch 51
Recovery Media Name: DSMSRVIMAGE

end [Link]

[Link]

Provides the recovery instructions for the server machine. This stanza is included
in the plan file if the machine recovery instructions are saved in the database.
begin [Link]
Purpose: Recovery instructions for machine DSMSRV1.

Primary Contact:
Jane Smith (wk 520-000-0000 hm 520-001-0001)
Secondary Contact:
John Adams (wk 520-000-0001 hm 520-002-0002)

end [Link]

[Link]

Provides the hardware and software characteristics for the server machine. This
stanza is included in the plan file if the machine characteristics are saved in the
database.

1072 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 begin [Link]
Purpose: Hardware and software characteristics of machine DSMSRV1.
Processor : x86 Family 5 Model 2 Stepping 11
Bus Type : AT 16-Bit bus
Keyboard Type : 101/102 Key Enhanced Keyboard
Pointer Type : PS/2 Mouse Buttons: 2
Equipment : 1 Parallel Port(s)
1 Serial Port(s)
1 CD Drive(s)
2 Fixed Disk(s)
Pointing Device
Math CoProcessor
Fixed Disk 1 : 609 MB
Total Physical Memory : 32,832 KB (33,619,968)

end [Link]

[Link]

Provides information about the media (for example, boot media) needed for
rebuilding the machine that contains the server. This stanza is included in the plan
file if recovery media information is saved in the database and it has been
associated with the machine that contains the server.
begin [Link]
Purpose: Recovery media for machine DSMSRV1.
Recovery Media Name: DSMSRV
Type: Other
Volume Names:
Location: IRONMNT
Description: Server Installation CD
Product:
Product Information:

end [Link]

Chapter 32. Disaster recovery manager 1073

1074 IBM Tivoli Storage Manager for Windows: Administrator's Guide
Chapter 33. Integrating disaster recovery manager and node
replication into your disaster recovery strategy
Many tools, architectures, and technologies are available for you to use to
implement a disaster recovery strategy. For example, you can use disaster recovery
manager (DRM) to vault of offsite volumes, generate a disaster recovery plan, and
automate the recovery of the Tivoli Storage Manager server. To maintain the same
level of files on two networked servers, you can use Tivoli Storage Manager node
replication.

The framework for evaluating disaster recovery strategies consists of the following
tiers:

Tier 0 -- Do nothing, no offsite data

Tier 1 -- Offsite vaulting

Tier 2 -- Offsite vaulting with a hot site

Tier 3 -- Electronic vaulting

Tier 4 -- Electronic vaulting to hot site (active secondary site)

Tier 5 -- Two-site, two-phase commit

Tier 6 -- Zero data loss

Figure 117. Tiers of disaster recovery

Each tier corresponds to different recovery times and potentials for data loss. For
example, in a tier 1 production site data is typically saved only selectively, and
volumes that are stored at an offsite facility can be difficult to track. In addition,
recovery time is unpredictable. After a disaster, hardware and software must be
restored, and storage volumes must be sent back to the production site.

A tier 6 strategy consists of synchronous remote-copy technologies between peer

production sites. Sites with tier 6 protection offer fast recovery times with no loss
of data. Data is considered lost only if a transaction started, for example if the user
presses Enter, but the transaction request was not completed at both sites.

© Copyright IBM Corp. 1993, 2013 1075

Plan for a disaster recovery strategy
For optimal results, plan the integration of disaster recovery manager (DRM) and
node replication.

Use the following questions as a guide to help you in the planning process:
Cost How much can you afford for your disaster recovery implementation?
Performance
Do you want a low or a high performance disaster recovery solution?
Recovery Time Objective (RTO) and Recovery Point Objective (RPO)
What are your system requirements?
Current disaster recovery strategy
What disaster recovery strategy is implemented in your environment?
Data What data do you need? Categorize and prioritize the data that you
require.

When you plan a disaster recovery strategy that might be suitable for your site,
consider using DRM and Tivoli Storage Manager node replication for these
reasons:
v DRM is an effective tool for managing offsite vaulting. With DRM, you can
configure and automatically generate a disaster recovery plan that contains the
information, scripts, and procedures that are required to automatically restore
the server and recover client data after a disaster.
DRM also manages and tracks the media on which client data is stored, whether
the data is on site, in-transit, or in a vault, so that the data can be more easily
located if disaster strikes. DRM also generates scripts that assist you in
documenting information-technology systems and recovery procedures that you
can use, including procedures to rebuild the server.
Use DRM alone to meet the disaster recovery objectives in tier 1, or use it
together with other backup-and-recovery tools and technologies in tiers 2, 3 and
4.
v Tivoli Storage Manager node replication meets the objectives of tier 5. After a
successful node replication, the target server contains all metadata updates and
data that is stored on the source server.
In addition to fast recovery and minimal potential data loss, Tivoli Storage
Manager node replication offers the following advantages:
– Node replication is easier to manage than device-based replication.
Device-based replication requires that you keep the database and the data it
represents synchronized. You manually schedule database backups to match
the point in time when the device synchronizes.
– Results for Tivoli Storage Manager operations are reported in terms such as
"node names" and "file names." In contrast, device-based replication results
are reported in terms of "disks," "sectors," and "blocks."

1076 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Tier 0: No disaster recovery capability
Tier 0 is defined as a single-site data center that does not back up data or, if data is
backed up, backups are left on site. The data center does not have a disaster
recovery plan or backup hardware and software.

In the following figure, the Tivoli Storage Manager server and database, tape
libraries, and tapes are in a single facility. If a disaster occurs, recovery time is
unpredictable. Tier 0 is not recommended and data might never be recovered.

Data center A

Figure 118. Tier 0: No disaster recovery capability

Tier 1: Offsite vaulting from a single production site

A tier-1 disaster-recovery strategy requires a disaster recovery plan and careful
management of volumes that are stored in an offsite facility. To manage offsite
volumes, generate a disaster recovery plan, and automate the Tivoli Storage
Manager server recovery process, use disaster recovery manager.

As shown in the following figure, storage volumes, such as tape cartridges and
media volumes, and are vaulted at an offsite location. Transportation is typically
handled by couriers. If a disaster occurs, the volumes are sent back to the
production site after hardware and the Tivoli Storage Manager server is restored.

Data center A

DRM
Offsite vault

Daily backups

Figure 119. Tier 1: Offsite vaulting from a single production site

Because it is relatively inexpensive, a tier 1 strategy is used by many organizations.

However, tier 1 capability is low and recovery time is slow. The recovery time
depends on how long it takes to restore the hardware and software and to
transport and load tapes into libraries. Before you implement a tier 1 strategy,

Chapter 33. Integrating disaster recovery manager and node replication into your disaster recovery strategy 1077
 consider that an extended recovery time can impact business operations for several
months or longer.

Tier 2: Offsite vaulting with a recovery site

A tier-2 disaster-recovery strategy includes a physically separate recovery site in
addition to offsite vaulting. The recovery site has the hardware and software that
are necessary to recover client data. To manage offsite volumes, generate a disaster
recovery plan, and automate the server recovery process, use disaster recovery
manager.

Like tier 1 organizations, tier 2 organizations rely on a courier to move storage

volumes from the production site to an offsite storage facility. However, if a
disaster occurs, tier 1 organizations move the data from the offsite storage facility
to the recovery site. A Tier 2 disaster recovery strategy has low capability and is
slow. See the following figure:

Data center A Data center B

(hotsite)
DRM
Offsite vault

Daily At recovery
backups time

Figure 120. Tier 2: Offsite vaulting with a recovery site

A dedicated recovery site can reduce recovery time compared to the single
production site in tier 1. The potential for data loss is also less. However, tier 2
architecture increases the cost of disaster recovery because more hardware and
software must be maintained. The recovery site must also have hardware and
software that are compatible with the hardware and software at the primary site.
For example, the recovery site must have compatible tape devices and Tivoli
Storage Manager server software. Before the production site can be recovered, the
hardware and software at the recovery site must be set up and running.
Transporting the storage volumes to the recovery site also affects recovery time.

Consider using Tivoli Storage Manager server-to-server communications for

enterprise configuration of the Tivoli Storage Manager servers at the primary and
recovery sites and routing commands from server to server.

Tier 3: Electronic vaulting of critical data

A tier-3 disaster-recovery strategy includes a recovery site with a running Tivoli
Storage Manager server. Critical data is vaulted electronically from the production
site to the recovery site. Disaster recovery manager is also used for offsite vaulting
of non-critical data.

Electronic vaulting moves critical data offsite faster and more frequently than
traditional courier methods. Recovery time is reduced because critical data is
already stored at the recovery site. The potential for lost or misplaced data is also
reduced. However, because the recovery site runs continuously, a tier 3 strategy is
relatively more expensive than a tier 1 or a tier 2 strategy.

1078 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 Critical data can include database backups, infrastructure setup files, and active
client-backup data. To vault data electronically, consider using Tivoli Storage
Manager virtual volumes over a TCP/IP connection. Data is stored as archive files
on the server at the recovery site.

As shown in the following figure, the recovery site is physically separated from the
production site. Often, the recovery site is a second data center that is operated by
the same organization or by a storage service provider. If a disaster occurs at the
primary site, storage media with the non-critical data are transported from the
offsite storage facility to the recovery site.

Electronic vaulting

Data center A Data center B

(hotsite)
DRM
Offsite vault

Daily At recovery
backups time

Figure 121. Tier 3: Electronic vaulting of critical data

If you implement a tier 3 strategy, you can use Tivoli Storage Manager
server-to-server communications for enterprise configuration of the Tivoli Storage
Manager servers and command routing.

Tier 4: Active data management at peer sites

In a tier-4 disaster-recovery strategy, data is actively managed at two physically
separate sites. The servers at each site are set up in a peer-to-peer relationship.
Critical data is replicated asynchronously. Copies of the data are available at both
sites, and each server is able to recover the server at the alternate site. As part of
this strategy, backup storage media are vaulted offsite and tracked by using
disaster recovery manager.

As shown in the following figure, critical data is replicated the two sites by using
high-bandwidth connections and data replication technology, for example,
Peer-to-Peer Remote Connection (PPRC). Data is transmitted over long distances
by using technologies such as extended storage area network (SAN), Dense Wave
Division Multiplexing (DWDM), and IP/WAN channel extenders.

Chapter 33. Integrating disaster recovery manager and node replication into your disaster recovery strategy 1079
 High bandwidth connections
Data center A Data center B
(hotsite)
DRM
Daily backups

Offsite
vault

At recovery time

Figure 122. Tier 4: Active data management at peer-to-peer sites

Non-critical backups from both sites are moved to a single offsite storage facility. If
a disaster occurs, the backup volumes are recovered by courier from the offsite
vault and transported the designated recovery site.

If you implement a tier-4 disaster-recovery strategy, you can use Tivoli Storage
Manager server-to-server communications for enterprise configuration of multiple
Tivoli Storage Manager servers and command routing.

Recovery time for a tier 4 strategy is faster than the recovery time for a tier 1, tier
2, or tier 3 strategy. Recovery time is faster because hardware, software, and data
are available or can be made available at two sites.

Tier 5: Synchronous replication

Tier 5 encompasses all the requirements of tier 4, including peer-peer sites.
However, data, including database backups and copy storage pools, is replicated
synchronously. Selected data is maintained in image status so that updates are
applied to both the local and the remote copies of the database. Data, metadata,
and the inventory information for the database are synchronized. The data at both
the sites must be updated before the update request is considered successful.
Physical media is not vaulted offsite.

As shown in the following figure, data is transmitted continuously between the

two sites and synchronized by using high-bandwidth connections such as
Peer-to-Peer Remote Copy (PPRC). Transmission over long distances is achieved by
using technologies such as extended storage area network (SAN), Dense Wave
Division Multiplexing (DWDM), and IP/WAN channel extenders. Copies of critical
data are available at both sites, and each server is able to recover the server at the
alternate site. With a minimum amount of data to recover, recovery time is reduced
significantly.

1080 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 Data center B
Data center A (hotsite)

High bandwidth
connections

Figure 123. Tier 5: Synchronous replication

Copies of critical data are available at both sites, and each server is able to recover
the server at the alternate site. Only the data transactions that are in-flight are lost
during a disaster.

If you implement a tier-5 disaster-recovery strategy, you can also use Tivoli Storage
Manager server-to-server communications to configure multiple Tivoli Storage
Manager servers and command routing.

Tier 6: Synchronous replication with coupling or clustering

applications
In Tier 6, local and remote copies of all data are updated synchronously, and dual
online storage is used with a full network-switching capability. Systems are
coupled with automated failover and failback capabilities when required.

As shown in the following figure, two sites are fully synchronized by using a
high-bandwidth connection.

Advanced Data center B

Data center A coupled/clustered (hotsite)
systems

High bandwidth connections

Data sharing

Figure 124. Tier 6: Synchronous replication with coupling or clustering applications

Tier 6 is the most expensive disaster recovery strategy because it requires coupling
or clustering applications, additional hardware to support data sharing, and
high-bandwidth connections over extended distances. However, this strategy also
offers the fastest recovery time and the least amount of data loss. The typical
length of time for recovery is normally a few minutes.

Chapter 33. Integrating disaster recovery manager and node replication into your disaster recovery strategy 1081
1082 IBM Tivoli Storage Manager for Windows: Administrator's Guide
Part 7. Appendixes

© Copyright IBM Corp. 1993, 2013 1083

1084 IBM Tivoli Storage Manager for Windows: Administrator's Guide
Appendix A. External media management interface
description
The programming interface that IBM Tivoli Storage Manager provides to external
media-management programs consists of request description strings that IBM
Tivoli Storage Manager sends and response strings that the external program
sends.

To use the interface, you must first define an EXTERNAL-type Tivoli Storage
Manager library that represents the media manager. You do not define drives, label
volumes, or check in media. Refer to your media manager's documentation for that
product's setup information and instructions for operational usage.

The details of the request types and the required processing are described in the
sections that follow. The request types are:
v Initialization of the external program
v Begin Batch
v End Batch
v Volume Query
v Volume Eject
v Volume Release
v Volume Mount
v Volume Dismount

The responses can be right-padded with any number of white-space characters.

The libraryname passed in a request must be returned in the response. The volume
specified in an eject request or a query request must be returned in the response.
The volume specified in a mount request (except for 'SCRTCH') must be returned
in the response. When 'SCRTCH' is specified in a mount request, the actual volume
mounted must be returned.

CreateProcess call
The server creates two anonymous unidirectional pipes and maps them to the
stdin and stdout streams during the CreateProcess call. When a standard handle is
redirected to refer to a file or a pipe, the handle can only be used by the ReadFile
and WriteFile functions.

This precludes normal C functions such as gets or printf. Since the server will
never terminate the external program process, it is imperative that the external
program recognize a read or write failure on the pipes and exit the process. In
addition, the external program should exit the process if it reads an unrecognized
command.

The external program may obtain values for the read and write handles using the
following calls:
readPipe=GetStdHandle(STD_INPUT-HANDLE)

and

© Copyright IBM Corp. 1993, 2013 1085

 writePipe=GetStdHandle(STD_OUTPUT_HANDLE)

Processing during server initialization

Ensure that the external media management program works with the server during
the server’s initialization.

About this task

For each external library defined to the server, the following must occur during
server initialization:

Procedure
1. The server loads the external program (CreateProcess) in a newly created
process and creates pipes to the external program.
2. The server sends an initialization request description string, in text form, into
the standard input (stdin) stream of the external program. The server waits for
the response.
3. When the external process completes the request, the process must write an
initialization response string, in text form, into its standard output (stdout)
stream.
4. The server closes the pipes.
5. When the agent detects that the pipes are closed, it performs any necessary
cleanup and calls the stdlib exit routine.

Processing for mount requests

To process the mount request, the server has to complete certain steps.

About this task

The server completes the following steps to process mount requests:

Procedure
1. The server loads the external program in a newly created process and creates
pipes to the external program.
2. The server sends an initialization request description string (in text form) into
the standard input (stdin) stream of the external program. The server waits for
the response.
3. When the external process completes the request, the process must write an
initialization response string (in text form) into its standard output (stdout)
stream.
4. The server sends the MOUNT request (stdin).
5. The agent sends the MOUNT response (stdout).
6. The agent waits.
7. The server sends the DISMOUNT request (stdin).
8. The agent sends the DISMOUNT response (stdout), performs any necessary
cleanup, and calls the stdlib exit routine.

1086 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Processing for release requests
To process release requests, the server has to complete certain steps.

About this task

The server completes the following steps to process release requests:

Procedure
1. The server loads the external program in a newly created process and creates
pipes to the external program.
2. The server sends an initialization request description string (in text form) into
the standard input (stdin) stream of the external program. The server waits for
the response.
3. When the external process completes the request, the process must write an
initialization response string (in text form) into its standard output (stdout)
stream.
4. The server sends the RELEASE request (stdin).
5. The agent sends the RELEASE response (stdout), performs any necessary
cleanup, and calls the stdlib exit routine.

Processing for batch requests

Batch processing is done during MOVE MEDIA, MOVE DRMEDIA, and QUERY
MEDIA command processing when performed on volumes in external libraries.

About this task

The move commands cause a QUERY to be issued for a volume. If the QUERY
indicates that the volume is in the library, a subsequent EJECT for that volume is
issued. Because the move commands can match any number of volumes, a QUERY
and an EJECT request is issued for each matching volume.

The QUERY MEDIA command results in QUERY requests being sent to the agent.
During certain types of processing, Tivoli Storage Manager might need to know if
a volume is present in a library. The external agent should verify that the volume
is physically present in the library.

Procedure
1. The server loads the external program in a newly created process and creates
pipes to the external program.
2. The server sends an initialization request description string (in text form) into
the standard input (stdin) stream of the external program. The server waits for
the response.
3. When the external process completes the request, the process must write an
initialization response string (in text form) into its standard output (stdout)
stream.
4. The server sends the BEGIN BATCH request (stdin).
5. The agent sends the BEGIN BATCH response (stdout).
6. The server sends 1 to n volume requests (n > 1). These can be any number of
QUERY or EJECT requests. For each request, the agent will send the applicable
QUERY response or EJECT response.
7. The server sends the END BATCH request (stdin).

Appendix A. External media management interface description 1087

 8. The agent sends the END BATCH response (stdout), performs any necessary
cleanup, and calls the stdlib exit routine.

Error handling
If the server encounters an error during processing, it closes the stdin and stdout
streams to the agent exit. The agent detects this when it attempts to read from
stdin or write to stdout. If this occurs, the agent performs any necessary cleanup
and calls the stdlib exit routine.

If the code for any response (except for EJECT and QUERY) is not equal to
SUCCESS, Tivoli Storage Manager does not proceed with the subsequent steps.
After the agent sends a non-SUCCESS return code for any response, the agent will
perform any necessary cleanup and call the stdlib exit routine.

However, even if the code for EJECT or QUERY requests is not equal to SUCCESS,
the agent will continue to send these requests.

If the server gets an error while trying to write to the agent, it will close the pipes,
perform any necessary cleanup, and terminate the current request.

Begin batch request

The programming interface includes a begin-batch description string and an
external-program response string.

The format of the begin batch request is:

BEGIN BATCH

The format of the external program response is:

BEGIN BATCH COMPLETE, RESULT=resultCode

where:
resultCode
One of the following:
v SUCCESS
v INTERNAL_ERROR

End batch request

The end batch request is sent by Tivoli Storage Manager to indicate that no more
requests are to be sent by the external library manager for the current process. The
external agent must send the end batch response and end by using the stdlib exit
routine.

The format of the end batch request is:

END BATCH

The format of the external program response is:

END BATCH COMPLETE, RESULT=resultCode

where resultCode is SUCCESS or INTERNAL_ERROR.

1088 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Volume query request
The programming interface includes a volume-query-request description string and
an external-program response string.

The format of the volume query request is:

QUERY libraryname volume

where:
libraryname
Specifies the name of the EXTERNAL library as defined to Tivoli Storage
Manager.
volume
Specifies the volume name to be queried.

The format of the external program response is:

QUERY libraryname volume COMPLETE, STATUS=statusValue, RESULT=resultCode

where:
libraryname
Specifies the name of the EXTERNAL library as defined to Tivoli Storage
Manager.
volume
Specifies the volume name queried.
resultCode
One of the following:
v SUCCESS
v LIBRARY_ERROR
v VOLUME_UNKNOWN
v VOLUME_UNAVAILABLE
v CANCELLED
v TIMED_OUT
v INTERNAL_ERROR

If resultCode is not SUCCESS, the exit must return statusValue set to UNDEFINED.
If resultCode is SUCCESS, STATUS must be one of the following values:
v IN_LIBRARY
v NOT_IN_LIBRARY

IN_LIBRARY means that the volume is currently in the library and available to be
mounted.

NOT_IN_LIBRARY means that the volume is not currently in the library.

Appendix A. External media management interface description 1089

Initialization requests
When the server is started, it sends an initialization request to the external media
management program for each EXTERNAL library. The external program processes
this request to ensure that the external program is present, functional, and ready to
process requests.

If the initialization request is successful, Tivoli Storage Manager informs its

operators that the external program reported its readiness for operations.
Otherwise, Tivoli Storage Manager reports a failure to its operators.

Tivoli Storage Manager does not attempt any other type of operation with that
library until an initialization request has succeeded. The server sends an
initialization request first. If the initialization is successful, the request is sent. If the
initialization is not successful, the request fails. The external media management
program can detect whether the initialization request is being sent by itself or with
another request by detecting end-of-file on the stdin stream. When end-of-file is
detected, the external program must end by using the stdlib exit routine (not the
return call).

When a valid response is sent by the external program, the external program must
end by using the exit routine.

Format of the request:

INITIALIZE libraryname

where libraryname is the name of the EXTERNAL library as defined to Tivoli

Storage Manager.

Format of the external program response:

INITIALIZE libraryname COMPLETE, RESULT=resultcode

where:
libraryname
Specifies the name of the EXTERNAL library as defined to Tivoli Storage
Manager.
resultcode
One of the following:
v SUCCESS
v NOT_READY
v INTERNAL_ERROR

Volume eject request

The Programming Interface includes a volume-eject-request description string and
an external-program response string.

The format of the volume eject request is:

EJECT libraryname volume ’location info’

where:

1090 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 libraryname
Specifies the name of the EXTERNAL library as defined to Tivoli Storage
Manager.
volume
Specifies the volume to be ejected.
'location info'
Specifies the location information associated with the volume from the Tivoli
Storage Manager inventory. It is delimited with single quotation marks. This
information is passed without any modification from the Tivoli Storage
Manager inventory. The customer is responsible for setting its contents with
the appropriate UPDATE MEDIA or UPDATE VOLUME command before the
move command is invoked. Set this field to some target location value that will
assist in placing the volume after it is ejected from the library. It is suggested
that the external agent post the value of this field to the operator.

The format of the external program response is:

EJECT libraryname volume COMPLETE, RESULT=resultCode

where:
libraryname
Specifies the name of the EXTERNAL library as defined to Tivoli Storage
Manager.
volume
Specifies the ejected volume.
resultCode
One of the following:
v SUCCESS
v LIBRARY_ERROR
v VOLUME_UNKNOWN
v VOLUME_UNAVAILABLE
v CANCELLED
v TIMED_OUT
v INTERNAL_ERROR

Volume release request

When the server returns a volume to scratch status, the server starts the external
media management program, issues a request to initialize, then issues a request to
release a volume.

The external program must send a response to the release request. No matter what
response is received from the external program, Tivoli Storage Manager returns the
volume to scratch. For this reason, Tivoli Storage Manager and the external
program can have conflicting information on which volumes are scratch. If an error
occurs, the external program should log the failure so that the external library
inventory can be synchronized later with Tivoli Storage Manager. The
synchronization can be a manual operation.

The format of the request is:

RELEASE libraryname volname

Appendix A. External media management interface description 1091

 where:
libraryname
Specifies the name of the EXTERNAL library as defined to Tivoli Storage
Manager.
volname
Specifies the name of the volume to be returned to scratch (released).

The format of the external program response is:

RELEASE libraryname volname COMPLETE, RESULT=resultcode

where:
libraryname
Specifies the name of the EXTERNAL library as defined to Tivoli Storage
Manager.
volname
Specifies the name of the volume returned to scratch (released).
resultcode
One of the following:
v SUCCESS
v VOLUME_UNKNOWN
v VOLUME_UNAVAILABLE
v INTERNAL_ERROR

Volume mount request

The Tivoli Storage Manager server can mount volumes as needed. Alternatively,
volumes can be manually mounted by issuing the MOUNT command.

When the server requires a volume mount, the server starts the external media
management program, issues a request to initialize, then issues a request to mount
a volume. The external program is responsible for verifying that this request is
coming from Tivoli Storage Manager and not from an unauthorized system.

The volume that is mounted by the external media management program must be
a tape with a standard IBM label that matches the external volume label. When the
external program completes the mount request, the program must send a response.
If the mount was successful, the external program must remain active. If the
mount failed, the external program must end immediately by using the stdlib exit
routine.

The request must be in the following format:

MOUNT libraryname volname accessmode devicetypes timelimit userid
volumenumber ’location’

where
libraryname
Specifies the name of the EXTERNAL library as defined to Tivoli Storage
Manager.
volname
Specifies the actual volume name if the request is for an existing volume. If a
scratch mount is requested, the volname is set to SCRTCH.

1092 IBM Tivoli Storage Manager for Windows: Administrator's Guide

accessmode
Specifies the access mode that is required for the volume. Possible values are
READONLY and READWRITE.
devicetypes
Specifies a list of device types that can be used to satisfy the request for the
volume and the FORMAT specified in the device class. The most preferred
device type must be specified first in the list. Items are separated by commas
with no intervening spaces. Possible values are:
v 3590
v 3590E
v 3590H
v 3592
v 4MM_DDS1
v 4MM_DDS1C
v 4MM_DDS2
v 4MM_DDS2C
v 4MM_DDS3
v 4MM_DDS3C
v 4MM_HP_DDS4
v 4MM_DDS4
v 4MM_HP_DDS4C
v 4MM_DDS4C
v 4MM_DDS5C
v 4MM_DDS6C
v 8MM_SAIT
v 8MM_VXA2
v 8MM_VXA3
v 8MM_AITC
v 8MM_AIT
v 8MM_8200
v 8MM_ELIANT
v 8MM_8500
v 8MM_8500C
v 8MM_8205
v 8MM_8900
v 8MM_M2
v DLT_2000
v DLT_4000
v DLT_7000
v DLT_8000
v SDLT
v SDLT320
v DLT1
v DLT2
v SDLT600
v GENERICTAPE

Appendix A. External media management interface description 1093

 v LTO_ULTRIUM
v LTO_ULTRIUM2
v LTO_ULTRIUM3
v LTO_ULTRIUM4
v LTO_ULTRIUM5
v LTO_ULTRIUM6
v REMOVABLEFILE
v M8100
v STK_SD3
v STK_9840
v STK_T9840C
v STK_9940
v STK_9940B
v STK_9840_VOLSAFE
v STK_T9840C_VOLSAFE
v STK_9940_VOLSAFE
v STK_9940B_VOLSAFE
v STK_T10000A
v STK_T10000B
v STK_T10000C
timelimit
Specifies the maximum number of minutes that the server waits for the
volume to be mounted. If the mount request is not completed within this time,
the external manager responds with the result code TIMED_OUT.
userid
Specifies the user ID of the process that needs access to the drive.
volumenumber
The volume number is 1.
'location'
Specifies the value of the location field from the Tivoli Storage Manager
inventory. For example, 'Room 617 Floor 2'. One blank character is inserted
between the volume number and the left single quotation mark in the location
information. If no location information is associated with a volume, nothing is
passed to the exit. If no volume information exists, the single quotation marks
are not passed. Also, if volume information is passed, the volume might have
been ejected from the library and it must be returned to the library before the
mount operation can proceed. The location information is posted by the agent
so that the operator can obtain the volume and return it to the library.

The external program response is in the following format:

MOUNT libraryname volname COMPLETE ON specialfile, RESULT=resultcode

where
libraryname
Specifies the name of the EXTERNAL library as defined to Tivoli Storage
Manager.
volname
Specifies the name of the volume that is mounted for the request.
1094 IBM Tivoli Storage Manager for Windows: Administrator's Guide
 specialfile
The fully qualified path name of the device special file for the drive in which
the volume was mounted. If the mount request fails, set the value to
/dev/null.
The external program must ensure that the special file is closed before the
response is returned to the server.
resultcode
The result code can be one of the following values:
v SUCCESS
v DRIVE_ERROR
v LIBRARY_ERROR
v VOLUME_UNKNOWN
v VOLUME_UNAVAILABLE
v CANCELLED
v TIMED_OUT
v INTERNAL_ERROR

Volume dismount request

When a successful mount operation completes, the external process must wait for a
request to dismount the volume. When the dismount operation completes, the
external program must send a response to the server.

After the dismount response is sent, the external process ends immediately by
using the stdlib exit routine.

Format of the request:

DISMOUNT libraryname volname

where:
libraryname
Specifies the name of the EXTERNAL library as defined to Tivoli Storage
Manager.
volname
Specifies the name of the volume to be dismounted.

Format of the external program response:

DISMOUNT libraryname volname COMPLETE, RESULT=resultcode

where:
libraryname
Specifies the name of the EXTERNAL library as defined to Tivoli Storage
Manager.
volname
Specifies the name of the volume dismounted.
resultcode
One of the following:
v SUCCESS
v DRIVE_ERROR

Appendix A. External media management interface description 1095

 v LIBRARY_ERROR
v INTERNAL_ERROR

1096 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Appendix B. User exit and file exit receivers
The data structure of the user exit receivers applies to the file exit receivers. To use
one of these exits with Tivoli Storage Manager, you must specify the corresponding
server option (FILEEXIT, FILETEXTEXIT, or USEREXIT) in the server options file.

The samples for the C, H, and make files are shipped with the server code in the
\win32app\ibm\adsm directory.

You can also use Tivoli Storage Manager commands to control event logging. For
details, see Chapter 27, “Logging IBM Tivoli Storage Manager events to receivers,”
on page 807 and Administrator's Reference.

Sample user-exit declarations

USEREXITSAMPLE.H contains declarations for a user-exit program.

Figure 125. Sample user exit declarations

/***********************************************************************
* Name: USEREXITSAMPLE.H
* Description: Declarations for a user exit
* Environment: WINDOWS NT
***********************************************************************/

#ifndef _H_USEREXITSAMPLE
#define _H_USEREXITSAMPLE

#include <stdio.h>
#include <sys/types.h>

/***** Do not modify below this line *****/

#define BASE_YEAR 1900

typedef short int16;

typedef int int32;

#ifndef uchar
typedef unsigned char uchar;
#endif
/* DateTime Structure Definitions - TSM representation of a timestamp */

typedef struct
{
uchar year; /* Years since BASE_YEAR (0-255) */
uchar mon; /* Month (1 - 12) */
uchar day; /* Day (1 - 31) */
uchar hour; /* Hour (0 - 23) */
uchar min; /* Minutes (0 - 59) */
uchar sec; /* Seconds (0 - 59) */
} DateTime;
/******************************************
* Some field size definitions (in bytes) *
******************************************/

#define MAX_SERVERNAME_LENGTH 64
#define MAX_NODE_LENGTH 64

© Copyright IBM Corp. 1993, 2013 1097

 #define MAX_COMMNAME_LENGTH 16
#define MAX_OWNER_LENGTH 64
#define MAX_HL_ADDRESS 64
#define MAX_LL_ADDRESS 32
#define MAX_SCHED_LENGTH 30
#define MAX_DOMAIN_LENGTH 30
#define MAX_MSGTEXT_LENGTH 1600

/**********************************************
* Event Types (in [Link]) *
**********************************************/

#define TSM_SERVER_EVENT 0x03 /* Server Events */

#define TSM_CLIENT_EVENT 0x05 /* Client Events */

/***************************************************
* Application Types (in [Link]) *
***************************************************/

#define TSM_APPL_BACKARCH 1 /* Backup or Archive client */

#define TSM_APPL_HSM 2 /* Space manage client */
#define TSM_APPL_API 3 /* API client */
#define TSM_APPL_SERVER 4 /* Server (ie. server to server ) */

/*****************************************************
* Event Severity Codes (in [Link]) *
*****************************************************/

#define TSM_SEV_INFO 0x02 /* Informational message. */

#define TSM_SEV_WARNING 0x03 /* Warning message. */
#define TSM_SEV_ERROR 0x04 /* Error message. */
#define TSM_SEV_SEVERE 0x05 /* Severe error message. */
#define TSM_SEV_DIAGNOSTIC 0x06 /* Diagnostic message. */
#define TSM_SEV_TEXT 0x07 /* Text message. */

/************************************************************
* Data Structure of Event that is passed to the User-Exit. *
* The same structure is used for a file receiver *
************************************************************/

typedef struct evRdata

{
int32 eventNum; /* the event number. */
int16 sevCode; /* event severity. */
int16 applType; /* application type (hsm, api, etc) */
int32 sessId; /* session number */
int32 version; /* Version of this structure (1) */
int32 eventType; /* event type */
*(TSM_CLIENT_EVENT, TSM_SERVER_EVENT)*/
DateTime timeStamp; /* timestamp for event data. */
uchar serverName[MAX_SERVERNAME_LENGTH+1]; /* server name */
uchar nodeName[MAX_NODE_LENGTH+1]; /* Node name for session */
uchar commMethod[MAX_COMMNAME_LENGTH+1]; /* communication method */
uchar ownerName[MAX_OWNER_LENGTH+1]; /* owner */
uchar hlAddress[MAX_HL_ADDRESS+1]; /* high-level address */
uchar llAddress[MAX_LL_ADDRESS+1]; /* low-level address */
uchar schedName[MAX_SCHED_LENGTH+1]; /*sched name if applicable */
uchar domainName[MAX_DOMAIN_LENGTH+1];/*omain name for node */
uchar event[MAX_MSGTEXT_LENGTH]; /* event text */
} elEventRecvData;

/************************************
* Size of the Event data structure *
************************************/

#define ELEVENTRECVDATA_SIZE sizeof(elEventRecvData)

1098 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 /*************************************
* User Exit EventNumber for Exiting *
*************************************/

#define USEREXIT_END_EVENTNUM 1822 /* Only user-exit receiver to exit*/

#define END_ALL_RECEIVER_EVENTNUM 1823 /* All receivers told to exit */

/**************************************
*** Do not modify above this line. ***
**************************************/

/********************** Additional Declarations **************************/

#endi

Sample user exit program

USEREXITSAMPLE.C is a sample user-exit program invoked by the Tivoli Storage
Manager server.

Figure 126. Sample user exit program

/***********************************************************************
* Name: USEREXITSAMPLE.C
* Description: Example user-exit program that is invoked by
* the TSM V3 Server
* Environment: *********************************************
* ** This is a platform-specific source file **
* ** versioned for: "WINDOWS NT" **
* *********************************************
***********************************************************************/

#include <stdio.h>
#include <stdlib.h>
#include <sys/types.h>
#include <io.h>
#include <windows.h>
#include "USEREXITSAMPLE.H"

/**************************************
*** Do not modify below this line. ***
**************************************/

#define DllExport __declspec(dllexport)

/****************
*** DLL MAIN ***
****************/

BOOL WINAPI
DllMain(HMODULE hMod, DWORD fdwReason, LPVOID lpvReserved)
{
return(TRUE);
} // End of WINAPI

/******************************************************************
* Procedure: adsmV3UserExit
* If the user-exit is specified on the server, a valid and
* appropriate event will cause an elEventRecvData structure
* (see USEREXITSAMPLE.H) to be passed to a procedure named
* adsmV3UserExit that returns a void.
*
* This procedure can be named differently:
* ----------------------------------------

Appendix B. User exit and file exit receivers 1099

 * The procedure name must match the function name specified in
* the server options file (4th arg). The DLL name generated from
* this module must also match in the server options file
* (3rd arg).
* INPUT : A (void *) to the elEventRecvData structure
* RETURNS: Nothing
******************************************************************/

DllExport void adsmV3UserExit( void *anEvent )

{
/* Typecast the event data passed */
elEventRecvData *eventData = (elEventRecvData *)anEvent;

/**************************************
*** Do not modify above this line. ***
**************************************/

fprintf(stderr,"NT UserExit: Got event

if( ( eventData->eventNum == USEREXIT_END_EVENTNUM ) ||

( eventData->eventNum == END_ALL_RECEIVER_EVENTNUM ) )
{
/* Server says to end this user-exit. Perform any cleanup, *
* but do NOT exit() !!! */
return;
}

/* Field Access: eventData->.... */

/* Your code here ... */

/* Be aware that certain function calls are process-wide and can cause
* synchronization of all threads running under the TSM Server process!
* Among these is the system() function call. Use of this call can
* cause the server process to hang and otherwise affect performance.
* Also avoid any functions that are not thread-safe. Consult your
* system’s programming reference material for more information.
*/

return; /* For picky compilers */

} /* End of adsmV3UserExit() */

Readable text file exit (FILETEXTEXIT) format

If you specify the readable text file exit (FILETEXTEXIT), each logged event is
written to a fixed-size, readable line.

The following table presents the format of the output. Fields are separated by
blank spaces.
Table 83. Readable text file exit (FILETEXTEXIT) format
Column Description
0001-0006 Event number (with leading zeros)
0008-0010 Severity code number
0012-0013 Application type number
0015-0023 Session ID number
0025-0027 Event structure version number
0029-0031 Event type number
0033-0046 Date/Time (YYYYMMDDDHHmmSS)

1100 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Table 83. Readable text file exit (FILETEXTEXIT) format (continued)
Column Description
0048-0111 Server name (right padded with spaces)
1
0113-0176 Node name
1
0178-0193 Communications method name
1
0195-0258 Owner name
1
0260-0323 High-level internet address (n.n.n.n)
1
0325-0356 Port number from high-level internet address
1
0358-0387 Schedule name
1
0389-0418 Domain name
0420-2019 Event text
2020-2499 Unused spaces
2500 New line character
1
Columns 113 - 418 contain data only for events that originate in a client or in another
Tivoli Storage Manager server. Otherwise, columns 113 - 418 contain blanks.

Appendix B. User exit and file exit receivers 1101

1102 IBM Tivoli Storage Manager for Windows: Administrator's Guide
Appendix C. Tivoli support information
You can find support information for Tivoli and other IBM products from various
sources.

From the IBM Support Portal at [Link] you

can select the products that you are interested in and search for a wide variety of
relevant information.

Communities and other learning resources

In addition to product documentation, many forms of assistance are available to
help you get started as you deploy and use the Tivoli Storage Manager family of
products. These resources can also help you to solve problems that you might
have.

You can use forums, wikis, and other social media tools to ask questions, talk to
experts, and learn from others.

User groups
Tivoli Global Storage Virtual User Group
Access this user group at [Link]
This group makes it possible for individuals from many different industries
and types of organizations to share information and work directly with the
IBM product experts. Local chapters also exist where members meet in
person to share experiences and hear from guest speakers.
[Link]
Access this mailing list at [Link]
This independently managed Storage Management discussion forum
started when Tivoli Storage Manager was known as ADSTAR Distributed
Storage Manager (ADSM). The members of this forum have many years of
experience with Tivoli Storage Manager in almost every type of IT
environment.
To subscribe to the forum, send an email to listserv@[Link]. The
body of the message must contain the following text: SUBSCRIBE ADSM-L
your_first_name your_family_name.

Tivoli Storage Manager community on Service Management

Connect

Access Service Management Connect at [Link]

servicemanagement. In the Storage Management community of Service
Management Connect, you can connect with IBM in the following ways:
v Become involved with transparent development, an ongoing, open engagement
between users and IBM developers of Tivoli products. You can access early
designs, sprint demonstrations, product roadmaps, and prerelease code.
v Connect one-on-one with the experts to collaborate and network about Tivoli
and the Tivoli Storage Manager community.
v Read blogs to benefit from the expertise and experience of others.

© Copyright IBM Corp. 1993, 2013 1103

 v Use wikis and forums to collaborate with the broader user community.

Tivoli Storage Manager wiki on developerWorks®

Access this wiki at [Link]

sm/[Link].

Find the latest best practices, white papers, and links to videos and other
resources. When you log on, you can comment on content, or contribute your own
content.

Tivoli Support Technical Exchange

Find information about upcoming Tivoli Support Technical Exchange webcasts at

[Link]
Replays of previous webcasts are also available.

Learn from technical experts who share their knowledge and then answer your
questions. The sessions are designed to address specific technical issues and
provide in-depth but narrowly focused training.

Other social media sites

LinkedIn
You can join groups on LinkedIn, a social media site for professionals. For
example:
v Tivoli Storage Manager Professionals: [Link]
groups/Tivoli-Storage-Manager-Professionals-54572
v TSM: [Link]
Twitter
Follow @IBMStorage on Twitter to see the latest news about storage and
storage software from IBM.

Tivoli education resources

Use these education resources to help you increase your Tivoli Storage Manager
skills:
Tivoli Education and Certification website
View available education at [Link]
education.
Use the Search for Training link to find local and online offerings of
instructor-led courses for Tivoli Storage Manager.
Education Assistant
Access resources at [Link]
tivv1r0/[Link].
Scroll to view the list of available training videos. Recorded product
demonstrations are also available on a YouTube channel.

1104 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Searching knowledge bases
If a problem occurs while you are using one of the Tivoli Storage Manager family
of products, you can search several knowledge bases.

Begin by searching the Tivoli Storage Manager Information Center at

[Link] Within the information center,
you can enter words, phrases, or message numbers in the Search field to find
relevant topics.

Searching the Internet

If you cannot find an answer to your question in the Tivoli Storage Manager
information center, search the Internet for the information that might help you
resolve the problem.

To search multiple Internet resources, go to the IBM support website at

[Link] You can search for information
without signing in.

Sign in using your IBM ID and password if you want to customize the site based
on your product usage and information needs. If you do not already have an IBM
ID and password, click Sign in at the top of the page and follow the instructions
to register.

From the support website, you can search various resources:

v IBM technotes.
v IBM downloads.
v IBM Redbooks® publications.
v IBM Authorized Program Analysis Reports (APARs). Select the product and click
Downloads to search the APAR list.

Using IBM Support Assistant

IBM Support Assistant is a complimentary software product that can help you with
problem determination. It is available for some Tivoli Storage Manager and Tivoli
Storage FlashCopy Manager products.

IBM Support Assistant helps you gather support information when you must open
a problem management record (PMR), which you can then use to track the
problem. The product-specific plug-in modules provide you with the following
resources:
v Support links
v Education links
v Ability to submit problem management reports

You can find more information and download the IBM Support Assistant web page
at [Link]

You can also install the stand-alone IBM Support Assistant application on any
workstation. You can then enhance the application by installing product-specific
plug-in modules for the IBM products that you use. Find add-ons for specific
products at [Link]

Appendix C. Tivoli support information 1105

 Finding product fixes
A product fix to resolve a software problem might be available from the IBM
software support website.

Procedure

Determine what fixes are available by checking the IBM software support website
at [Link]
If you previously customized the site based on your product usage:
1. Click the link for the product, or a component for which you want to
find a fix.
2. Click Downloads, and then click Search for recommended fixes.
If you have not previously customized the site:
Click Downloads and search for the product.

Receiving notification of product fixes

You can receive notifications about fixes, flashes, upgrades, and other news about
IBM products.

Procedure
1. From the support page at [Link] click
Sign in and sign in using your IBM ID and password. If you do not have an
ID and password, click register now and complete the registration process.
2. Click Manage all my subscriptions in the Notifications pane.
3. Click the Subscribe tab, and then click Tivoli.
4. Select the products for which you want to receive notifications and click
Continue.
5. Specify your notification preferences and click Submit.

Contacting IBM Software Support

You can contact IBM Software Support if you have an active IBM subscription and
support contract, and if you are authorized to submit problems to IBM.

Procedure
1. Ensure that you have completed the following prerequisites:
a. Set up a subscription and support contract.
b. Determine the business impact of the problem.
c. Describe the problem and gather background information.
2. Follow the instructions in “Submitting the problem to IBM Software Support”
on page 1108.

1106 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Setting up and managing support contracts
You can set up and manage your Tivoli support contracts by enrolling in IBM
Passport Advantage. The type of support contract that you need depends on the
type of product you have.

Procedure

Enroll in IBM Passport Advantage in one of the following ways:

v Online: Go to the Passport Advantage website at [Link]
software/lotus/passportadvantage/, click How to enroll, and follow the
instructions.
v By telephone: For critical, system-down, or high-severity issues, you can call
1-800-IBMSERV (1-800-426-7378) in the United States. For the telephone number
to call in your country, go to the IBM Software Support Handbook web page at
[Link]
and click Contacts.

Determining the business impact

When you report a problem to IBM, you are asked to supply a severity level.
Therefore, you must understand and assess the business impact of the problem
you are reporting.

Severity level Description

Severity 1 Critical business impact: You are unable to use the program,
resulting in a critical impact on operations. This condition
requires an immediate solution.
Severity 2 Significant business impact: The program is usable but is
severely limited.
Severity 3 Some business impact: The program is usable with less
significant features (not critical to operations) unavailable.
Severity 4 Minimal business impact: The problem causes little impact on
operations, or a reasonable circumvention to the problem has
been implemented.

Describing the problem and gathering background information

When explaining a problem to IBM, it is helpful to be as specific as possible.
Include all relevant background information so that IBM Software Support
specialists can help you solve the problem efficiently.

To save time, know the answers to these questions:

v What software versions were you running when the problem occurred?
v Do you have logs, traces, and messages that are related to the problem
symptoms? IBM Software Support is likely to ask for this information.
v Can the problem be re-created? If so, what steps led to the failure?
v Have any changes been made to the system? For example, hardware, operating
system, networking software, and so on.
v Are you using a workaround for this problem? If so, be prepared to explain it
when you report the problem.

Appendix C. Tivoli support information 1107

 Submitting the problem to IBM Software Support
You can submit the problem to IBM Software Support online or by telephone.
Online
Go to the IBM Software Support website at [Link]
support/entry/portal/Open_service_request/Software/
Software_support_(general). Sign in to access IBM Service Requests and
enter your information into the problem submission tool.
By telephone
For critical, system-down, or severity 1 issues, you can call 1-800-IBMSERV
(1-800-426-7378) in the United States. For the telephone number to call in
your country, go to the IBM Software Support Handbook web page at
[Link]
[Link] and click Contacts.

1108 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Appendix D. Accessibility features for the Tivoli Storage
Manager product family
Accessibility features help users who have a disability, such as restricted mobility
or limited vision to use information technology products successfully.

Accessibility features

The IBM Tivoli Storage Manager family of products includes the following
accessibility features:
v Keyboard-only operation using standard operating-system conventions
v Interfaces that support assistive technology such as screen readers

The command-line interfaces of all products in the product family are accessible.

Tivoli Storage Manager Operations Center provides the following additional

accessibility features when you use it with a Mozilla Firefox browser on a
Microsoft Windows system:
v Screen magnifiers and content zooming
v High contrast mode

The Operations Center and the Tivoli Storage Manager Server can be installed in
console mode, which is accessible.

The Tivoli Storage Manager Information Center is enabled for accessibility. For
information center accessibility information, see “Accessibility features in the
information center” ( [Link]
[Link]/iehs36_accessibility.html).

Vendor software

The Tivoli Storage Manager product family includes certain vendor software that is
not covered under the IBM license agreement. IBM makes no representation about
the accessibility features of these products. Contact the vendor for the accessibility
information about its products.

IBM and accessibility

See the IBM Human Ability and Accessibility Center ([Link]

for information about the commitment that IBM has to accessibility.

© Copyright IBM Corp. 1993, 2013 1109

1110 IBM Tivoli Storage Manager for Windows: Administrator's Guide
Notices
This information was developed for products and services offered in the U.S.A.

IBM may not offer the products, services, or features discussed in this document in
other countries. Consult your local IBM representative for information on the
products and services currently available in your area. Any reference to an IBM
product, program, or service is not intended to state or imply that only that IBM
product, program, or service may be used. Any functionally equivalent product,
program, or service that does not infringe any IBM intellectual property right may
be used instead. However, it is the user's responsibility to evaluate and verify the
operation of any non-IBM product, program, or service.

IBM may have patents or pending patent applications covering subject matter
described in this document. The furnishing of this document does not grant you
any license to these patents. You can send license inquiries, in writing, to:

IBM Director of Licensing

IBM Corporation
North Castle Drive
Armonk, NY 10504-1785
U.S.A.

For license inquiries regarding double-byte character set (DBCS) information,

contact the IBM Intellectual Property Department in your country or send
inquiries, in writing, to:

Intellectual Property Licensing

Legal and Intellectual Property Law
IBM Japan, Ltd.
19-21, Nihonbashi-Hakozakicho, Chuo-ku
Tokyo 103-8510, Japan

The following paragraph does not apply to the United Kingdom or any other
country where such provisions are inconsistent with local law:
INTERNATIONAL BUSINESS MACHINES CORPORATION PROVIDES THIS
PUBLICATION “AS IS” WITHOUT WARRANTY OF ANY KIND, EITHER
EXPRESS OR IMPLIED, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
WARRANTIES OF NON-INFRINGEMENT, MERCHANTABILITY OR FITNESS
FOR A PARTICULAR PURPOSE. Some states do not allow disclaimer of express or
implied warranties in certain transactions, therefore, this statement may not apply
to you.

This information could include technical inaccuracies or typographical errors.

Changes are periodically made to the information herein; these changes will be
incorporated in new editions of the publication. IBM may make improvements
and/or changes in the product(s) and/or the program(s) described in this
publication at any time without notice.

Any references in this information to non-IBM websites are provided for

convenience only and do not in any manner serve as an endorsement of those
websites. The materials at those websites are not part of the materials for this IBM
product and use of those websites is at your own risk.

© Copyright IBM Corp. 1993, 2013 1111

 IBM may use or distribute any of the information you supply in any way it
believes appropriate without incurring any obligation to you.

Licensees of this program who want to have information about it for the purpose
of enabling: (i) the exchange of information between independently created
programs and other programs (including this one) and (ii) the mutual use of the
information which has been exchanged, should contact:

IBM Corporation
2Z4A/101
11400 Burnet Road
Austin, TX 78758
U.S.A.

Such information may be available, subject to appropriate terms and conditions,

including in some cases, payment of a fee.

The licensed program described in this information and all licensed material
available for it are provided by IBM under terms of the IBM Customer Agreement,
IBM International Program License Agreement, or any equivalent agreement
between us.

Any performance data contained herein was determined in a controlled

environment. Therefore, the results obtained in other operating environments may
vary significantly. Some measurements may have been made on development-level
systems and there is no guarantee that these measurements will be the same on
generally available systems. Furthermore, some measurements may have been
estimated through extrapolation. Actual results may vary. Users of this document
should verify the applicable data for their specific environment.

Information concerning non-IBM products was obtained from the suppliers of

those products, their published announcements or other publicly available sources.
IBM has not tested those products and cannot confirm the accuracy of
performance, compatibility or any other claims related to non-IBM products.
Questions on the capabilities of non-IBM products should be addressed to the
suppliers of those products.

This information is for planning purposes only. The information herein is subject to
change before the products described become available.

This information contains examples of data and reports used in daily business
operations. To illustrate them as completely as possible, the examples include the
names of individuals, companies, brands, and products. All of these names are
fictitious and any similarity to the names and addresses used by an actual business
enterprise is entirely coincidental.

COPYRIGHT LICENSE:

This information contains sample application programs in source language, which

illustrate programming techniques on various operating platforms. You may copy,
modify, and distribute these sample programs in any form without payment to
IBM, for the purposes of developing, using, marketing or distributing application
programs conforming to the application programming interface for the operating
platform for which the sample programs are written. These examples have not
been thoroughly tested under all conditions. IBM, therefore, cannot guarantee or
imply reliability, serviceability, or function of these programs. The sample

1112 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 programs are provided "AS IS", without warranty of any kind. IBM shall not be
liable for any damages arising out of your use of the sample programs.

Each copy or any portion of these sample programs or any derivative work, must
include a copyright notice as follows: © (your company name) (year). Portions of
this code are derived from IBM Corp. Sample Programs. © Copyright IBM Corp.
_enter the year or years_.

If you are viewing this information in softcopy, the photographs and color
illustrations may not appear.

Trademarks
IBM, the IBM logo, and [Link]® are trademarks or registered trademarks of
International Business Machines Corp., registered in many jurisdictions worldwide.
Other product and service names might be trademarks of IBM or other companies.
A current list of IBM trademarks is available on the Web at “Copyright and
trademark information” at [Link]

Adobe and PostScript are either registered trademarks or trademarks of Adobe

Systems Incorporated in the United States, and/or other countries.

Java and all Java-based trademarks and logos are trademarks or registered
trademarks of Oracle and/or its affiliates.

LTO and Ultrium are trademarks of HP, IBM Corp. and Quantum in the U.S. and
other countries.

Linux is a registered trademark of Linus Torvalds in the United States, other

countries, or both.

Microsoft, Windows, Windows NT, and the Windows logo are trademarks of
Microsoft Corporation in the United States, other countries, or both.

UNIX is a registered trademark of The Open Group in the United States and other
countries.

Privacy policy considerations

IBM Software products, including software as a service solutions, (“Software
Offerings”) may use cookies or other technologies to collect product usage
information, to help improve the end user experience, to tailor interactions with
the end user or for other purposes. In many cases no personally identifiable
information is collected by the Software Offerings. Some of our Software Offerings
can help enable you to collect personally identifiable information. If this Software
Offering uses cookies to collect personally identifiable information, specific
information about this offering’s use of cookies is set forth below.

This Software Offering does not use cookies or other technologies to collect
personally identifiable information.

Notices 1113
 If the configurations deployed for this Software Offering provide you as customer
the ability to collect personally identifiable information from end users via cookies
and other technologies, you should seek your own legal advice about any laws
applicable to such data collection, including any requirements for notice and
consent.

For more information about the use of various technologies, including cookies, for
these purposes, see IBM’s Privacy Policy at [Link] and
IBM’s Online Privacy Statement at [Link] the
section entitled “Cookies, Web Beacons and Other Technologies” and the “IBM
Software Products and Software-as-a-Service Privacy Statement” at
[Link]

1114 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Glossary
This glossary provides terms and definitions for backup data. See also server storage,
Tivoli Storage Manager, Tivoli Storage FlashCopy storage pool, storage pool volume.
Manager, and associated products.
active file system
A file system to which space management
The following cross-references are used in this
has been added. With space management,
glossary:
tasks for an active file system include
v See refers you from a nonpreferred term to the automatic migration, reconciliation,
preferred term or from an abbreviation to the selective migration, and recall. See also
spelled-out form. inactive file system.
v See also refers you to a related or contrasting
active policy set
term.
The activated policy set that contains the
policy rules currently in use by all client
For other terms and definitions, see the IBM
nodes assigned to the policy domain. See
Terminology website at [Link]/software/
also policy domain, policy set.
globalization/terminology.
active version
The most recent backup copy of a file
A stored. The active version of a file cannot
absolute mode be deleted until a backup process detects
In storage management, a backup that the user has either replaced the file
copy-group mode that specifies that a file with a newer version or has deleted the
is considered for incremental backup even file from the file server or workstation.
if the file has not changed since the last See also backup version, inactive version.
backup. See also mode, modified mode. activity log
access control list (ACL) A log that records normal activity
In computer security, a list associated messages that are generated by the server.
with an object that identifies all the These messages include information about
subjects that can access the object and server and client operations, such as the
their access rights. start time of sessions or device I/O errors.

access mode adaptive subfile backup

An attribute of a storage pool or a storage A type of backup that sends only changed
volume that specifies whether the server portions of a file to the server, instead of
can write to or read from the storage pool sending the entire file. Adaptive subfile
or storage volume. backup reduces network traffic and
increases the speed of the backup.
ACK See acknowledgment.
administrative client
acknowledgment (ACK) A program that runs on a file server,
The transmission of acknowledgment workstation, or mainframe that
characters as a positive response to a data administrators use to control and monitor
transmission. the server. See also backup-archive client.
ACL See access control list. administrative command schedule
activate A database record that describes the
To validate the contents of a policy set planned processing of an administrative
and then make it the active policy set. command during a specific time period.
See also central scheduler, client schedule,
active-data pool schedule.
A named set of storage pool volumes that
contain only active versions of client administrative privilege class
See privilege class.

© Copyright IBM Corp. 1993, 2013 1115

administrative session node and a client schedule. An association
A period of time during which an identifies the name of a schedule, the
administrator user ID communicates with name of the policy domain to which the
a server to perform administrative tasks. schedule belongs, and the name of a
See also client node session, session. client node that performs scheduled
operations.
administrator
A person responsible for administrative audit To check for logical inconsistencies
tasks such as access authorization and between information that the server has
content management. Administrators can and the actual condition of the system.
also grant levels of authority to users. The storage manager can audit
information about items such as volumes,
agent node
libraries, and licenses. For example, when
A client node that has been granted proxy
a storage manager audits a volume, the
authority to perform operations on behalf
server checks for inconsistencies between
of another client node, which is the target
information about backed-up or archived
node.
files that are stored in the database and
aggregate the actual data that are associated with
An object, stored in one or more storage each backup version or archive copy in
pools, consisting of a group of logical files server storage.
that are packaged together. See also
authentication rule
logical file, physical file.
A specification that another user can use
aggregate data transfer rate to either restore or retrieve files from
A performance statistic that indicates the storage.
average number of bytes that were
authority
transferred per second while processing a
The right to access objects, resources, or
given operation.
functions. See also privilege class.
application client
authorization rule
A program that is installed on a system to
A specification that permits another user
protect an application. The server
to either restore or retrieve a user's files
provides backup services to an
from storage.
application client.
authorized user
archive
A user who has administrative authority
To copy programs, data, or files to
for the client on a workstation. This user
another storage media, usually for
changes passwords, performs open
long-term storage or security. See also
registrations, and deletes file spaces.
retrieve.
AutoFS
archive copy
See automounted file system.
A file or group of files that was archived
to server storage automatic detection
A feature that detects, reports, and
archive copy group
updates the serial number of a drive or
A policy object containing attributes that
library in the database when the path
control the generation, destination, and
from the local server is defined.
expiration of archived files. See also copy
group. automatic migration
The process that is used to automatically
archive-retention grace period
move files from a local file system to
The number of days that the storage
storage, based on options and settings
manager retains an archived file when the
that are chosen by a root user on a
server is unable to rebind the file to an
workstation. See also demand migration,
appropriate management class. See also
threshold migration.
bind.
automounted file system (AutoFS)
association
A file system that is managed by an
The defined relationship between a client

1116 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 automounter daemon. The automounter
daemon monitors a specified directory
path, and automatically mounts the file
system to access data.

B
backup-archive client
A program that runs on a workstation or
file server and provides a means for users
to back up, archive, restore, and retrieve
files. See also administrative client.
backup copy group
A policy object containing attributes that
control the generation, destination, and
expiration of backup versions of files. A
backup copy group belongs to a
management class. See also copy group.
backup retention grace period
The number of days the storage manager
retains a backup version after the server
is unable to rebind the file to an
appropriate management class.
backup set
A portable, consolidated group of active
versions of backup files that are generated
for a backup-archive client.
backup set collection
A group of backup sets that are created at
the same time and which have the same
backup set name, volume names,
description, and device classes. The server
identifies each backup set in the collection
by its node name, backup set name, and
file type.
backup version
A file or directory that a client node
backed up to storage. More than one
backup version can exist in storage, but
only one backup version is the active
version. See also active version, copy
group, inactive version.
bind To associate a file with a management
class name. See also archive-retention
grace period, management class, rebind.

Glossary 1117
 the server and used on client nodes in
C conjunction with client options files.
cache To place a duplicate copy of a file on client options file
random access media when the server An editable file that identifies the server
migrates a file to another storage pool in and communication method, and
the hierarchy. provides the configuration for backup,
cache file archive, hierarchical storage management,
A snapshot of a logical volume created by and scheduling.
Logical Volume Snapshot Agent. Blocks client-polling scheduling mode
are saved immediately before they are A method of operation in which the client
modified during the image backup and queries the server for work. See also
their logical extents are saved in the cache server-prompted scheduling mode.
files.
client schedule
CAD See client acceptor daemon. A database record that describes the
central scheduler planned processing of a client operation
A function that permits an administrator during a specific time period. The client
to schedule client operations and operation can be a backup, archive,
administrative commands. The operations restore, or retrieve operation, a client
can be scheduled to occur periodically or operating system command, or a macro.
on a specific date. See also administrative See also administrative command
command schedule, client schedule. schedule, central scheduler, schedule.
client A software program or computer that client/server
requests services from a server. See also Pertaining to the model of interaction in
server. distributed data processing in which a
program on one computer sends a request
client acceptor to a program on another computer and
A service that serves the Java applet for awaits a response. The requesting
the web client to web browsers. On program is called a client; the answering
Windows systems, the client acceptor is program is called a server.
installed and run as a service. On AIX,
UNIX, and Linux systems, the client client system-options file
acceptor is run as a daemon. A file, used on AIX, UNIX, or Linux
system clients, containing a set of
client acceptor daemon (CAD) processing options that identify the
See client acceptor. servers to be contacted for services. This
client domain file also specifies communication methods
The set of drives, file systems, or volumes and options for backup, archive,
that the user selects to back up or archive hierarchical storage management, and
data, using the backup-archive client. scheduling. See also client user-options
file, options file.
client node
A file server or workstation on which the client user-options file
backup-archive client program has been A file that contains the set of processing
installed, and which has been registered options that the clients on the system use.
to the server. The set can include options that
determine the server that the client
client node session contacts, and options that affect backup
A session in which a client node operations, archive operations,
communicates with a server to perform hierarchical storage management
backup, restore, archive, retrieve, migrate, operations, and scheduled operations.
or recall requests. See also administrative This file is also called the [Link] file.
session. For AIX, UNIX, or Linux systems, see also
client option set client system-options file. See also client
A group of options that are defined on system-options file, options file.

1118 IBM Tivoli Storage Manager for Windows: Administrator's Guide

closed registration copy backup
A registration process in which only an A full backup in which the transaction log
administrator can register workstations as files are not deleted so that backup
client nodes with the server. See also open procedures that use incremental or
registration. differential backups are not disrupted.
collocation copy group
The process of keeping all data belonging A policy object containing attributes that
to a single-client file space, a single client control how backup versions or archive
node, or a group of client nodes on a copies are generated, where backup
minimal number of sequential-access versions or archive copies are initially
volumes within a storage pool. located, and when backup versions or
Collocation can reduce the number of archive copies expire. A copy group
volumes that must be accessed when a belongs to a management class. See also
large amount of data must be restored. archive copy group, backup copy group,
backup version,management class.
collocation group
A user-defined group of client nodes copy storage pool
whose data is stored on a minimal A named set of volumes that contain
number of volumes through the process copies of files that reside in primary
of collocation. storage pools. Copy storage pools are
used only to back up the data that is
commit point
stored in primary storage pools. A copy
A point in time when data is considered
storage pool cannot be a destination for a
to be consistent.
backup copy group, an archive copy
communication method group, or a management class (for
The method by which a client and server space-managed files). See also destination,
exchange information. See also primary storage pool, server storage,
Transmission Control Protocol/Internet storage pool, storage pool volume.
Protocol.
communication protocol D
A set of defined interfaces that permit
computers to communicate with each daemon
other. A program that runs unattended to
perform continuous or periodic functions,
compression such as network control.
A function that removes repetitive
characters, spaces, strings of characters, or damaged file
binary data from the data being processed A physical file in which read errors have
and replaces characters with control been detected.
characters. Compression reduces the database backup series
amount of storage space that is required One full backup of the database, plus up
for data. to 32 incremental backups made since
configuration manager that full backup. Each full backup that is
A server that distributes configuration run starts a new database backup series.
information, such as policies and A number identifies each backup series.
schedules, to managed servers according See also database snapshot, full backup.
to their profiles. Configuration database snapshot
information can include policy and A complete backup of the entire database
schedules. See also enterprise to media that can be taken off-site. When
configuration, managed server, profile. a database snapshot is created, the current
conversation database backup series is not interrupted.
A connection between two programs over A database snapshot cannot have
a session that allows them to incremental database backups associated
communicate with each other while with it. See also database backup series,
processing a transaction. See also session. full backup.

Glossary 1119
data center out-of-space condition on a file system for
In a virtualized environment, a container which hierarchical storage management
that holds hosts, clusters, networks, and (HSM) is active. Files are migrated to
data stores. server storage until space usage drops to
the low threshold that was set for the file
data deduplication
system. If the high threshold and low
A method of reducing storage needs by
threshold are the same, one file is
eliminating redundant data. Only one
migrated. See also automatic migration,
instance of the data is retained on storage
selective migration, threshold migration.
media. Other instances of the same data
are replaced with a pointer to the retained desktop client
instance. The group of backup-archive clients that
includes clients on Microsoft Windows,
data manager server
Apple, and Novell NetWare operating
A server that collects metadata
systems.
information for client inventory and
manages transactions for the storage destination
agent over the local area network. The A copy group or management class
data manager server informs the storage attribute that specifies the primary storage
agent with applicable library attributes pool to which a client file will be backed
and the target volume identifier. up, archived, or migrated. See also copy
storage pool.
data mover
A device that moves data on behalf of the device class
server. A network-attached storage (NAS) A named set of characteristics that are
file server is a data mover. applied to a group of storage devices.
Each device class has a unique name and
data storage-management application-
represents a device type of disk, file,
programming interface (DSMAPI)
optical disk, or tape.
A set of functions and semantics that can
monitor events on files, and manage and device configuration file
maintain the data in a file. In an HSM 1. For a storage agent, a file that contains
environment, a DSMAPI uses events to the name and password of the storage
notify data management applications agent, and information about the
about operations on files, stores arbitrary server that is managing the
attribute information with a file, supports SAN-attached libraries and drives that
managed regions in a file, and uses the storage agent uses.
DSMAPI access rights to control access to
2. For a server, a file that contains
a file object.
information about defined device
data store classes, and, on some servers, defined
In a virtualized environment, the location libraries and drives. The information
where virtual machine data is stored. is a copy of the device configuration
information in the database.
deduplication
The process of creating representative disaster recovery manager (DRM)
records from a set of records that have A function that assists in preparing and
been identified as representing the same using a disaster recovery plan file for the
entities. server.
default management class disaster recovery plan
A management class that is assigned to a A file that is created by the disaster
policy set. This class is used to govern recover manager (DRM) that contains
backed up or archived files when a file is information about how to recover
not explicitly associated with a specific computer systems if a disaster occurs and
management class through the scripts that can be run to perform some
include-exclude list. recovery tasks. The file includes
information about the software and
demand migration
The process that is used to respond to an

1120 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 hardware that is used by the server, and failure of an operation, a user action, or
the location of recovery media. the change in state of a process. See also
enterprise logging, receiver.
domain
A grouping of client nodes with one or event record
more policy sets, which manage data or A database record that describes actual
storage resources for the client nodes. See status and results for events.
also policy domain.
event server
DRM See disaster recovery manager. A server to which other servers can send
events for logging. The event server
DSMAPI
routes the events to any receivers that are
See data storage-management
enabled for the sending server's events.
application-programming interface.
exabyte (EB)
dynamic serialization
For processor, real and virtual storage
Copy serialization in which a file or
capacities and channel volume, 2 to the
folder is backed up or archived on the
power of 60 or 1 152 921 504 606 846 976
first attempt regardless of whether it
bytes. For disk storage capacity and
changes during a backup or archive. See
communications volume, 1 000 000 000
also shared dynamic serialization, shared
000 000 000 bytes.
static serialization, static serialization.
exclude
The process of identifying files in an
E include-exclude list. This process prevents
EA See extended attribute. the files from being backed up or
migrated whenever a user or schedule
EB See exabyte. enters an incremental or selective backup
EFS See Encrypted File System. operation. A file can be excluded from
backup, from space management, or from
Encrypted File System (EFS) both backup and space management.
A file system that uses file system-level
encryption. exclude-include list
See include-exclude list.
enterprise configuration
A method of setting up servers so that the expiration
administrator can distribute the The process by which files, data sets, or
configuration of one of the servers to the objects are identified for deletion because
other servers, using server-to-server their expiration date or retention period
communication. See also configuration has passed.
manager, managed server, profile, expiring file
subscription. A migrated or premigrated file that has
enterprise logging been marked for expiration and removal
The process of sending events from a from storage. If a stub file or an original
server to a designated event server. The copy of a premigrated file is deleted from
event server routes the events to a local file system, or if the original copy
designated receivers, such as to a user of a premigrated file is updated, the
exit. See also event. corresponding migrated or premigrated
file is marked for expiration the next time
error log reconciliation is run.
A data set or file that is used to record
error information about a product or extend
system. To increase the portion of available space
that can be used to store database or
estimated capacity recovery log information.
The available space, in megabytes, of a
storage pool. extended attribute (EA)
Names or value pairs that are associated
event An occurrence of significance to a task or with files or directories. There are three
system. Events can include completion or

Glossary 1121
 classes of extended attributes: user management support that is required. If
attributes, system attributes, and trusted no space management support is
attributes. required, the operation is passed to the
operating system, which performs its
external library
normal functions. The file system
A collection of drives that is managed by
migrator is mounted over a file system
the media-management system other than
when space management is added to the
the storage management server.
file system.
file system state
F The storage management mode of a file
file access time system that resides on a workstation on
On AIX, UNIX, or Linux systems, the which the hierarchical storage
time when the file was last accessed. management (HSM) client is installed. A
file system can be in one of these states:
file age native, active, inactive, or global inactive.
For migration prioritization purposes, the
number of days since a file was last frequency
accessed. A copy group attribute that specifies the
minimum interval, in days, between
file device type incremental backups.
A device type that specifies the use of
sequential access files on disk storage as FSID See file space ID.
volumes. FSM See file system migrator.
file server full backup
A dedicated computer and its peripheral The process of backing up the entire
storage devices that are connected to a server database. A full backup begins a
local area network that stores programs new database backup series. See also
and files that are shared by users on the database backup series, database
network. snapshot, incremental backup.
file space fuzzy backup
A logical space in server storage that A backup version of a file that might not
contains a group of files that have been accurately reflect what is currently in the
backed up or archived by a client node, file because the file was backed up at the
from a single logical partition, file system, same time as it was being modified.
or virtual mount point. Client nodes can
restore, retrieve, or delete their file spaces fuzzy copy
from server storage. In server storage, A backup version or archive copy of a file
files belonging to a single file space are that might not accurately reflect the
not necessarily stored together. original contents of the file because it was
backed up or archived the file while the
file space ID (FSID) file was being modified.
A unique numeric identifier that the
server assigns to a file space when it is
stored in server storage. G
file state GB See gigabyte.
The space management mode of a file
General Parallel File System (GPFS)
that resides in a file system to which
A high-performance shared-disk file
space management has been added. A file
system that can provide data access from
can be in one of three states: resident,
nodes in a clustered system environment.
premigrated, or migrated. See also
See also information lifecycle
migrated file, premigrated file, resident
management.
file.
gigabyte (GB)
file system migrator (FSM)
For processor storage, real and virtual
A kernel extension that intercepts all file
storage, and channel volume, 10 to the
system operations and provides any space

1122 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 power of nine or 1,073,741,824 bytes. For
disk storage capacity and communications
volume, 1,000,000,000 bytes.
global inactive state
The state of all file systems to which
space management has been added when
space management is globally deactivated
for a client node.
Globally Unique Identifier (GUID)
An algorithmically determined number
that uniquely identifies an entity within a
system. See also Universally Unique
Identifier.
GPFS See General Parallel File System.
GPFS node set
A mounted, defined group of GPFS file
systems.
group backup
The backup of a group containing a list of
files from one or more file space origins.
GUID See Globally Unique Identifier.

H
hierarchical storage management (HSM)
A function that automatically distributes
and manages data on disk, tape, or both
by regarding devices of these types and
potentially others as levels in a storage
hierarchy that range from fast, expensive
devices to slower, cheaper, and possibly
removable devices. The objectives are to
minimize access time to data and
maximize available media capacity. See
also hierarchical storage management
client, recall, storage hierarchy.
hierarchical storage management client (HSM
client) A client program that works with the
server to provide hierarchical storage
management (HSM) for a system. See also
hierarchical storage management,
management class.
HSM See hierarchical storage management.
HSM client
See hierarchical storage management
client.

Glossary 1123
 systems. An inode contains the node,
I type, owner, and location of a file.
ILM See information lifecycle management. inode number
image A file system or raw logical volume that A number specifying a particular inode
is backed up as a single object. file in the file system.
image backup IP address
A backup of a full file system or raw A unique address for a device or logical
logical volume as a single object. unit on a network that uses the Internet
Protocol standard.
inactive file system
A file system for which space
management has been deactivated. See J
also active file system.
job file
inactive version A generated file that contains
A backup version of a file that is either configuration information for a migration
not the most recent backup version, or job. The file is XML format and can be
that is a backup version of a file that no created and edited in the hierarchical
longer exists on the client system. Inactive storage management (HSM) client for
backup versions are eligible for expiration Windows client graphical user interface.
processing according to the management See also migration job.
class assigned to the file. See also active
journal-based backup
version, backup version.
A method for backing up Windows clients
include-exclude file and AIX clients that exploits the change
A file containing statements to determine notification mechanism in a file to
the files to back up and the associated improve incremental backup performance
management classes to use for backup or by reducing the need to fully scan the file
archive. See also include-exclude list. system.
include-exclude list journal daemon
A list of options that include or exclude On AIX, UNIX, or Linux systems, a
selected files for backup. An exclude program that tracks change activity for
option identifies files that should not be files residing in file systems.
backed up. An include option identifies
journal service
files that are exempt from the exclusion
In Microsoft Windows, a program that
rules or assigns a management class to a
tracks change activity for files residing in
file or a group of files for backup or
file systems.
archive services. See also include-exclude
file.
incremental backup
K
The process of backing up files or KB See kilobyte.
directories, or copying pages in the
database, that are new or changed since kilobyte (KB)
the last full or incremental backup. See For processor storage, real and virtual
also selective backup. storage, and channel volume, 2 to the
power of 10 or 1,024 bytes. For disk
individual mailbox restore storage capacity and communications
See mailbox restore. volume, 1,000 bytes.
information lifecycle management (ILM)
A policy-based file-management system
for storage pools and file sets. See also
General Parallel File System.
inode The internal structure that describes the
individual files on AIX, UNIX, or Linux

1124 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 local shadow volume
L Data that is stored on shadow volumes
LAN See local area network. localized to a disk storage subsystem.
LAN-free data movement LOFS See loopback virtual file system.
The movement of client data between a logical file
client system and a storage device on a A file that is stored in one or more server
storage area network (SAN), bypassing storage pools, either by itself or as part of
the local area network. an aggregate. See also aggregate, physical
LAN-free data transfer file, physical occupancy.
See LAN-free data movement. logical occupancy
leader data The space that is used by logical files in a
Bytes of data, from the beginning of a storage pool. This space does not include
migrated file, that are stored in the file's the unused space created when logical
corresponding stub file on the local file files are deleted from aggregate files, so it
system. The amount of leader data that is might be less than the physical
stored in a stub file depends on the stub occupancy. See also physical occupancy.
size that is specified. logical unit number (LUN)
library In the Small Computer System Interface
1. A repository for demountable (SCSI) standard, a unique identifier used
recorded media, such as magnetic to differentiate devices, each of which is a
disks and magnetic tapes. logical unit (LU).
2. A collection of one or more drives, logical volume
and possibly robotic devices A portion of a physical volume that
(depending on the library type), which contains a file system.
can be used to access storage volumes. logical volume backup
library client A back up of a file system or logical
A server that uses server-to-server volume as a single object.
communication to access a library that is Logical Volume Snapshot Agent (LVSA)
managed by another storage management Software that can act as the snapshot
server. See also library manager. provider for creating a snapshot of a
library manager logical volume during an online image
A server that controls device operations backup.
when multiple storage management loopback virtual file system (LOFS)
servers share a storage device. See also A file system that is created by mounting
library client. a directory over another local directory,
local also known as mount-over-mount. A
1. Pertaining to a device, file, or system LOFS can also be generated using an
that is accessed directly from a user automounter.
system, without the use of a LUN See logical unit number.
communication line. See also remote.
LVSA See Logical Volume Snapshot Agent.
2. For hierarchical storage management
products, pertaining to the destination
of migrated files that are being
moved. See also remote.
local area network (LAN)
A network that connects several devices
in a limited area (such as a single
building or campus) and that can be
connected to a larger network.

Glossary 1125
 storage for Tivoli Storage Manager servers
M that run on operating systems other than
macro file z/OS.
A file that contains one or more storage megabyte (MB)
manager administrative commands, For processor storage, real and virtual
which can be run only from an storage, and channel volume, 2 to the
administrative client using the MACRO 20th power or 1,048,576 bytes. For disk
command. See also Tivoli Storage storage capacity and communications
Manager command script. volume, 1,000,000 bytes.
mailbox restore metadata
A function that restores Microsoft Data that describes the characteristics of
Exchange Server data (from IBM Data data; descriptive data.
Protection for Microsoft Exchange
backups) at the mailbox level or migrate
mailbox-item level. To move data to another location, or an
application to another computer system.
managed object
A definition in the database of a managed migrated file
server that was distributed to the A file that has been copied from a local
managed server by a configuration file system to storage. For HSM clients on
manager. When a managed server UNIX or Linux systems, the file is
subscribes to a profile, all objects that are replaced with a stub file on the local file
associated with that profile become system. On Windows systems, creation of
managed objects in the database of the the stub file is optional. See also file state,
managed server. premigrated file, resident file, stub file.

managed server migration

A server that receives configuration The process of moving data from one
information from a configuration manager computer system to another, or an
using a subscription to one or more application to another computer system.
profiles. Configuration information can migration job
include definitions of objects such as A specification of files to migrate, and
policy and schedules. See also actions to perform on the original files
configuration manager, enterprise after migration. See also job file, threshold
configuration, profile, subscription. migration.
management class migration threshold
A policy object that users can bind to each High and low capacities for storage pools
file to specify how the server manages the or file systems, expressed as percentages,
file. The management class can contain a at which migration is set to start and
backup copy group, an archive copy stop.
group, and space management attributes.
See also bind, copy group, hierarchical mirroring
storage management client, policy set, The process of writing the same data to
rebind. multiple disks at the same time. The
mirroring of data protects it against data
maximum transmission unit (MTU) loss within the database or within the
The largest possible unit of data that can recovery log.
be sent on a given physical medium in a
single frame. For example, the maximum mode A copy group attribute that specifies
transmission unit for Ethernet is 1500 whether to back up a file that has not
bytes. been modified since the last time the file
was backed up. See also absolute mode,
MB See megabyte. modified mode.
media server modified mode
In a z/OS® environment, a program that In storage management, a backup
provides access to z/OS disk and tape copy-group mode that specifies that a file

1126 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 is considered for incremental backup only storage (NAS) file server. Data for the
if it has changed since the last backup. A NAS node is transferred by a NAS file
file is considered a changed file if the server that is controlled by the network
date, size, owner, or permissions of the data management protocol (NDMP). A
file have changed. See also absolute NAS node is also called a NAS file server
mode, mode. node.
mount limit native file system
The maximum number of volumes that A file system that is locally added to the
can be simultaneously accessed from the file server and is not added for space
same device class. The mount limit management. The hierarchical storage
determines the maximum number of manager (HSM) client does not provide
mount points. See also mount point. space management services to the file
system.
mount point
A logical drive through which volumes native format
are accessed in a sequential access device A format of data that is written to a
class. For removable media device types, storage pool directly by the server. See
such as tape, a mount point is a logical also non-native data format.
drive associated with a physical drive. For
NDMP
the file device type, a mount point is a
See Network Data Management Protocol.
logical drive associated with an I/O
stream. See also mount limit. NetBIOS (Network Basic Input/Output System)
A standard interface to networks and
mount retention period
personal computers that is used on local
The maximum number of minutes that
area networks to provide message,
the server retains a mounted
print-server, and file-server functions.
sequential-access media volume that is
Application programs that use NetBIOS
not being used before it dismounts the
do not have to handle the details of LAN
sequential-access media volume.
data link control (DLC) protocols.
mount wait period
network-attached storage file server (NAS file
The maximum number of minutes that
server)
the server waits for a sequential-access
A dedicated storage device with an
volume mount request to be satisfied
operating system that is optimized for
before canceling the request.
file-serving functions. A NAS file server
MTU See maximum transmission unit. can have the characteristics of both a
node and a data mover.
N Network Basic Input/Output System
See NetBIOS.
Nagle algorithm
An algorithm that reduces congestion of Network Data Management Protocol (NDMP)
TCP/IP networks by combining smaller A protocol that allows a network
packets and sending them together. storage-management application to
control the backup and recovery of an
named pipe NDMP-compliant file server, without
A type of interprocess communication installing vendor-acquired software on
that permits message data streams to pass that file server.
between peer processes, such as between
a client and a server. network data-transfer rate
A rate that is calculated by dividing the
NAS file server total number of bytes that are transferred
See network-attached storage file server. by the data transfer time. For example,
NAS file server node this rate can be the time that is spent
See NAS node. transferring data over a network.
NAS node node A file server or workstation on which the
A client node that is a network-attached

Glossary 1127
 backup-archive client program has been contacting for space management services.
installed, and which has been registered For example, a stub file can be orphaned
to the server. when the client system-options file is
modified to contact a server that is
node name
different than the one to which the file
A unique name that is used to identify a
was migrated.
workstation, file server, or PC to the
server.
node privilege class
P
A privilege class that gives an packet In data communication, a sequence of
administrator the authority to remotely binary digits, including data and control
access backup-archive clients for a specific signals, that are transmitted and switched
client node or for all clients in a policy as a composite whole.
domain. See also privilege class.
page A defined unit of space on a storage
non-native data format medium or within a database volume.
A format of data that is written to a
storage pool that differs from the format partial-file recall mode
that the server uses for operations. See A recall mode that causes the hierarchical
also native format. storage management (HSM) function to
read just a portion of a migrated file from
storage, as requested by the application
O accessing the file.
offline volume backup password generation
A backup in which the volume is locked A process that creates and stores a new
so that no other system applications can password in an encrypted password file
access it during the backup operation. when the old password expires.
Automatic generation of a password
online volume backup
prevents password prompting.
A backup in which the volume is
available to other system applications path An object that defines a one-to-one
during the backup operation. relationship between a source and a
destination. Using the path, the source
open registration
accesses the destination. Data can flow
A registration process in which users can
from the source to the destination, and
register their workstations as client nodes
back. An example of a source is a data
with the server. See also closed
mover (such as a network-attached
registration.
storage [NAS] file server), and an
operator privilege class example of a destination is a tape drive.
A privilege class that gives an
pattern-matching character
administrator the authority to disable or
See wildcard character.
halt the server, enable the server, cancel
server processes, and manage removable physical file
media. See also privilege class. A file that is stored in one or more
storage pools, consisting of either a single
options file
logical file, or a group of logical files that
A file that contains processing options.
are packaged together as an aggregate.
See also client system-options file, client
See also aggregate, logical file, physical
user-options file.
occupancy.
originating file system
physical occupancy
The file system from which a file was
The amount of space that is used by
migrated. When a file is recalled, it is
physical files in a storage pool. This space
returned to its originating file system.
includes the unused space that is created
orphaned stub file when logical files are deleted from
A file for which no migrated file can be aggregates. See also logical file, logical
found on the server that the client node is occupancy, physical file.

1128 IBM Tivoli Storage Manager for Windows: Administrator's Guide

plug-in primary storage pool
A separately installable software module A named set of volumes that the server
that adds function to an existing program, uses to store backup versions of files,
application, or interface. archive copies of files, and files migrated
from client nodes. See also copy storage
policy domain
pool, server storage, storage pool, storage
A grouping of policy users with one or
pool volume.
more policy sets, which manage data or
storage resources for the users. The users privilege class
are client nodes that are associated with A level of authority that is granted to an
the policy domain. See also active policy administrator. The privilege class
set, domain. determines which administrative tasks the
administrator can perform. See also
policy privilege class
authority, node privilege class, operator
A privilege class that gives an
privilege class, policy privilege class,
administrator the authority to manage
storage privilege class, system privilege
policy objects, register client nodes, and
class.
schedule client operations for client
nodes. Authority can be restricted to profile
certain policy domains. See also privilege A named group of configuration
class. information that can be distributed from a
configuration manager when a managed
policy set
server subscribes. Configuration
A group of rules in a policy domain. The
information can include registered
rules specify how data or storage
administrator IDs, policies, client
resources are automatically managed for
schedules, client option sets,
client nodes in the policy domain. Rules
administrative schedules, storage manager
can be contained in management classes.
command scripts, server definitions, and
See also active policy set, management
server group definitions. See also
class.
configuration manager, enterprise
premigrated file configuration, managed server.
A file that has been copied to server
profile association
storage, but has not been replaced with a
On a configuration manager, the defined
stub file on the local file system. An
relationship between a profile and an
identical copy of the file resides both on
object such as a policy domain. Profile
the local file system and in server storage.
associations define the configuration
Premigrated files occur on UNIX and
information that is distributed to a
Linux file systems to which space
managed server when it subscribes to the
management has been added. See also file
profile.
state, migrated file, resident file.
premigrated files database
A database that contains information
Q
about each file that has been premigrated quota
to server storage. 1. For HSM on AIX, UNIX, or Linux
premigration systems, the limit (in megabytes) on
The process of copying files that are the amount of data that can be
eligible for migration to server storage, migrated and premigrated from a file
but leaving the original file intact on the system to server storage.
local file system. 2. For HSM on Windows systems, a
premigration percentage user-defined limit to the space that is
A space management setting that controls occupied by recalled files.
whether the next eligible candidates in a
file system are premigrated following
threshold or demand migration.

Glossary 1129
 During the reconciliation process, data
R that is identified as no longer needed is
randomization removed.
The process of distributing schedule start recovery log
times for different clients within a A log of updates that are about to be
specified percentage of the schedule's written to the database. The log can be
startup window. used to recover from system and media
raw logical volume failures. The recovery log consists of the
A portion of a physical volume that is active log (including the log mirror) and
comprised of unallocated blocks and has archive logs.
no journaled file system (JFS) definition. register
A logical volume is read/write accessible To define a client node or administrator
only through low-level I/O functions. ID that can access the server.
rebind registry
To associate all backed-up versions of a A repository that contains access and
file with a new management class name. configuration information for users,
For example, a file that has an active systems, and software.
backup version is rebound when a later
version of the file is backed up with a remote
different management class association. For hierarchical storage management
See also bind, management class. products, pertaining to the origin of
migrated files that are being moved. See
recall To copy a migrated file from server also local.
storage back to its originating file system
using the hierarchical storage resident file
management client. See also selective On a Windows system, a complete file on
recall. a local file system that might also be a
migrated file because a migrated copy can
receiver exist in server storage. On a UNIX or
A server repository that contains a log of Linux system, a complete file on a local
server and client messages as events. For file system that has not been migrated or
example, a receiver can be a file exit, a premigrated, or that has been recalled
user exit, or the server console and from server storage and modified. See
activity log. See also event. also file state.
reclamation restore
The process of consolidating the To copy information from its backup
remaining data from many location to the active storage location for
sequential-access volumes onto fewer, use. For example, to copy information
new sequential-access volumes. from server storage to a client
reclamation threshold workstation.
The percentage of space that a retention
sequential-access media volume must The amount of time, in days, that inactive
have before the server can reclaim the backed-up or archived files are kept in the
volume. Space becomes reclaimable when storage pool before they are deleted.
files are expired or are deleted. Copy group attributes and default
reconciliation retention grace periods for the domain
The process of ensuring consistency define retention.
between the original data repository and retrieve
the larger system where the data is stored To copy archived information from the
for backup. Examples of larger systems storage pool to the workstation for use.
where the data is stored for backup are The retrieve operation does not affect the
storage servers or other storage systems. archive version in the storage pool. See
also archive.

1130 IBM Tivoli Storage Manager for Windows: Administrator's Guide

root user selective recall
A system user who operates without The process of copying user-selected files
restrictions. A root user has the special from server storage to a local file system.
rights and privileges needed to perform See also recall, transparent recall.
administrative tasks.
serialization
The process of handling files that are
S modified during backup or archive
processing. See also shared dynamic
SAN See storage area network. serialization, shared static serialization,
schedule static serialization.
A database record that describes client server A software program or a computer that
operations or administrative commands to provides services to other software
be processed. See also administrative programs or other computers. See also
command schedule, client schedule. client.
scheduling mode server options file
The type of scheduling operation for the A file that contains settings that control
server and client node that supports two various server operations. These settings
scheduling modes: client-polling and affect such things as communications,
server-prompted. devices, and performance.
scratch volume server-prompted scheduling mode
A labeled volume that is either blank or A client/server communication technique
contains no valid data, that is not defined, where the server contacts the client node
and that is available for use. See also when tasks must be done. See also
volume. client-polling scheduling mode.
script A series of commands, combined in a file, server storage
that carry out a particular function when The primary, copy, and active-data storage
the file is run. Scripts are interpreted as pools that are used by the server to store
they are run. See also Tivoli Storage user files such as backup versions, archive
Manager command script. copies, and files migrated from
Secure Sockets Layer (SSL) hierarchical storage management client
A security protocol that provides nodes (space-managed files). See also
communication privacy. With SSL, active-data pool, copy storage pool,
client/server applications can primary storage pool, storage pool
communicate in a way that is designed to volume, volume.
prevent eavesdropping, tampering, and session
message forgery. A logical or virtual connection between
selective backup two stations, software programs, or
The process of backing up certain files or devices on a network that allows the two
directories from a client domain. The files elements to communicate and exchange
that are backed up are those that are not data for the duration of the session. See
excluded in the include-exclude list. The also administrative session.
files must meet the requirement for session resource usage
serialization in the backup copy group of The amount of wait time, processor time,
the management class that is assigned to and space that is used or retrieved during
each file. See also incremental backup. a client session.
selective migration shadow copy
The process of copying user-selected files A snapshot of a volume. The snapshot
from a local file system to server storage can be taken while applications on the
and replacing the files with stub files on system continue to write data to the
the local file system. See also demand volumes.
migration, threshold migration.

Glossary 1131
shadow volume special file
The data stored from a snapshot of a On AIX, UNIX, or Linux systems, a file
volume. The snapshot can be taken while that defines devices for the system, or
applications on the system continue to temporary files that are created by
write data to the volumes. processes. There are three basic types of
special files: first-in, first-out (FIFO);
shared dynamic serialization
block; and character.
A value for serialization that specifies that
a file must not be backed up or archived SSL See Secure Sockets Layer.
if it is being modified during the
stabilized file space
operation. The backup-archive client
A file space that exists on the server but
retries the backup or archive operation a
not on the client.
number of times; if the file is being
modified during each attempt, the stanza A group of lines in a file that together
backup-archive client will back up or have a common function or define a part
archive the file on its last try. See also of the system. Stanzas are usually
dynamic serialization, serialization, shared separated by blank lines or colons, and
static serialization, static serialization. each stanza has a name.
shared library startup window
A library device that is used by multiple A time period during which a schedule
storage manager servers. See also library. must be initiated.
shared static serialization static serialization
A copy-group serialization value that A copy-group serialization value that
specifies that a file must not be modified specifies that a file must not be modified
during a backup or archive operation. The during a backup or archive operation. If
client attempts to retry the operation a the file is in use during the first attempt,
number of times. If the file is in use the backup-archive client cannot back up
during each attempt, the file is not backed or archive the file. See also dynamic
up or archived. See also dynamic serialization, serialization, shared dynamic
serialization, serialization, shared dynamic serialization, shared static serialization.
serialization, static serialization.
storage agent
snapshot A program that enables the backup and
An image backup type that consists of a restoration of client data directly to and
point-in-time view of a volume. from storage attached to a storage area
network (SAN).
space-managed file
A file that is migrated from a client node storage area network (SAN)
by the hierarchical storage management A dedicated storage network tailored to a
(HSM) client. The HSM client recalls the specific environment, combining servers,
file to the client node on demand. systems, storage products, networking
products, software, and services.
space management
See hierarchical storage management. storage hierarchy
A logical order of primary storage pools,
space monitor daemon
as defined by an administrator. The order
A daemon that checks space usage on all
is typically based on the speed and
file systems for which space management
capacity of the devices that the storage
is active, and automatically starts
pools use. The storage hierarchy is
threshold migration when space usage on
defined by identifying the next storage
a file system equals or exceeds its high
pool in a storage pool definition. See also
threshold.
storage pool.
sparse file
storage pool
A file that is created with a length greater
A named set of storage volumes that is
than the data it contains, leaving empty
the destination that is used to store client
spaces for the future addition of data.

1132 IBM Tivoli Storage Manager for Windows: Administrator's Guide

 data. See also active-data pool, copy
storage pool, primary storage pool,
storage hierarchy.
storage pool volume
A volume that has been assigned to a
storage pool. See also active-data pool,
copy storage pool, primary storage pool,
server storage, volume.
storage privilege class
A privilege class that gives an
administrator the authority to control how
storage resources for the server are
allocated and used, such as monitoring
the database, the recovery log, and server
storage. See also privilege class.
stub A shortcut on the Windows file system
that is generated by the hierarchical
storage management (HSM) client for a
migrated file that allows transparent user
access. A stub is the sparse file
representation of a migrated file, with a
reparse point attached.
stub file
A file that replaces the original file on a
local file system when the file is migrated
to storage. A stub file contains the
information that is necessary to recall a
migrated file from server storage. It also
contains additional information that can
be used to eliminate the need to recall a
migrated file. See also migrated file,
resident file.
stub file size
The size of a file that replaces the original
file on a local file system when the file is
migrated to server storage. The size that
is specified for stub files determines how
much leader data can be stored in the
stub file. The default for stub file size is
the block size defined for a file system
minus 1 byte.
subscription
In a storage environment, the process of
identifying the subscribers to which the
profiles are distributed. See also
enterprise configuration, managed server.
system privilege class
A privilege class that gives an
administrator the authority to issue all
server commands. See also privilege class.

Glossary 1133
 tombstone object
T A small subset of attributes of a deleted
tape library object. The tombstone object is retained
A set of equipment and facilities that for a specified period, and at the end of
support an installation's tape the specified period, the tombstone object
environment. The tape library can include is permanently deleted.
tape storage racks, mechanisms for Transmission Control Protocol/Internet Protocol
automatic tape mounting, a set of tape (TCP/IP)
drives, and a set of related tape volumes An industry-standard, nonproprietary set
mounted on those drives. of communication protocols that provides
tape volume prefix reliable end-to-end connections between
The high-level-qualifier of the file name applications over interconnected networks
or the data set name in the standard tape of different types. See also communication
label. method.
target node transparent recall
A client node for which other client nodes The process that is used to automatically
(called agent nodes) have been granted recall a migrated file to a workstation or
proxy authority. The proxy authority file server when the file is accessed. See
allows the agent nodes to perform also selective recall.
operations such as backup and restore on trusted communications agent (TCA)
behalf of the target node, which owns the A program that handles the sign-on
data. password protocol when clients use
TCA See trusted communications agent. password generation.
TCP/IP
See Transmission Control U
Protocol/Internet Protocol.
UCS-2 A 2-byte (16-bit) encoding scheme based
threshold migration on ISO/IEC specification 10646-1. UCS-2
The process of moving files from a local defines three levels of implementation:
file system to server storage based on the Level 1-No combining of encoded
high and low thresholds that are defined elements allowed; Level 2-Combining of
for the file system. See also automatic encoded elements is allowed only for
migration, demand migration, migration Thai, Indic, Hebrew, and Arabic; Level
job, selective migration. 3-Any combination of encoded elements
are allowed.
throughput
In storage management, the total bytes in UNC See Universal Naming Convention.
the workload, excluding overhead, that
Unicode
are backed up or restored, divided by
A character encoding standard that
elapsed time.
supports the interchange, processing, and
timeout display of text that is written in the
A time interval that is allotted for an common languages around the world,
event to occur or complete before plus many classical and historical texts.
operation is interrupted.
Unicode-enabled file space
Tivoli Storage Manager command script Unicode file space names provide support
A sequence of Tivoli Storage Manager for multilingual workstations without
administrative commands that are stored regard for the current locale.
in the database of the Tivoli Storage
Universally Unique Identifier (UUID)
Manager server. The script can run from
The 128-bit numeric identifier that is used
any interface to the server. The script can
to ensure that two components do not
include substitution for command
have the same identifier. See also Globally
parameters and conditional logic. See also
Unique Identifier.
macro file, script.

1134 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Universal Naming Convention (UNC) volume history file
The server name and network name A file that contains information about
combined. These names together identify volumes that have been used by the
the resource on the domain. server for database backups and for
export of administrator, node, policy, or
UTF-8 Unicode Transformation Format, 8-bit
server data. The file also has information
encoding form, which is designed for ease
about sequential-access storage pool
of use with existing ASCII-based systems.
volumes that have been added, reused, or
The CCSID value for data in UTF-8
deleted. The information is a copy of
format is 1208. See also UCS-2.
volume information that is recorded in
UUID See Universally Unique Identifier. the server database.
Volume Shadow Copy Service (VSS)
V A set of Microsoft application-
programming interfaces (APIs) that are
validate used to create shadow copy backups of
To check a policy set for conditions that volumes, exact copies of files, including
can cause problems if that policy set all open files, and so on.
becomes the active policy set. For
example, the validation process checks VSS See Volume Shadow Copy Service.
whether the policy set contains a default VSS Backup
management class. A backup operation that uses Microsoft
version Volume Shadow Copy Service (VSS)
A backup copy of a file stored in server technology. The backup operation
storage. The most recent backup copy of a produces an online snapshot
file is the active version. Earlier copies of (point-in-time consistent copy) of
the same file are inactive versions. The Microsoft Exchange data. This copy can
number of versions retained by the server be stored on local shadow volumes or on
is determined by the copy group Tivoli Storage Manager server storage.
attributes in the management class. VSS Fast Restore
virtual file space An operation that restores data from a
A representation of a directory on a local snapshot. The snapshot is the VSS
network-attached storage (NAS) file backup that resides on a local shadow
system as a path to that directory. volume. The restore operation retrieves
the data by using a file-level copy
virtual mount point method.
A directory branch of a file system that is
defined as a virtual file system. The VSS Instant Restore
virtual file system is backed up to its own An operation that restores data from a
file space on the server. The server local snapshot. The snapshot is the VSS
processes the virtual mount point as a backup that resides on a local shadow
separate file system, but the client volume. The restore operation retrieves
operating system does not. the data by using a hardware assisted
restore method (for example, a FlashCopy
virtual volume operation).
An archive file on a target server that
represents a sequential media volume to a VSS offloaded backup
source server. A backup operation that uses a Microsoft
Volume Shadow Copy Service (VSS)
volume hardware provider (installed on an
A discrete unit of storage on disk, tape or alternate system) to move IBM Data
other data recording medium that Protection for Microsoft Exchange data to
supports some form of identifier and the Tivoli Storage Manager server. This
parameter list, such as a volume label or type of backup operation shifts the
input/output control. See also scratch backup load from the production system
volume, server storage, storage pool, to another system.
storage pool volume.

Glossary 1135
VSS Restore
A function that uses a Microsoft Volume
Shadow Copy Service (VSS) software
provider to restore VSS Backups (IBM
Data Protection for Microsoft Exchange
database files and log files) that reside on
Tivoli Storage Manager server storage to
their original location.

W
wildcard character
A special character such as an asterisk (*)
or a question mark (?) that can be used to
represent one or more characters. Any
character or set of characters can replace
the wildcard character.
workload partition (WPAR)
A partition within a single operating
system instance.
workstation
A terminal or personal computer at which
a user can run applications and that is
usually connected to a mainframe or a
network.
worldwide name (WWN)
A 64-bit, unsigned name identifier that is
unique.
WPAR See workload partition.
WWN See worldwide name.

1136 IBM Tivoli Storage Manager for Windows: Administrator's Guide

Index
Special characters active log (continued)
increasing the size 678
$$CONFIG_MANAGER$$ 720 move to another directory 681
out of space 678
space requirements 656
Numerics active log mirror 877
3480 tape drive description 648
cleaner cartridge 167 active log size
device support 36 reducing the size 679
mixing drive generations 183 ACTIVE policy set
3490 tape drive creating 500, 510
cleaner cartridge 167 replacing 478
device support 36 active-data pool
mixing drive generations 183 auditing volumes in 890
3494 automated library device 38, 80, 96 backup-set file source 543
3494 library 80 collocation on 364
configuration with a single drive device 98 defining 410
defining devices 106 export file source 747, 756, 757
3494SHARED server option 62 import operations 766
3570 tape drive overview 241, 265
ASSISTVCRRECOVERY server option 62 reclamation of 374
device support 36 simultaneous-write function 331
3590 tape drive specifying in policy definition 498
defining device class 58, 177, 179 storage pool search-and-selection order 243
device support 97, 129 ACTIVELOGDIRECTORY server option 678, 681
3592 drives and media ACTIVELOGSIZE 679
as element of storage hierarchy 260 ACTIVELOGSIZE server option 678, 679
cleaning 165 activity log
data encryption 154, 185, 536 description of 799
defining device class 58 logging events to 809
enabling for WORM media 133 monitoring 799
mixing drive generations 183 querying 800
4mm tape device support 179 setting size limit 801
8mm tape device support 179 setting the retention period 801
adding space to the database 674
administrative client
A description of 3
viewing information after IMPORT or EXPORT 771
absolute mode, description of 503 administrative clients
access authority, client 446, 447 preventing from accessing the server 744
access mode, volume administrative commands
changing 257 ACCEPT DATE 609
description 259 ASSIGN DEFMGMTCLASS 510
determining for storage pool 245, 410 AUDIT LIBRARY 137
access, managing 831, 844 AUDIT LICENSE 597
accessibility features 1109 AUDIT VOLUME 894
accounting record BACKUP NODE 228, 230
description of 803 BEGIN EVENTLOGGING 809
monitoring 803 CANCEL PROCESS 610
accounting variable 803 CANCEL REQUEST 144
ACSLS (Automated Cartridge System Library Software) CANCEL RESTORE 472
StorageTek library 37 CANCEL SESSION 469
configuring 109 CHECKIN LIBVOLUME 130
description 39 CHECKOUT LIBVOLUME 136
mixing 3592 drive generations 183 CLEAN DRIVE 164
Tivoli Storage Manager server options for 62 COMMIT 640
ACTIVATE POLICYSET command 510 COPY ACTIVEDATA 242, 265
active data 944 COPY CLOPTSET 467
active data, protecting with active-data pools 241 COPY DOMAIN 500
active files, storage-pool search order 243 COPY POLICYSET 500
active log 877 COPY SCHEDULE 570
description 647

© Copyright IBM Corp. 1993, 2013 1137

administrative commands (continued) administrative commands (continued)
COPY SCRIPT 634 EXPIRE INVENTORY 27
COPY SERVERGROUP 730 EXPORT ADMIN 741
DEFINE ASSOCIATION 562 EXPORT NODE 755
DEFINE BACKUPSET 548 EXPORT POLICY 755
DEFINE CLIENTACTION 583 EXPORT SERVER 755
DEFINE CLIENTOPT 583 EXTEND DBSPACE 672, 676
DEFINE CLOPTSET 464 GENERATE BACKUPSET 544
DEFINE COPYGROUP 502, 507, 508 GRANT AUTHORITY 845
DEFINE DATAMOVER 175, 224 HALT 607
DEFINE DEVCLASS HELP 616
3592 183 IMPORT 770, 771
FILE device classes 187 IMPORT ADMIN 758
LTO device classes 191 IMPORT NODE 758, 767
REMOVEABLEFILE device classes 186 IMPORT POLICY 758
SERVER device classes 194 IMPORT SERVER 758, 767
tape device classes 179 LOCK ADMIN 852
VOLSAFE device classes 195 LOCK NODE 440
DEFINE DRIVE 174 LOCK PROFILE 711, 712
DEFINE GRPMEMBER 729 MACRO 138
DEFINE LIBRARY 38, 173 MOVE DATA 401
DEFINE MACHINE 1013 MOVE MEDIA 138
DEFINE MACHNODEASSOCIATION 1013 MOVE NODEDATA 406
DEFINE PATH 176 NOTIFY SUBSCRIBERS 712
DEFINE POLICYSET 500, 501 PING SERVER 732
DEFINE PROFASSOCIATION 707, 708 PREPARE 1018
DEFINE PROFILE 707 QUERY ACTLOG 800
DEFINE RECMEDMACHASSOCIATION 1017 QUERY BACKUPSETCONTENTS 551
DEFINE RECOVERYMEDIA 1017 QUERY CONTENT 388
DEFINE SCHEDULE 621 QUERY COPYGROUP 529, 765
DEFINE SCRIPT 627 QUERY DB 671
DEFINE SERVER 694, 726, 732 QUERY DBSPACE 671
DEFINE SERVERGROUP 729 QUERY DEVCLASS 198
DEFINE STGPOOL 249, 250, 261, 262 QUERY DOMAIN 531
DEFINE SUBSCRIPTION 719 QUERY DRIVE 152
DEFINE VIRTUALFSMAPPING 236 QUERY DRMSTATUS 1006
DEFINE VOLUME 46, 256 QUERY ENABLED 822
DELETE ASSOCIATION 573 QUERY EVENT 563
DELETE BACKUPSET 552 QUERY FILESPACE 462
DELETE COPYGROUP 532 QUERY LIBRARY 150
DELETE DOMAIN 532 QUERY LICENSE 597
DELETE DRIVE 168 QUERY MEDIA 138
DELETE EVENT 575 QUERY MGMTCLASS 530
DELETE GRPMEMBER 731 QUERY MOUNT 144
DELETE KEYRING 837 QUERY NODE 443
DELETE LIBRARY 152 QUERY NODEDATA 398
DELETE MGMTCLASS 532 QUERY OCCUPANCY
DELETE POLICYSET 532 backed-up, archived, and space-managed files 399
DELETE PROFASSOCIATION 712 client file spaces 397
DELETE PROFILE 714 client nodes 397
DELETE SCHEDULE 570 device classes 398
DELETE SCRIPT 636 storage pools 398
DELETE SERVER 700 QUERY OPTION 792
DELETE SERVERGROUP 731 QUERY POLICYSET 530
DELETE STGPOOL 413 QUERY PROCESS 404
DELETE SUBSCRIBER 724 QUERY REQUEST 144
DELETE SUBSCRIPTION 714, 720 QUERY RESTORE 472
DELETE VOLHISTORY 615 QUERY RPFCONTENT 1022
DELETE VOLUME 415, 416 QUERY RPFILE 1022
DISABLE EVENTS 808 QUERY SCHEDULE 563
DISABLE SESSIONS 471 QUERY SCRIPT 635
DISMOUNT VOLUME 145 QUERY SERVERGROUP 730
DSMSERV DISPLAY DBSPACE 671 QUERY STGPOOL 382, 392, 761
DSMSERV DISPLAY LOG 671 QUERY SUBSCRIPTION 719
ENABLE EVENTS 808 QUERY SYSTEM 793
ENABLE SESSIONS 471 QUERY VOLUME 384, 405
END EVENTLOGGING 809 RECONCILE VOLUMES 739

1138 IBM Tivoli Storage Manager for Windows: Administrator's Guide

administrative commands (continued) administrative commands (continued)
REGISTER ADMIN 847 UPDATE LIBRARY 151
REGISTER LICENSE 596 UPDATE LIBVOLUME 46, 135
REMOVE ADMIN 851 UPDATE MGMTCLASS 501
REMOVE NODE 440 UPDATE NODE 434
RENAME ADMIN 850 UPDATE POLICYSET 500
RENAME FILESPACE 769 UPDATE RECOVERYMEDIA 1017
RENAME NODE 439 UPDATE SCHEDULE 621
RENAME SCRIPT 636 UPDATE SCRIPT 633, 634
RENAME SERVERGROUP 730 UPDATE SERVER 700
RENAME STGPOOL 409 UPDATE SERVERGROUP 730
REPLY 144 UPDATE STGPOOL 138
RESTORE DB 608 UPDATE VOLUME 139, 256
RESTORE NODE 228, 230 VALIDATE LANFREE 123
RESTORE STGPOOL 916 administrative privilege class
ROLLBACK 641 description 845
RUN 636 granting authority 845
SELECT 794 reducing 850
SET ACCOUNTING 803 revoking all 850
SET AUTHENTICATION 865 administrative user ID
SET CLIENTACTDURATION 583 creating automatically 448
SET CONFIGMANAGER 703, 706 description of 420
SET CONFIGREFRESH 719 preventing automatic creation of 448
SET CONTEXTMESSAGING 808 administrative web interface
SET CROSSDEFINE 695, 698 description 17
SET DBREPORTMODE 671 administrative Web interface
SET DRMCHECKLABEL 1009 limitation of browser for script definitions 627
SET DRMCOPYSTGPOOL 1006 administrator
SET DRMCOURIERNAME 1009 authorizing to manage a policy domain 845
SET DRMDBBACKUPRXPIREDAYS 1009 locking 852
SET DRMFILEPROCESS 1009 managing registration 595
SET DRMINSTPREFIX 1006 querying 848
SET DRMNOTMOUNTABLE 1009 registering 847
SET DRMPLANPOSTFIX 1006 removing 851
SET DRMPLANPREFIX 1006 renaming 850
SET DRMPRIMSTGPOOL 1006 restrictions when registering 847
SET DRMRPFEXPIREDAYS 1022 unlocking 852
SET DRMVAULTNAME 1009 updating 847, 850
SET EVENTRETENTION 575, 626 viewing information about 848
SET INVALIDPWLIMIT 863 administrators
SET LICENSEAUDITPERIOD 597 managing 844
SET MAXCMDRETRIES 582 aggregates
SET MAXSCHEDSESSIONS 579 controlling the size of 262
SET MINPWLENGTH 864 estimating size 388
SET PASSEXP 860 RECONSTRUCT parameter 405
SET QUERYSCHEDPERIOD 581 reconstructing 368, 375, 405
SET RANDOMIZE 580 viewing information about 390, 397
SET REGISTRATION 421 alert monitoring 587, 590
SET RETRYPERIOD 582 alias name, device 75
SET SCHEDMODES 577 ANR9999D message 808
SET SERVERHLADDRESS 695, 698 application client
SET SERVERLLADDRESS 695, 698 adding node for 420
SET SERVERNAME 694, 698, 792 description 4
SET SERVERPASSWORD 694, 695, 698 policy for 523
SET SUBFILE 552 application program interface (API)
SET SUMMARYRETENTION 798 client, registering 424
SETOPT 615 compression option 424
UNLOCK PROFILE 711, 712 deletion option 424
UPDATE ADMIN 847 registering to server 424
UPDATE ARCHIVE 556 simultaneous-write function, version support for 333
UPDATE BACKUPSET 550 application programming interface (API)
UPDATE CLIENTOPT 467 description of 3
UPDATE CLOPTSET 467 ARCHFAILOVERLOGDIRECTORY server option 684
UPDATE COPYGROUP 502, 508 archive
UPDATE DEVCLASS 179 allowing while file changes 508
UPDATE DOMAIN 500 backup set, uses for 9, 13
UPDATE DRIVE 153 determining storage usage 399

Index 1139
archive (continued) Automated Cartridge System Library Software (ACSLS)
directory 555 StorageTek library (continued)
increasing throughput using DSMMAXSG utility 126 mixing 3592 drive generations 183
instant 9, 13 Tivoli Storage Manager server options for 62
package 555 automated library device
policy, defining 497 adding, using commands 91
policy, introduction 24 auditing 137
process description 495 checking in media 130
storage usage, minimizing 555, 556 defining 38
storage usage, reducing 555, 556 managing media 135
uses for 8, 12 overflow location 245
archive copy group replacing tape drive 159
defining 508, 510 reusing media 141
deleting 532 scratch and private volumes 46
description of 482 updating 151
archive data volume inventory 47, 128
expiration 514 automatic deployment 434
managing 554 verifying 438
protection 514 automatic failover
archive failover log 877 overview 920
description 649 automatic failover scenario 968
move to another directory 681 automatically renaming file spaces 456
archive log 877 automating
description 648 administrative commands 27
move to another directory 681 client operations 560
space requirements 656 server operations 620
archiving awk script 1048
file 481, 495
file management 481
FILE-type volume, archiving many small objects to 188
ASCII restriction for browser script definition 627
B
background processes 609
ASSIGN DEFMGMTCLASS command 510
backup
association, client with schedule
amount of space used by client 399
defining 562
comparison of types 10, 13
deleting 573
default policy 477
association, file with management class 489, 490
defining criteria for client files 497
association, object with profile
differential, for NAS node 10, 51
administrative command schedule 711
file 480, 492, 494
administrator 707, 723
file management 480
client option set 708
file while open 502
deleting 712
FILE-type volume, backing up small objects to 188
policy domain 709
frequency for file 503
script 708
full, for NAS node 51
AUDIT LIBRARY command 137
group 12
AUDIT LICENSE command 597
increasing throughput using DSMMAXSG utility 126
AUDIT VOLUME command 887, 894
incremental 480, 492
auditing
logical volume 495
LDAP directory server 914
NAS file server 213
library's volume inventory 137
NAS file server to Tivoli Storage Manager server 232
license, automatic by server 597
NAS filer to Tivoli Storage Manager server 231
multiple volumes in sequential access storage pool 894
policy 24
single volume in sequential access storage pool 895
selective 480, 494
volume in disk storage pool 894
SnapMirror to Tape 237
volume, reasons for 887
snapshot, using hardware 9, 11
volumes by date 895
subfiles, server set-up 23, 552
volumes by storage pool 896
types available 10, 13
authority
BACKUP command 680
client access 447
backup copy group
granting to administrators 845
defining 502, 507
privilege classes 845
deleting 532
server options 845
description of 482
AUTOFSRENAME parameter 456
frequency 492
AUTOLABEL parameter for tape volumes 129
mode 492
Automated Cartridge System Library Software (ACSLS)
serialization 492
StorageTek library
backup data, protecting active 241
configuring 109
BACKUP DEVCONFIG command 879
description 39

1140 IBM Tivoli Storage Manager for Windows: Administrator's Guide

backup media cartridge
move 1024 cleaner cartridge 167
backup period, specifying for incremental 579 device support 36
backup set mixing drive generations 183
adding subfiles to 554 categories, 349X library 38, 96
deleting 552 CD-ROM media (REMOVABLEFILE device class) 89
description of 544 Celerra
displaying contents of 551 file server integrated checkpoints 238
example of generating 546 Centera libraries 83
generating 544 Centera SDK
how the server manages and tracks 549 installing 83
media, selecting 545 Centera storage device
moving to other servers 548 concurrent access 197
OST extension on 545 overview 42
selecting a name for 545 restore improve 197
selecting retention period for 549 unsupported functions 249
suggested usage 9, 23 unsupported server operations 197
updating 550 Centera storage pool, backing up 886
use as archive 12, 15, 23 central monitoring 689
backup sets Central Processing Unit
displaying information 550 CPU 598
generate to a point-in-time 546 central scheduling
generate with multiple data types 547 client operations 535, 559, 569, 576
backup sets, managing 549 controlling the workload 579
BACKUP VOLHISTORY command 878 coordinating 576
backup volumes description of 25, 27, 559
create single set 547 server operations 620
backup-archive client certificate
description of 3 adding to the key database 836, 837
operations summary 10 homegrown certificate authority 837
performing operations for 535, 569, 576 changing date and time on server 609
policy for 486 changing hostname 613
registering node 420 characteristics, machine 1013
scheduling operations for 560 check in
unified logon 865 cleaner cartridge 166
using to back up NAS file server 205, 230 library volume 130
base file 552 setting a time interval for volume 183
batch file, scheduling on client 564 VolSafe-enabled volumes 195
binding a file to a management class 489 CHECKIN LIBVOLUME command 130
browser, limited to ASCII entry for script definition 627 checking the log file generated by processed schedules 575
checklist for DRM project plan 1043
CHECKOUT LIBVOLUME command 136
C CHECKTAPEPOS server option 62
class, administrator privilege
cache
description 845
deleting files from 284, 401
granting authority 845
description of 20
reducing 850
disabling for disk storage pools 284
revoking all 850
effect on performance 284
class, device
effect on statistics 285
3590 177, 179
enabling for disk storage pools 245, 284
3592 179
monitoring utilization on disk 395
4MM 177, 179
calculate replication 942
8MM 177, 179
CANCEL PROCESS command 393, 610
amount of space used 398
CANCEL RESTORE command 472
CENTERA 42
CANCEL SESSION command 469
defining 177
capacity planning
description of 19
database space requirements
DISK 177
estimates based on number of files 652
DLT 177, 179
estimates based storage pool capacity 654
ECARTRIDGE 179
starting size 652
FILE 177
recovery log space requirements
for devices using native Windows device driver 88
active and archive logs 656
FORMAT parameter 180
active log mirror 668
GENERICTAPE 88, 177, 179
archive failover log 669
LTO 191
archive log space for database reorganization 668
overview 94
capacity, tape 199
REMOVABLEFILE 186

Index 1141
class, device (continued) client node (continued)
requesting information about 198 importing 767
selecting for import and export 754 locking 440
sequential 179 managing registration 420, 431, 595
SERVER 177, 179, 734 options file 425
StorageTek devices 179, 195 performing operations for 535, 569, 576
tape 179, 186 privilege class for scheduling operations for 561
Ultrium, LTO 179 proxy node relationships 441
updating 179, 186 querying 443
VOLSAFE 195 reducing archive packages for 556
WORM 177, 179 registering 424
class, policy privilege removing 440
description 844, 845 renaming 439
granting 849 scheduling operations for 560
revoking 849, 850 setting password authentication 865
class, storage privilege setting scheduling mode 578
description 845 setting up subfile backups 553
granting 849 target 442
reducing 849 unlocking 440
revoking 850 updating 434
CLEAN DRIVE command 164 viewing information about 443
cleaner cartridge client nodes
checking in 166 defining 451
operations with 167 file spaces 451
CLI 587 managing 22, 431
client managing across a firewall 432
access user ID 447 prevent from accessing server 744
administrative 3 client option
API (application program interface) 424 TXNBYTELIMIT 263
API (application programming interface) 4 VIRTUALMOUNTPOINT 451
application client 4, 523 client option set
backup-archive 3 adding client options to 465
how to protect 8 assigning clients to 466
operations summary 10 copying 467
options file 425 creating 465
restore without primary volumes available 913 deleting 467
Tivoli Storage Manager for Space Management (HSM deleting an option from 467
client) 4, 486 for NAS node 213
using to back up NAS file server 213, 230 requesting information about 467
client file updating description for 467
allowing archive while changing 477 client options
allowing backup while changing 477, 502 UNIX and Linux 426
archive package 555 client options file
associating with management class 489, 490 create 426
damaged 913 update 426
delaying migration of 278 client point-in-time restore, enabling 527
deleting 414 client queries to the server, setting the frequency 581
deleting from a storage pool 413 client restartable restore session
deleting from cache 284 canceling 472
deleting when deleting a volume 414 interrupting, active 473
duplication when restoring 913 requesting information about 472
eligible for archive 477, 492 client schedule
eligible for backup 477, 492 define 561
eligible for expiration 479 client session
eligible for space management 496 canceling 469
how IBM Tivoli Storage Manager stores 262 DSMC loop 467
on a volume, querying 388 held volume 467
server migration of 271 managing 467
client migration 496 querying 468, 790
client node viewing information about 468, 790
adding 419 client system options file 424
agent 442 client-polling scheduling 577, 581
amount of space used 397 client, application
creating backup sets for 544 adding node for 420
file spaces, QUERY OCCUPANCY command 397 description 4
finding tapes used by 391 policy for 523
immediate processing 583 client/server, description of 3

1142 IBM Tivoli Storage Manager for Windows: Administrator's Guide

clients commands, administrative (continued)
adding through the command line client 428 COPY POLICYSET 500
closed registration COPY SCHEDULE 570
description 421 COPY SCRIPT 634
process 421 DEFINE ASSOCIATION 562
setting 421 DEFINE BACKUPSET 548
cluster 1000, 1002 DEFINE CLIENTACTION 583
administrators tasks 1001 DEFINE CLIENTOPT 583
configuration 1000, 1001 DEFINE CLOPTSET 464
configuration considerations 996 DEFINE COPYGROUP 502, 507, 508
managing SCSI tape failover 1003 DEFINE DATAMOVER 175, 224
server migration 1001 DEFINE DEVCLASS
setup 999 3592 183
cluster configuration 991, 992 FILE device classes 187
cluster failover 994 LTO device classes 191
cluster group 993 REMOVEABLEFILE device classes 186
cluster logs 1003 SERVER device classes 194
cluster nodes 992 tape device classes 179
cluster overview 991 VOLSAFE device classes 195
clustered environment 991, 1002 DEFINE DRIVE 174
administrative tasks 996 DEFINE GRPMEMBER 729
clustering 998 DEFINE LIBRARY 38, 173
hardware and software configuration 997 DEFINE MACHINE 1013
clusters 999 DEFINE MACHNODEASSOCIATION 1013
collocation DEFINE PATH 176
active-data pools 364 DEFINE POLICYSET 500, 501
changing, effect of 363 DEFINE PROFASSOCIATION 707, 708
copy storage pools 364 DEFINE PROFILE 707
definition 245, 357, 410 DEFINE RECMEDMACHASSOCIATION 1017
description of 20 DEFINE RECOVERYMEDIA 1017
determining whether to use collocation 245, 357, 410 DEFINE SCHEDULE 621
effects on operations 359 DEFINE SCRIPT 627
effects on volume reclamation 378 DEFINE SERVER 694, 726, 732
enabling 365 DEFINE SERVERGROUP 729
enabling for sequential storage pool 245, 357, 410 DEFINE STGPOOL 249, 250, 261, 262
how the server selects volumes when disabled 363 DEFINE SUBSCRIPTION 719
migration thresholds 279 DEFINE VIRTUALFSMAPPING 236
planning 365 DEFINE VOLUME 46, 256
selecting volumes when enabled 361 DELETE ASSOCIATION 573
command file, scheduling on client 564 DELETE BACKUPSET 552
command line 587 DELETE COPYGROUP 532
command retry attempts DELETE DOMAIN 532
setting the amount of time between 582 DELETE DRIVE 168
setting the number of 582 DELETE EVENT 575
command routing 726 DELETE GRPMEMBER 731
command script 627 DELETE LIBRARY 152
commands DELETE MGMTCLASS 532
grant authority 848 DELETE POLICYSET 532
commands, administrative DELETE PROFASSOCIATION 712
ACCEPT DATE 609 DELETE PROFILE 714
ASSIGN DEFMGMTCLASS 510 DELETE SCHEDULE 570
AUDIT LIBRARY 137 DELETE SCRIPT 636
AUDIT LICENSE 597 DELETE SERVER 700
AUDIT VOLUME 894 DELETE SERVERGROUP 731
BACKUP NODE 228, 230 DELETE STGPOOL 413
BEGIN EVENTLOGGING 809 DELETE SUBSCRIBER 724
CANCEL PROCESS 610 DELETE SUBSCRIPTION 714, 720
CANCEL REQUEST 144 DELETE VOLHISTORY 615
CANCEL RESTORE 472 DELETE VOLUME 415, 416
CANCEL SESSION 469 DISABLE EVENTS 808
CHECKIN LIBVOLUME 130 DISABLE SESSIONS 471
CHECKOUT LIBVOLUME 136 DISMOUNT VOLUME 145
CLEAN DRIVE 164 DSMSERV DISPLAY DBSPACE 671
COMMIT 640 DSMSERV DISPLAY LOG 671
COPY ACTIVEDATA 242, 265 ENABLE EVENTS 808
COPY CLOPTSET 467 ENABLE SESSIONS 471
COPY DOMAIN 500 END EVENTLOGGING 809

Index 1143
commands, administrative (continued) commands, administrative (continued)
EXPIRE INVENTORY 27 REGISTER ADMIN 847
EXPORT ADMIN 741 REGISTER LICENSE 596
EXPORT NODE 755 REMOVE ADMIN 851
EXPORT POLICY 755 REMOVE NODE 440
EXPORT SERVER 755 RENAME ADMIN 850
EXTEND DBSPACE 672, 676 RENAME FILESPACE 769
GENERATE BACKUPSET 544 RENAME NODE 439
GRANT AUTHORITY 845 RENAME SCRIPT 636
HALT 607 RENAME SERVERGROUP 730
HELP 616 RENAME STGPOOL 409
IMPORT 770, 771 REPLY 144
IMPORT ADMIN 758 RESTORE DB 608
IMPORT NODE 758, 767 RESTORE NODE 228, 230
IMPORT POLICY 758 RESTORE STGPOOL 916
IMPORT SERVER 758, 767 ROLLBACK 641
LOCK ADMIN 852 RUN 636
LOCK NODE 440 SELECT 794
LOCK PROFILE 711, 712 SET ACCOUNTING 803
MACRO 138 SET AUTHENTICATION 865
MOVE DATA 401 SET CLIENTACTDURATION 583
MOVE MEDIA 138 SET CONFIGMANAGER 703, 706
MOVE NODEDATA 406 SET CONFIGREFRESH 719
NOTIFY SUBSCRIBERS 712 SET CONTEXTMESSAGING 808
PING SERVER 732 SET CROSSDEFINE 695, 698
PREPARE 1018 SET DBREPORTMODE 671
QUERY ACTLOG 800 SET DRMCHECKLABEL 1009
QUERY BACKUPSETCONTENTS 551 SET DRMCOPYSTGPOOL 1006
QUERY CONTENT 388 SET DRMCOURIERNAME 1009
QUERY COPYGROUP 529, 765 SET DRMDBBACKUPRXPIREDAYS 1009
QUERY DB 671 SET DRMFILEPROCESS 1009
QUERY DBSPACE 671 SET DRMINSTPREFIX 1006
QUERY DEVCLASS 198 SET DRMNOTMOUNTABLE 1009
QUERY DOMAIN 531 SET DRMPLANPOSTFIX 1006
QUERY DRIVE 152 SET DRMPLANPREFIX 1006
QUERY DRMSTATUS 1006 SET DRMPRIMSTGPOOL 1006
QUERY ENABLED 822 SET DRMRPFEXPIREDAYS 1022
QUERY EVENT 563 SET DRMVAULTNAME 1009
QUERY FILESPACE 462 SET EVENTRETENTION 575, 626
QUERY LIBRARY 150 SET INVALIDPWLIMIT 863
QUERY LICENSE 597 SET LICENSEAUDITPERIOD 597
QUERY MEDIA 138 SET MAXCMDRETRIES 582
QUERY MGMTCLASS 530 SET MAXSCHEDSESSIONS 579
QUERY MOUNT 144 SET MINPWLENGTH 864
QUERY NODE 443 SET PASSEXP 860
QUERY NODEDATA 398 SET QUERYSCHEDPERIOD 581
QUERY OCCUPANCY SET RANDOMIZE 580
backed-up, archived, and space-managed files 399 SET REGISTRATION 421
client file spaces 397 SET RETRYPERIOD 582
client nodes 397 SET SCHEDMODES 577
device classes 398 SET SERVERHLADDRESS 695, 698
storage pools 398 SET SERVERLLADDRESS 695, 698
QUERY OPTION 792 SET SERVERNAME 694, 698, 792
QUERY POLICYSET 530 SET SERVERPASSWORD 694, 695, 698
QUERY PROCESS 404 SET SUBFILE 552
QUERY REQUEST 144 SET SUMMARYRETENTION 798
QUERY RESTORE 472 SETOPT 615
QUERY RPFCONTENT 1022 UNLOCK PROFILE 711, 712
QUERY RPFILE 1022 UPDATE ADMIN 847
QUERY SCHEDULE 563 UPDATE ARCHIVE 556
QUERY SCRIPT 635 UPDATE BACKUPSET 550
QUERY SERVERGROUP 730 UPDATE CLIENTOPT 467
QUERY STGPOOL 382, 392, 761 UPDATE CLOPTSET 467
QUERY SUBSCRIPTION 719 UPDATE COPYGROUP 502, 508
QUERY SYSTEM 793 UPDATE DEVCLASS 179
QUERY VOLUME 384, 405 UPDATE DOMAIN 500
RECONCILE VOLUMES 739 UPDATE DRIVE 153

1144 IBM Tivoli Storage Manager for Windows: Administrator's Guide

commands, administrative (continued) configuring (continued)
UPDATE LIBRARY 151 devices not supported by the IBM Tivoli Storage Manager
UPDATE LIBVOLUME 46, 135 device driver 88
UPDATE MGMTCLASS 501 devices, automated library example 97
UPDATE NODE 434 hub server 593
UPDATE POLICYSET 500 NDMP operations for NAS file servers 211
UPDATE RECOVERYMEDIA 1017 Operations Center 591, 593
UPDATE SCHEDULE 621 planning your storage environment 60
UPDATE SCRIPT 633, 634 shared library 102, 119
UPDATE SERVER 700 VTL 117
UPDATE SERVERGROUP 730 configuring storage 85
UPDATE STGPOOL 138 console mode 771
UPDATE VOLUME 139, 256 contents of a volume 388
VALIDATE LANFREE 123 context messaging for ANR9999D 808
commands, administrative COPY SERVERGROUP command continuation characters, using 631
COPY SERVERGROUP 730 controlling devices 82
COMMIT command 640 conventions
Common Agent Services typographic xv
CAS 598 COPY CLOPTSET command 467
Common Inventory Technology 598 COPY DOMAIN command 500
COMMTIMEOUT server option 468, 470 copy group
communication set up archive, description of 482
among servers 692 backup, description of 482
command routing, for 696 defining archive 508
cross definition 693, 694, 698 defining backup 502
enterprise configuration, for 692 deleting 532
enterprise event logging, for 692, 820 COPY MGMTCLASS command 501
security 695 COPY POLICYSET command 500
server-to-server virtual volumes 734 COPY SCHEDULE command 570, 624
using Secure Sockets Layer 833 COPY SCRIPT command 634
using Transport Layer Security (TLS) 833 copy storage pool
compression compared with primary 412
choosing client or drive 200 defining a 410
option for API 424 restore from multiple 913
options for clients 422 role in storage pool migration 283
setting 422 simultaneous-write function 331
tape volume capacity, effect on 200 creating
concurrent access backup sets 23
Centera volumes 197 server scripts 627
configuration file, device creating backup sets
unrecognized by IBM Tivoli Storage Manager 88 benefits of 544
configuration information, enterprise management example for 546
administrative command schedule 704, 711 cross definition 693, 694, 698
administrator 707, 723 customer support
changing 711 contacting 1106
client option set 704, 708 cyclic redundancy check
client schedule 704, 705, 709 during a client session 535
copy group 704, 709 for storage pool volumes 890
deleting 712, 714 for virtual volumes 732
distributing 701, 708, 712 performance considerations for nodes 536
management class 709 performance considerations for storage pools 893
policy domain 704, 705, 709
refreshing 712, 719, 721
script 704, 708
server 710
D
daily monitoring disk storage pools 782
server group 710
daily monitoring of databases 779
configuration manager
daily monitoring of server processes 778
communication setup 692
daily monitoring scheduled operations 786
default profile 703, 710
daily monitoring sequential access storage pools 783
scenario 703
damaged files 896, 897
setting up 703, 706, 707
data
configuration settings 587
active backup versions, storing 241
configuration wizard
considering user needs for recovering 59
Operations Center 593
exporting 741
configuring
importing 741
connect 349x library to server 221
data compression 422
connecting ACSLS library to server 222
data deduplication xx, 317, 318

Index 1145
data deduplication (continued) data shredding (continued)
checklist for configuration 293 EXPIRE INVENTORY command 541
client-side 314 EXPORT NODE command 541, 742
changing location 316 EXPORT SERVER command 541, 742
client and server settings 285, 310 GENERATE BACKUPSET command 541, 543
multiple nodes 315 MOVE DATA command 402, 541
overview 286 setting up 540
single node 315 UPDATE STGPOOL command 541
controlling duplicate-identification manually 312 data storage
data deduplication 295, 296, 297, 298, 299, 300, 301, 324, active-data pools 241
325, 326, 327, 328, 329, 330 client files, process for storing 5
DEDUPLICATION parameter 310 concepts overview 15
DEDUPREQUIRESBACKUP server option 307 considering user needs for recovering 59
definition 285 deleting files from 414
detecting security attacks 302 evaluating 60
duplicate-identification processes 306, 311, 313, 314 example 243
IDENTIFY DUPLICATES command 312 managing 18
limitations 288 monitoring 887
managing 306 planning 60
memory xx server options affecting 62
moving or copying data 308 tailoring definitions 765
node replication 931, 987 using another IBM Tivoli Storage Manager server 732
options for 318 using devices not recognized by IBM Tivoli Storage
planning 291 Manager 88
processing 305 using disk devices 63
protecting data 307 using removable media devices 89
reclamation 307 using the storage hierarchy 270
requirements 293 data validation
server-side 285, 286, 310 during a client session 535
specifying the size of objects to be deduplicated 316 for storage pool volumes 890
statistics for virtual volumes 732
displaying information about files with links to a performance considerations for nodes 536
volume 319 performance considerations for storage pools 893
querying a duplicate-identification process 319, 320, database
321, 323 adding directories 674
querying a storage pool 318 audits 644
testing backup 870, 871, 875, 876
restore operations 303 buffer size 644
space savings 304 description of 643
Tivoli Storage Manager for Virtual Environments 326 increasing the size 672, 676
turning off 309 log files, alternative locations 682
virtual volumes, server-to-server managing 643
data deduplication 308 monitoring 671
data format move to another directory 680
NATIVE 238 moving 608
data format for storage pool 206, 209, 238 protecting 898
definition 245 relocating on a server 679
operation restrictions 249 reorganization, table and index
data movement, querying 404 archive log, space requirements 668
data mover index 678
defining 175, 224 overview 676
description 45 restrictions 677
managing 208 table 678
NAS file server 45 restoring 900, 903, 905
data protection with WORM media 133 most current state 875, 901, 903
data retention protection 514 point in time 875, 901
data retention using Centera setting user data limit 650
overview 42 SQL queries 644
unsupported functions 249 statistics collection 644
data shredding transactions 643, 685
BACKUP STGPOOL command 541 database backup 872
COPY ACTIVEDATA command 541 database backup and recovery
DEFINE STGPOOL command 541 general strategy 732
DELETE FILESPACE, command 541 methods 732
DELETE VOLUME, command 541 providing 732
description 539 database manager 654, 655
enforcing 541 database space 654

1146 IBM Tivoli Storage Manager for Windows: Administrator's Guide

database, IBM Tivoli Storage Manager DELETE SCRIPT command 636
description of 28 DELETE SERVER command 701
ensuring integrity of 31 DELETE SERVERGROUP command 731
querying using SQL 794 DELETE STGPOOL command 413
date and time, changing on the server 609 DELETE SUBSCRIBER command 724
day of the week parameter 622 DELETE SUBSCRIPTION command 720
DB2 tools, use of 643 DELETE VOLHISTORY command 615
DB2 Universal Database DELETE VOLUME command 415, 416
Enterprise Extended Edition 441 deleting
deactivating policy 478 cached files on disk 401
deduplicate-identification processes 311 empty volume 415, 615
DEDUPTIER2FILESIZE 317, 318 file spaces 463
DEDUPTIER3FILESIZE 317, 318 files 414, 512
default management class scratch volume 254, 615
assigning for a policy set 510 storage volume 415
binding files to 491 subfile 554
description of 482 volume history information 615
purpose 487 volume with residual data 416
recommendation for using 489 deletion hold 515
default policy 477 deployment
default profile 703, 710, 717 command-line interface 434, 435
default scheduling mode importing 436
modifying 576 schedule 437
DEFINE ASSOCIATION command 562 verifying 438
DEFINE BACKUPSET command 548 descriptions, for archive packages 555, 556
DEFINE CLIENTACTION command 583 DESTINATION parameter (storage pool) 477, 502
DEFINE CLIENTOPT command 465 destroyed volume access mode 259, 911
DEFINE CLOPTSET command 464 determining
DEFINE COPYGROUP command 502, 507, 508, 510 cause of ANR9999D messages 808
DEFINE DEVCLASS command 179, 186 the time interval for volume check in 183
define drive 118 device 76
DEFINE DRIVE command 174 adding using commands 91
DEFINE GRPMEMBER command 729 attaching to server 91, 219
DEFINE LIBRARY command 173 configuring removable media 89
DEFINE MACHNODEASSOCIATION command 1013 defining to IBM Tivoli Storage Manager 92
DEFINE MGMTCLASS command 501 multiple types in a library 52
define path 118 selecting device drivers 79
DEFINE POLICYSET command 500 supported by IBM Tivoli Storage Manager 36
DEFINE PROFASSOCIATION command 708 troubleshooting 123
DEFINE PROXYNODE command 442 unsupported 88
DEFINE RECMEDMACHASSOCIATION command 1017 Windows device drivers 81
DEFINE RECOVERYMEDIA command 1017 device class
DEFINE SCHEDULE command 27, 621 3590 177, 179
DEFINE SCRIPT command 627 3592 179
DEFINE SERVER command 694, 726, 732 4MM 177, 179
DEFINE STGPOOL command 249, 250, 261, 262 8MM 177, 179
DEFINE SUBSCRIPTION command 719 amount of space used 398
DEFINE VIRTUALFSMAPPING command 236 CENTERA 42
DEFINE VOLUME command 256 defining 177
defining description of 19
client nodes 451 DISK 177
delaying migration for files 278 DLT 177, 179
delaying reuse of volumes 378 ECARTRIDGE 179
DELETE ASSOCIATION command 573 FILE 177
DELETE BACKUPSET command 552 for devices using native Windows device driver 88
DELETE CLIENTOPT command 467 FORMAT parameter 180
DELETE COPYGROUP command 532 GENERICTAPE 88, 177, 179
DELETE DOMAIN command 532 LTO 191
DELETE EVENT command 575, 626 overview 94
DELETE FILESPACE command 463 REMOVABLEFILE 186
DELETE GRPMEMBER command 731 requesting information about 198
DELETE KEYRING command 837 selecting for import and export 754
DELETE MGMTCLASS command 532 sequential 179
DELETE POLICYSET command 532 SERVER 177, 179, 734
DELETE PROFASSOCIATION command 712 StorageTek devices 179, 195
DELETE PROFILE command 714 tape 179, 186
DELETE SCHEDULE command 570, 625 Ultrium, LTO 179

Index 1147
device class (continued) disaster recovery (continued)
updating 179, 186 providing 732
VOLSAFE 195 server
WORM 177, 179 disaster recovery 1034
device classes server recovery 1034
database backups 871 disaster recovery manager
device configuration file 879, 906 awk script 1048
unrecognized by IBM Tivoli Storage Manager 88 client recovery information 1005
device diagnostics 123 creating a disaster recovery plan 1018
device driver customizing 1006
for manual tape devices 74 displaying a disaster recovery plan 1022
IBM Tivoli Storage Manager, installing 74 expiring a disaster recovery plan 1022
installing 73 features 1005
overview 79 moving volumes back on-site 1028
requirements 73 project plan, checklist 1043
Windows 81 querying a disaster recovery plan 1022
device sharing 60 recovery media 1017
device special file names 76 saving machine characteristics 1013
device support 15 stanzas, recovery instructions 1011
device type storing a disaster recovery plan 1018
3590 179 disaster, protecting against 30
4MM 177, 179 disk device class, defined 177
8MM 177, 179 disk devices
CENTERA 42 random access 65
DISK 177 sequential access 65
DLT 177, 179 disk space
ECARTRIDGE 179 for the database and recovery log 650
FILE 177 freeing 71
GENERICTAPE 177, 179 disk storage
LTO 179, 191 and file-system requirements 63
multiple in a single library 52 random access (DISK) 40
REMOVABLEFILE 177 sequential access (FILE) 41
SERVER 177, 179, 734, 736 disk storage pool
VOLSAFE 195 cache, use of 284
WORM 177, 179 deleting cached files from 401
device, storage estimating space 380
adding, overview 86 estimating space for archived files 381
alias 75 estimating space for backed up files 380
automated library device 97 migration threshold 273
device information 123 setting up 63
device name 75 disk subsystems
disk 63 requirements 63
removable media device 89, 186 disk support 63
replacing tape drive 159 disk systems support 63
required IBM Tivoli Storage Manager definitions 58 dismounting
supported devices 36 idle volumes 145
devices volumes from REMOVABLEFILE devices 146
defining 173 display information 444
diagnosing ANR9999D messages 808 specific client node 444
diagnostics, for device 123 DISPLAY OBJNAME command 452
differential backup distribute workloads
compared to incremental 13 to reduce backup and restore time 441
of image, description 10, 51 DLT WORM media 133
direct-to-tape, policy for 522 domain, policy
directories active-data pools, specifying 498
deleting from archive packages 556 assigning client node 512
directory-level backup 235 changing 478
preventing archive of 557 creating 500
storage usage for archive packages 555 deleting 532
disability 1109 description of 482
DISABLE EVENTS command 808 distributing via profile 528, 705
DISABLE SESSIONS command 471 for NAS file server node 212
disaster recovery querying 531
auditing storage pool volumes 897 updating 496, 498
general strategy 732 drive
methods 31, 732 cleaning 164
node replication as a method for 988 defining 174

1148 IBM Tivoli Storage Manager for Windows: Administrator's Guide

drive (continued) enterprise configuration (continued)
defining path for 176 profile for 703
deleting 168 scenario 690, 702
detecting changes on a SAN 125 subscription to 705
element address 175, 177 enterprise event logging 692, 820
querying 152 environment variable, accounting 803
serial number 174 error analysis 793
server retries for acquiring 62 error checking for drive cleaning 168
simultaneous-write function, requirements for 353 error reporting for ANR9999D messages 808
updating 153 error reports for volumes 386
updating to use for NDMP operations 208 establishing server-to-server communications
drive configuration 117 enterprise configuration 692
DRIVEACQUIRERETRY server option 62 enterprise event logging 692
DRIVEENCRYPTION parameter virtual volumes 700
3592 device class 185 estimate network bandwidth 941
ECARTRIDGE device class 196, 197 estimate replication 941
LTO device class 193 estimated capacity for storage pools 382
driver, device estimated capacity for tape volumes 386
for manual tape devices 74 event logging 807, 814
IBM Tivoli Storage Manager, installing 74 event record (for a schedule)
installing 73 deleting 575, 626
requirements 73 description of 563, 573
Windows 81 managing 625
drivers querying 625
for IBM devices 78 removing from the database 575, 626
drives 117 setting retention period 575, 626
defining in the library 93 event server 820
[Link] file 425, 464, 559 example
[Link] 803 assigning a default management class 510
DSMADMC command 750, 764, 771 register three client nodes with CLI 428
DSMC loop session 467 validating and activating a policy set 512
DSMMAXSG utility 126 expiration 71
[Link] file 575 expiration date, setting 623
DSMSERV DISPLAY DBSPACE command 671 expiration processing 27
DSMSERV DISPLAY LOG command 671, 678 description 887
DSMSERV_ACCOUNTING_DIR 803 files eligible 479, 512
duplicate-identification processes 306, 311 of subfiles 479, 504, 512, 554
duplication of restored data 913 starting 512
DVD-RAM support using disaster recovery manager 514
configuring 91 EXPIRE INVENTORY command 27
defining and updating a device class 186 duration of process 513
dismounting volumes from DVD-RAM devices 146 export
dynamic serialization, description of 502, 508 administrator information 751
client node information 751
data from virtual volumes 773
E decided when 743
directly to another server 744
ECARTRIDGE device class 179
labeling tapes 746, 754
element address 174, 175
monitoring 770
embedded help 587
options to consider 745
ENABLE EVENTS command 808
planning for sequential media 754
ENABLE SESSIONS command 471
policy information 752
encoding events to UTF-8 814
PREVIEW parameter 753
encryption
previewing results 749
changing method 538
querying about a process 770
choosing a method 537
querying the activity log 772
DRIVEENCRYPTION parameter
replacing definitions before 746
3592 Generation 2 185
server data 752
ECARTRIDGE 196, 197
using scratch media 754
LTO-4 193
viewing information about a process 770
methods 154, 536
EXPORT ADMIN command 755
END EVENTLOGGING command 809
export and import data
Enterprise Administration
sequential media volumes 753
description 687
EXPORT commands 770, 771
enterprise configuration
EXPORT NODE command 756
communication setup 692
EXPORT POLICY command 757
description 688, 701
EXPORT SERVER command 753, 757
procedure for setup 702

Index 1149
exporting file, client
administrator data 755 allowing archive while changing 477
client node data 756 allowing backup while changing 477, 502
data to tape 755 archive package 555
description of 741 associating with management class 489, 490
policy data 757 damaged 913
server data 757 delaying migration of 278
subfiles 554 deleting 414
EXPQUIET server option 513 deleting from a storage pool 413
EXTEND DBSPACE command 672, 676 deleting from cache 284
EXTERNAL library type 1090 deleting when deleting a volume 414
external media management duplication when restoring 913
IBM Tivoli Storage Manager setup 149 eligible for archive 477, 492
initialization requests 1090 eligible for backup 477, 492
interface description 1085 eligible for expiration 479
overview 148 eligible for space management 496
processing during server initialization 1086 how IBM Tivoli Storage Manager stores 262
volume dismount requests 1095 on a volume, querying 388
volume mount requests 1092 server migration of 271
volume release requests 1091 files, damaged 896, 897, 913
files, unreadable 896, 897
firewall, client nodes
F client-initiated sessions 432
server-initiated sessions 433
failover 968
fixes, obtaining 1106
failover cluster 998
format for storage pool 206, 209
installing Tivoli Storage Manager 1000
definition 245
fiber tape failover 995
operation restrictions 249
file data, importing 741
formatting
file deletion option
event classes 812
setting 425
storage pool volume 69, 255
FILE device type
frequency of backup 503
active-data pools 241
FSID 462
backing up or archiving many small objects 188
full image backup, NAS node 51
benefits 41
concurrent access to FILE volumes 41
defining device class 177
deleting scratch volumes 615 G
free space in directories 400 GENERATE BACKUPSET command 544
setting up storage pool 70 GENERICTAPE device type 186
file exit 807 Global Security Kit (GSKit) 833
logging events to 810 glossary 1115
file path name 452 GRANT AUTHORITY command 845
file retrieval date 285 group backup, on the client 12
file server, network-attached storage (NAS) group, server
backup methods 213 copying 730
registering a NAS node for 223 defining 729
using NDMP operations 50, 203 deleting 731
file size, determining maximum for storage pool 245 member, deleting 731
file space moving a member 731
deleting, effect on reclamation 368 querying 730
deleting, overview 463 renaming 730
description of 451 updating description 730
merging on import 745, 759
names that do not display correctly 462
QUERY OCCUPANCY command 397
querying 451
H
HALT command 607
renaming 769
halting the server 607
Unicode enabled 462
hardware scans
viewing information about 451
VMware environment 450, 603
file space identifier (FSID) 462
held volume in a client session 467
file spaces
help 587
defining 451
HELP command 616
file-level restore
hierarchy, storage 19
managing 234
copying active backup data 241
planning 233
defining in reverse order 249, 261
file-system requirements for disk storage 63
establishing 260

1150 IBM Tivoli Storage Manager for Windows: Administrator's Guide

hierarchy, storage (continued) incomplete copy storage pool, using to restore 913
example 243 incremental backup 492
for LAN-free data movement 261 incremental backup, client
how the server stores files in 262 file eligibility for 492
next storage pool frequency, specifying 579
definition 261 full 492
deleting 414 partial 493
migration to 271, 392 progressive 13
restrictions 261 incremental replication 945
staging data on disk for tape storage 270 inheritance model for the simultaneous-write function 340
HL ADDRESS 433 initial replication 944
Host Bus Adapter (HBA), increasing maximum transfer length initial start date for schedule 622
for 126 initial start time for schedule 622
hostname initializing 1000
changing 613 installing 80
how to cause the server to accept date and time 609 installing IBM Tivoli Storage Manager 420
hub server instance user ID 606
changing 591 instant archive
configuring 593 creating on the server 543
restoring to preconfigured state 592 description of 9, 13
interface, application program
client, registering 424
I compression option 424
deletion option 424
IBM error analysis 793
description of 3
IBM Support Assistant 1105
registering to server 424
IBM Tivoli Monitoring 805
simultaneous-write function, version support for 333
IBM Tivoli Storage Manager (Tivoli Storage Manager)
interfaces to IBM Tivoli Storage Manager 17
introduction 3
Internet, searching for problem resolution 1105, 1106
server network 29
introduction to IBM Tivoli Storage Manager 3
IBM Tivoli Storage Manager device driver 78
iPad
IBM Tivoli Storage Manager MACRO command 138
monitoring the storage environment 589
IDLETIMEOUT server option 468, 470
image backup
policy for 523, 524
suggested use 8, 11 K
import keepalive, TCP
data from virtual volumes 773 enabling 211
monitoring 770 overview 210
options to consider 759 specifying connection idle time 211
PREVIEW parameter 753, 761 key database
querying about a process 770 adding certificates 836, 837
querying the activity log 772 password change 836, 837
recovering from an error 769 keyboard 1109
replace existing definitions 760 knowledge bases, searching 1105
viewing information about a process 770
IMPORT ADMIN command 758
IMPORT commands 770, 771
IMPORT NODE command 758, 767
L
label
IMPORT POLICY command 758
automatic labeling in SCSI libraries 129
IMPORT SERVER command 758, 767
sequential storage pools 255
importing
LABEL LIBVOLUME command
active-data pools 766
restrictions for VolSafe-enabled drives 195
data 758
using an automated library 101, 115
data storage definitions 763, 765
LAN-free data movement
date of creation 760, 767
configuration 122
description of 741
description 14, 49
directing messages to an output file 750, 764
storage pool hierarchy restriction 261
duplicate file spaces 766
suggested usage 10
file data 766
LDAP-authenticated password
node replication restriction 933
configuring an LDAP directory server 855
policy definitions 763
configuring the server 857
server control data 765
policy 856
subfiles 554
query admin 860
subsets of information 768
query node 860
include-exclude file
register nodes and admin IDs 858
description of 24, 488
scenarios 861
for policy environment 482, 488

Index 1151
LDAP-authenticated password (continued) logical volume on client (continued)
transport layer security 839 management class for 489
update node or admin 859 policy for 492, 523
libraries process for backup 495
defining 92 restore 481
NDMP operations 217 logical volume, raw 21
virtual tape library 116 loop session, DSMC 467
library low migration thresholds 277
ACSLS (Automated Cartridge System Library LTO Ultrium devices and media
Software) 39, 109 device class, defining and updating 191
attaching for NAS file server backup 219 encryption 154, 193, 536
auditing volume inventory 137 WORM 133, 195
automated 135 LUN
categories for volumes in IBM 3494 38, 96 using in paths 176, 177
configuration example 97
configure for more than one device type 52
defining 152, 173
defining path for 176
M
machine characteristics 1013
deleting 152
machine recovery information 1014
detecting changes to, on a SAN 125, 174
macro
external 39
commit individual commands 640
IBM 3494 38, 96
continuation characters 638
managing 150
controlling command processing 640
manual 37
running 640
mixing device types 52, 183, 191
scheduling on client 564
mode, random or sequential 74
substitution variables 639
overflow location 245
testing 641
querying 150
using 637
SCSI 38
writing commands 637
serial number 174
writing comments 638
sharing among servers 102, 119
MACRO administrative command, using 428
type 47
magnetic disk devices 40, 63
updating 151
maintenance plan
volume inventory 47, 128
modify 629
VTL 38
maintenance script
library client, shared library 48, 103, 114, 121, 905, 916
create 628
library drive usage
custom 628
determining 215
modify 27
library manager, shared library 48, 103, 113, 119, 903, 915
managed server
library sharing 54
changing the configuration manager 718, 724
license
communication setup 692
compliance 597
deleting a subscription 720
features 596
description 688
monitoring 597
managed objects 688, 717
registering 596
refreshing configuration information 721
using 595
renaming 725
limitation for script definition on administrative Web
returning managed objects to local control 722
interface 627
setting up 705
LL ADDRESS 433
subscribing to a profile 705, 717, 718, 719
Local System account 604
management class
location, volume
assigning a default 510
changing 257
associating a file with 489
overflow for storage pool 245
binding a file to 489
querying volume 387
configuration 486
LOCK ADMIN command 852
controlling user access 486
LOCK NODE command 440
copying 496, 501
LOCK PROFILE command 711, 712
default 487
log mirror 647, 678
define new 527
logical block protection
defining 501
enabling 156
deleting 532
overview 155
description of 482, 486
read/write operations 158
querying 530
storage pool management 159
rebinding a file 491
supported drives 156
updating 490, 496, 501
logical devices 70
managingserver operation 26
logical volume on client
manual drive
backup 480
attaching 73

1152 IBM Tivoli Storage Manager for Windows: Administrator's Guide

manual library device mode (continued)
adding, using commands 91 library (random or sequential) 74
managing media 142 scheduling 577
reusing media 142 modified mode, description of 503
manually configuring a device modifying schedules 570
Windows 86 monitoring
MAXSCRATCH parameter 245, 257, 410 server-to-server export 750
media monitoring the storage environment 587
checking in 130 mount
labeling for automated library devices 129 count of number of times per volume 387
mount operations 143 library 181
reusing 141 limit 181
tape rotation 55, 146 retention period 182
media label wait period 183
for devices unrecognized by IBM Tivoli Storage mount point 978
Manager 129 preemption 611
merging file spaces 745, 759 queue, server option 62
messages relationship to mount limit in a device class 181, 190, 198
determining cause of ANR9999D message 808 requirements for simultaneous-write operations 352
directing import messages to an output file 750, 764 settings for a client session 421
for drive cleaning 168 MOVE DATA command 401
getting help on 616 MOVE DRMEDIA command 1029
severe 808 MOVE NODEDATA 406
Microsoft Failover Cluster 998 moving a backup set
Microsoft Failover Cluster Manager 993 benefits of 548
MIGRATE STGPOOL command 282 to another server 548
migrating a file 481, 496 moving data 463
migration, client from offsite volume in a copy storage pool 402
automatic, for HSM client monitoring the movement of 405
demand 481 procedure 403
files, eligible 496 requesting processing information 404
threshold 481 to another storage pool 402
premigration for HSM client 481 to other volumes in same storage pool 401
reconciliation 482 multipath I/O 79
selective, for HSM client 481 multiple
stub file on HSM client 481 copy storage pools, restoring from 913
migration, server managing IBM Tivoli Storage Manager servers 29
canceling the server process 393 managing Tivoli Storage Manager servers 687
controlling by file age 278 multiple server instances 606
controlling duration 282 multiple servers 725
controlling start of, server 276 completing tasks 725
copy storage pool, role of 283 multistreaming, concurrent for database backups and
defining threshold for disk storage pool 276 restores 872, 900
defining threshold for tape storage pool 279
delaying by file age 278
description, server process 273
minimizing access time to migrated files 279
N
NAS file server, NDMP operations
monitoring thresholds for storage pools 392
backing up a NAS file server 230
multiple concurrent processes
backing up a NAS file server to native pools 231, 232
random access storage pool 245, 273
configuration checklist 211
sequential access storage pool 245, 282
data format 206
problems, diagnosing and fixing 271
data mover, description 45, 175
providing additional space for server process 394
defining a data mover 175, 224
starting manually 282
defining a storage pool 217
starting server process 270, 276
defining paths to drives
threshold for a storage pool
drives attached only to file server 225
random access 273
drives attached to file server and Tivoli Storage
sequential access 279, 280
Manager server 224
mirroring 877
obtaining names for devices attached to file server 226
description of 31
defining paths to libraries 227
MIRRORLOGDIRECTORY server option 684
differential image backup, description 51
mixed device types in a library 52, 183, 191
full image backup, description 51
mobile client support 552
interfaces used with 205
mobile device
managing NAS nodes 207
monitoring the storage environment 589
path, description 45, 176
mode
planning 215
client backup 503
policy configuration 212, 524

Index 1153
NAS file server, NDMP operations (continued) network-attached storage (NAS) file server
registering a NAS node 223, 423 backup methods 213
requirements for set up 203 registering a NAS node for 223
restoring a NAS file server 230 using NDMP operations 50, 203
scheduling a backup 228 new tape drive 159
storage pools for NDMP operations 217 next storage pool
NAS node definition 261
defining 223 deleting 414
deleting 208 migration to 271, 392
registering 223 node
renaming 207 registering 448, 478
NAS nodes 238 node privilege class
NATIVE data format 206 description of 446
NDMP granting 447
operations 206 node replication 945, 978
NDMP backup data 238 configuration
NDMP operations 238 effectiveness, measuring 986
NDMP operations for Celerra data movers 238 removing 989
NDMP operations for NAS file servers setting up the default 948, 950
backing up a NAS file server 230 source and target replication servers 921
backing up a NAS file server to native pools 231, 232 validating 970
configuration checklist 211 database requirements 940
data format 206 database restore, replicating after 983
data mover, description 45, 175 deduplicated data 931, 987
defining a data mover 175, 224 disabling and enabling
defining a storage pool 217 rules 982
defining paths to drives disabling and enabling replication
drives attached only to file server 225 all client nodes 980, 981
drives attached to file server and Tivoli Storage data types in file spaces 978
Manager server 224 individual client nodes 979
obtaining names for devices attached to file server 226 disaster recovery
defining paths to libraries 227 recovering data from the target 988
differential image backup, description 51 store operations on the target replication server 989
full image backup, description 51 file spaces
interfaces used with 205 data types, disabling and enabling 978
managing NAS nodes 207 purging data in 983
path, description 45, 176 querying replication results 985
planning 215 rules, changing 952
policy configuration 212, 524 import-and-export operations
prevent closing of inactive connections converting from 961
enabling TCP keepalive 211 restriction 933
overview 210 migration by HSM for Windows client 933
specifying connection idle time 211 mode, replication 930
registering a NAS node 223, 423 nodes
requirements for set up 203 adding for replication 962
restoring a NAS file server 230 all client nodes, disabling and enabling replication 980,
scheduling a backup 228 981
storage pools for NDMP operations 217 attributes updated during replication 932
NetApp file server individual client nodes, disabling and enabling
data format for backup 206 replication 979
international characters 234 removing from replication 963
NETAPPDUMP data format 206, 217 resetting the replication mode 964
NetView 807 overview 919
Network Appliance file server planning 939
backup methods 213 policy management 922
requirements 203 previewing results 970
storage pool for backup 249 process information
tape device for backup 203 activity log 986
using NDMP operations 50, 203 file spaces 985
network attached storage record retention 988
virtual file spaces 228 summary records 986
network bandwidth 941 records
network of IBM Tivoli Storage Manager servers 29 displaying 985
network of Tivoli Storage Manager servers 687 retaining 988
network-attached nodes replicating
comparing to local nodes 427 canceling processes 985
data by file space 972

1154 IBM Tivoli Storage Manager for Windows: Administrator's Guide

node replication (continued) node, client (continued)
replicating (continued) target 442
data by priority 974 unlocking 440
data by type 973 updating 434
scheduling or starting manually 971 viewing information about 443
throughput, managing 976 nodes
restoring, retrieving, and recalling data from the moving nodes from schedule 572
target 988 overview of client and server 419
results, previewing 970 NOPREEMPT server option 611
retention protection, archive 933 NORETRIEVEDATE server option 285
rules NOTIFY SUBSCRIBERS command 712
attributes 924 NT event log receiver 820
definitions 923 number of times mounted, definition 387
disabling and enabling 982
file spaces 952
hierarchy 924
nodes, individual 954
O
occupancy, querying 397
processing example 925
off-site volume
server 956
limiting the number to be reclaimed 377
Secure Sockets Layer (SSL) 968, 969
offsite recovery media
servers
specify defaults 1009
communications, setting up 948
offsite recovery media (for DRM)
configurations 921
volumes
source, adding 965
moving back on-site 1028
target 965, 989
sending offsite 1026
settings, displaying
states 1024
file spaces 985
offsite volume access mode 260
nodes 985
offsite volumes
rules 985
limiting the number to be reclaimed 245
SSL (Secure Sockets Layer) 968, 969
moving data in a copy storage pool 402
state, replication 927
one-drive library, volume reclamation 245, 372
task tips
open registration
monitoring processes 938
description 421
nodes, adding and removing 936
enabling 428
previewing results 937
process 422
processing, managing 937
setting 421
rules, changing replication 935
operations available to client 10
servers, managing 936
Operations Center 587
validating a configuration 937
configuring 591, 593
verifying results 938
opening 589
node replication method 943, 944
overview xvii, 587
node replication synchronization 943
port number 589
node replication tiering 1076
restoring to preconfigured state 592
node, client
URL 589
adding 419
web server 593
agent 442
operator privilege class
amount of space used 397
reducing 850
creating backup sets for 544
revoking 850
file spaces, QUERY OCCUPANCY command 397
optical device
finding tapes used by 391
reclamation for media 373
immediate processing 583
option set, client
importing 767
adding client options to 465
locking 440
assigning clients to 466
managing registration 420, 431, 595
copying 467
options file 425
creating 465
performing operations for 535, 569, 576
deleting 467
privilege class for scheduling operations for 561
deleting an option from 467
proxy node relationships 441
for NAS node 213
querying 443
requesting information about 467
reducing archive packages for 556
updating description for 467
registering 424
option, server
removing 440
3494SHARED 62
renaming 439
ACSLS options 62
scheduling operations for 560
ASSISTVCRRECOVERY 62
setting password authentication 865
AUDITSTORAGEstorage audit 597
setting scheduling mode 578
changing with SETOPT command 615
setting up subfile backups 553
CHECKTAPEPOS 62

Index 1155
option, server (continued) performance (continued)
COMMTIMEOUTcommunication timeout 468, 470 volume frequently used, improve with longer mount
DRIVEACQUIRERETRY 62 retention 182
EXPINTERVAL 513 period, specifying for an incremental backup 579
EXPQUIET 513 plan
IDLETIMEOUTidle timeout 468, 470, 790 Disaster Recovery Manager 1076
NOPREEMPT 62, 611 DRM 1076
NORETRIEVEDATEfile retrieval date 285 planning, capacity
overview 27 database space requirements
QUERYAUTH 845 estimates based on number of files 652
REQSYSAUTHOUTFILE 845 estimates based storage pool capacity 654
RESOURCETIMEOUT 62 starting size 652
RESTOREINTERVALrestore interval 471, 479, 513 recovery log space requirements
SEARCHMPQUEUE 62 active and archive logs 656
THROUGHPUTDATATHRESHOLD 470 active log mirror 668
THROUGHPUTTIMETHRESHOLD 470 archive failover log 669
TXNGROUPMAXmaximum transaction group size 262 archive log space for database reorganization 668
options file, client 425 point-in-time restore
options, querying enable for clients 9, 527
VIRTUALMOUNTPOINT client option 452 policy
overflow location 245 default 5, 477
overview deleting 531
Operations Center xvii description of 482
overviewOperations Center 587 distributing with enterprise management 528
owner authority, client 446, 448 effect of changing 510, 511
for application clients 523
for clients using SAN devices 525
P for direct-to-tape backup 522
for logical volume backups 523
PARALLEL command 630
for NAS file server node 212
Passport Advantage 1107
for point-in-time restore 527
password
for server as client 527
changing the key database 836, 837
for space management 477, 496, 501
LDAP-authenticated policy 856
importing 763
resetting an administrative 850
managing 475
setting authentication for a client 865
operations controlled by 480
setting expiration 860
planning 476
setting invalid limit 863
querying 529
setting minimum length 864
policy domain
update for scheduling operations 426
active-data pools, specifying 498
using with unified logon 865
assigning client node 512
path
changing 478
defining 176
creating 500
description 45, 215
define 526
paths
deleting 532
defining 173
description of 482
pending, volume state 387
distributing via profile 528, 705
per product ID (PID) 601
for NAS file server node 212
PERFORM LIBACTION 118
querying 531
performance
updating 496, 498
cache, considerations for using 71, 284
policy privilege class
clients, optimizing restore 241
description 844, 845
concurrent client/server operation considerations 579
granting 849
data protection, increasing with simultaneous-write
revoking 849, 850
function 331
policy set
data validation for nodes 536
activating 511
data validation for storage pools 893
changing, via the active policy set 478
file system effects on 69, 255
copying 478, 496, 500
random-access disk (DISK) 63
defining 500
FILE-type volumes, backing up many small objects to 188
deleting 532
fragmentation, private FILE volumes for reducing disk 41
description of 482
migration, multiple concurrent processes 245, 282
querying 530
mobile client 552
updating 500
reclamation, multiple concurrent processes
validating 510, 512
copy storage pools 375
pool, storage
primary sequential access storage pools 245, 371
3592, special considerations for 183
storage pool backup, reducing time required for 331
active-data pool 241
storage pool volume 279

1156 IBM Tivoli Storage Manager for Windows: Administrator's Guide

pool, storage (continued) privilege class, administrator
amount of space used 398 description 845
auditing a volume 887 granting authority 845
comparing primary and copy types 412 reducing 850
copy 241 revoking all 850
creating a hierarchy 260 privilege class, policy
data format 206, 245, 249 description 844, 845
defining 245 granting 849
defining a copy storage pool 410 revoking 849, 850
defining for disk, example 249, 261 privilege class, storage
defining for NDMP operations 217 description 845
defining for tape, example 249, 261 granting 849
deleting 413 reducing 849
description of 240 revoking 850
destination in copy group 502, 508 problem determination
determining access mode 245, 410 describing problem for IBM Software Support 1107
determining maximum file size 245 determining business impact for IBM Software
determining whether to use collocation 245, 357, 410 Support 1107
disk 21 migration 271
duplicate, using to restore 913 submitting a problem to IBM Software 1108
enabling cache for disk 245, 284 process
estimating space for archived files on disk 381 background 609
estimating space for backed up files on disk 380 canceling 610
estimating space for disk 380 drive clean error checking 168
estimating space for sequential 381 expiration 887
estimating space in multiple 260 number for migration 245, 273
incomplete, using to restore 913 reclamation 367, 376
increasing sizes 21 processor value unit 598
LTO Ultrium, special considerations for 191 Product ID (PID) 598
managing 239 profile
monitoring 382 associating configuration information with 707
moving files 401 changing 707, 711, 712
moving files between 402 default 710, 717
multiple, using to restore 913 defining 707, 708
next storage pool deleting 712, 714
definition 261 description 707
deleting 414 getting information about 714
migration to 271, 392 locking 711
overview 43 problems with synchronization 723
policy use 502, 508 unlocking 711
primary 240 progressive incremental backup 13
querying 382 protecting your data 30, 133
renaming 409 active-data pools 241
search-and-selection order for active files 243 data deduplication 307
simultaneous-write function 331 simultaneous-write function 331
updating 245 protection options
updating for disk, example 250, 262 client 8
using cache on disk 245, 284 server 14, 31
validation of data 890 proxy node relationships 442
viewing information about 382 publications
port number download xv
Operations Center 589 PVU
portable media estimation 598
description of 6, 8, 543 VMware environment 450, 603
restoring from 548
preemption
mount point 611
volume access 612
Q
query
prefix, for recovery instructions 1006
authority 845
prefix, for recovery plan file 1006
for general information 385
prefix, server 727
policy objects 529
premigration 481
QUERY ACTLOG command 772, 800
PREPARE command 1018
QUERY ADMIN command 848
PREVIEW parameter 753, 761
query association output 572
primary volumes unavailable for restore 913
QUERY BACKUPSETCONTENTS command 551
private category, 349X library 38, 96
QUERY CONTENT command 388
private volumes 46
QUERY COPYGROUP command 529, 765

Index 1157
QUERY DB command 671 reclamation (continued)
QUERY DBSPACE command 671 setting a threshold for sequential storage pool 245, 367,
QUERY DEVCLASS command 754 411
QUERY DIRSPACE 400 starting reclamation manually 371
QUERY DOMAIN command 531 storage pool for 245
QUERY DRMSTATUS command 1006 virtual volumes 374
QUERY ENABLED command 822 with single drive 372
QUERY EVENT command 573, 625 RECONCILE VOLUMES command 739
QUERY FILESPACE command 451 reconstructing aggregates 368, 375, 405
QUERY LICENSE command 597 recovery instructions file 1057
QUERY MEDIA command 138 recovery log 643, 647
QUERY MGMTCLASS command 530 active log 646, 647, 649
QUERY NODE command 443 active log mirror 648
QUERY OCCUPANCY command 397, 398, 399 alternative file locations
QUERY OPTION command 792 overview 682
QUERY POLICYSET command 530 specifying with ARCHFAILOVERLOGDIRECTORY
QUERY PROCESS command 610, 770, 791 option or parameter 683
identification numbers of migration processes 393 specifying with ARCHLOGDIRECTORY
information about data movement process 404 parameter 683
QUERY PVUESTIMATE 601 specifying with RECOVERYDIR parameter 684
QUERY RESTORE command 472 archive failover log 646, 649
QUERY RPFCONTENT command 1022 archive log 646, 648, 649
QUERY RPFILE command 1022 description of 28, 643
QUERY SCHEDULE command 563 increasing the size 678
QUERY SCRIPT command 635 log mirror 646, 648, 649
QUERY SERVERGROUP command 730 managing 643
QUERY SESSION command 468, 790 monitoring 671
QUERY SHREDSTATUS command 540 out of space 678
QUERY STATUS command 792 recovery logs
QUERY STGPOOL command 382, 392, 395 move to another directory 680
QUERY SUBSCRIPTION command 719 relocating on a server 679
QUERY SYSTEM command 793 recovery plan file
QUERY VOLHISTORY command 615 break out stanzas 1048
QUERY VOLUME command 384, 385, 405 using VBScript procedure 1048
QUERYAUTH server option 845 creating 1018
example 1054
prefix 1006
R stanzas 1051
recovery, disaster
random mode for libraries 74
auditing storage pool volumes 897
randomize, description of 580
general strategy 732
raw logical volume 21
media 1017
read-only access mode 259
methods 732
read/write access mode 259
providing 732
rebinding
redistributing data for database directories 674
description of 491
REGISTER ADMIN command 847
file to a management class 491
REGISTER LICENSE command 596
recalling a file
REGISTER NODE command 448
selective 482
registering
transparent 482
client option sets 423
receiver 807
workstation 424
RECLAIM STGPOOL command 371
registration
reclaiming storage space 674
description of 421
reclamation 377
licensing for a client node 595
active-data pools 374
licensing for an administrator 595
aggregate reconstruction 368
managing client node 420, 431
controlling duration 371
setting for a client node 421
delayed start of process 368
source server 424
delaying reuse of volumes 378, 887
relationships
description of 20
among clients, storage, and policy 484
effects of collocation 378
remote access to clients 445
effects of DELETE FILESPACE 368
removable file system device
multiple concurrent processes
example of setting up 89
copy storage pools 375
REMOVABLEFILE device type, defining and
primary sequential access storage pools 245, 371
updating 186
off-site volume
support for 89, 186
controlling when reclamation occurs 376
removable media 40

1158 IBM Tivoli Storage Manager for Windows: Administrator's Guide

removable media device, adding 89 retention grace period (continued)
REMOVE ADMIN command 851 description of backup 498
REMOVE NODE command 440 for backup sets 546
remove replication server 967 using archive 498
REMOVE REPLSERVER 967 using backup 498
RENAME ADMIN command 850 RETEXTRA parameter 477, 504
RENAME FILESPACE command 769 RETONLY parameter 477, 504
RENAME NODE command 439 retrieval date for files 285
RENAME SCRIPT command 636 retrieval from archive
RENAME SERVERGROUP command 730 archive package 555
RENAME STGPOOL command 409 file 481
renamed file spaces 461 reuse of sequential volumes
renaming delaying 378, 887
administrator ID 850 volume pending state 387
NAS node 207 reusing media 141
storage pool 409 roll-forward recovery 647
renaming the host 613 ROLLBACK command 641
renaming the server 613 routing commands 689
replacing tape drive 159 routing commands to servers 726
replicate data 941 RUN command 636
replicate NAS node 238
replication 941, 978
node replication 919
recovering an LDAP server 914
S
SAN (storage area network)
replication method 944, 945
client access to devices 49
replication performance 977, 978
device changes, detecting 125
replication target server 966
LAN-free data movement 49
replication time 942
NDMP operations 50, 203
replication workload 978
policy for clients using LAN-free data movement 525
reporting and monitoring 805
sharing a library among servers 47, 102, 119
reporting ANR9999D messages 808
storage agent role 49
REQSYSAUTHOUTFILE server option 845
scale capacity 185
requirements for disk subsystems 63
scenarios
resetting
policy configuration 522
administrative password 850
schedule
user password expiration 860
administrative command 619
RESOURCETIMEOUT server option 62
associating client node with 562
restartable export 747
checking the log file 575
restartable restore session, client
coordinating 576
canceling 472
copying 570, 624
interrupting, active 473
database backups 876
requesting information about 472
day of the week 622
restore
defining 561, 613, 621
increasing throughput using DSMMAXSG utility 126
deleting 570, 625
selecting individual files 548
description of 559
RESTORE DB command 608
expiration date 623
restore operations 872
failed, querying 563, 574
RESTORE STGPOOL command 916
for NAS file server backup 228
restore to point-in-time, enabling for clients 527
frequency of service 622
RESTOREINTERVAL server optionrestore interval for
initial start date 622
restartable restore sessions 471, 479, 513
initial time 622
restoring
maintenance 629
clients, optimizing restore 241
mode, setting 577
file 480
priority 623
storage pools with incomplete volumes 913
querying 563
restoring image data
results of 573, 625
from backup sets 548
server administrative command 620
restriction
startup window 579, 622
ASCII characters in administrative Web interface 627
type of action 623
serial number detection 126
uncertain status 574, 626
retain extra versions, description of 477, 504
updating 621
retain only version, description of 477, 504
viewing information about 563
retaining data using Centera
schedule event
overview 42
managing 573, 625
unsupported functions 249
querying 573, 625
retention grace period
viewing information about 573, 625
description of archive 498
schedule replication 945

Index 1159
scheduled operations, setting the maximum 579 security (continued)
scheduler workload, controlling 579 data encryption (continued)
scheduling Oracle StorageTek T10000B 154, 536
administrative commands 27 Oracle StorageTek T10000C 154, 536
scheduling mode data encryption, 3592 Generation 2 and later 154
client-polling 577 features, overview 23
overview of 577 for the server 831
selecting 577 locking and unlocking administrators 852
server-prompted 577 locking and unlocking nodes 440
setting on a client node 578 managing access 831, 844
setting on the server 577 password expiration for nodes 860
scratch category, 349X library 38, 96 privilege class authority for administrators 845
scratch volume Secure Sockets Layer (SSL) for node replication 968, 969
deleting 254, 615 server options 845
description 46 security, replicating node data 919
FILE volumes 72 SELECT command 794
number allowed in a storage pool 245, 410 customizing queries 796
using in storage pools 257 selective backup 480, 494
script selective recall 482
maintenance 629 sending commands to servers 726
script, scheduling on client 564 sequence number 465, 467
script, server sequential mode for libraries 74
continuation characters 631 sequential storage pool
copying 634 auditing a single volume in 895
defining 627 auditing multiple volumes in 894
deleting 636 collocation 364
EXIT statement 632 criteria for migrating 281
GOTO statement 633 estimating space 381
IF clause 632 migration threshold 279
querying 635 reclamation 367
renaming 636 SERIAL command 630
routing commands in 727 serial number
running 636 automatic detection by the server 125, 174, 177
running commands in parallel 630 for a drive 175
running commands serially 630 for a library 174, 177
substitution variables 631 serialization parameter 477, 502, 508
updating 633, 634 server
used with SNMP 815 backing up subfiles on 552
Web browser, restricted to ASCII entry 627 canceling process 610
scripts 797 changing the date and time 609
SCSI deleting 700
automatic labeling of volumes 129 description of 3
library with different tape technologies 183 disabling access 471
SCSI library disaster recovery 31
connect to NAS file server 220 enabling access 471
connecting to the server 220 halting 607
SCSI tape library importing subfiles from 554
setting up for NDMP operations 217 instances
SEARCHMPQUEUE server option 62 multiple on single system 606
secure sockets layer 838 owner ID 606
configuration 838 maintaining, overview 17
Secure Sockets Layer 832 managing multiple 29
Secure Sockets Layer (SSL) managing operations 595
certificate managing processes 609
adding CA-signed 837 messages 808
adding to key database 836 monitoring 689
communication using 831 multiple instances 606
digital certificate file protection 881 network of IBM Tivoli Storage Manager 29
Global Security Kit 833 network of Tivoli Storage Manager servers 687
security options
client access, controlling 447 adding or updating 615
data encryption prefix 727
3592 generation 2 536 protecting 31
3592 Generation 2 185 querying about processes 610, 791
3592 generation 3 536 querying options 792
ECARTRIDGE 196, 197 querying status 792
IBM LTO Generation 4 154, 193, 536 running multiple servers 606

1160 IBM Tivoli Storage Manager for Windows: Administrator's Guide

server (continued) server script (continued)
setting the server name 613 used with SNMP 815
starting 604 Web browser, restricted to ASCII entry 627
starting as a service 604 server scripts 626
stopping 607 server session
updating 700 states 468
viewing information about 792 server storage
viewing information about processes 610, 791 active-data pools 241
server console client files, process for storing 5
logging events to 809 concepts overview 15
server console, description of 843 considering user needs for recovering 59
SERVER device type 241, 732 deleting files from 414
server group evaluating 60
copying 730 example 243
defining 729 managing 18
deleting 731 monitoring 887
member, deleting 731 planning 60
moving a member 731 server options affecting 62
querying 730 tailoring definitions 765
renaming 730 using another IBM Tivoli Storage Manager server 732
updating description 730 using devices not recognized by IBM Tivoli Storage
server option Manager 88
3494SHARED 62 using disk devices 63
ACSLS options 62 using removable media devices 89
ACTIVELOGDIRECTORY 678, 681 using the storage hierarchy 270
ACTIVELOGSIZE 678 server-free data movement
ASSISTVCRRECOVERY 62 suggested usage 10
AUDITSTORAGEstorage audit 597 server-prompted scheduling 577
changing with SETOPT command 615 server-to-server communications, establishing
CHECKTAPEPOS 62 enterprise configuration 692
COMMTIMEOUTcommunication timeout 468, 470 enterprise event logging 692
DRIVEACQUIRERETRY 62 virtual volumes 700
EXPINTERVAL 513 server-to-server virtual volumes
EXPQUIET 513 deduplication 732
IDLETIMEOUTidle timeout 468, 470, 790 reclaiming 374
NOPREEMPT 62, 611 using to store data 732
NORETRIEVEDATEfile retrieval date 285 session
overview 27 canceling 469
QUERYAUTH 845 negative number 472
REQSYSAUTHOUTFILE 845 server-initiated 433
RESOURCETIMEOUT 62 setting the maximum percentage for scheduled
RESTOREINTERVALrestore interval 471, 479, 513 operations 579
SEARCHMPQUEUE 62 session, client
THROUGHPUTDATATHRESHOLD 470 canceling 469
THROUGHPUTTIMETHRESHOLD 470 DSMC loop 467
TXNGROUPMAXmaximum transaction group size 262 held volume 467
server options 614 managing 467
ARCHFAILOVERLOGDIRECTORY 684 querying 468, 790
MIRRORLOGDIRECTORY 684 viewing information about 468, 790
TECUTF8EVENT 814 sessions, maximum number for scheduled operations 976
server options file 880 SET ACCOUNTING command 803
server script SET ACTLOGRETENTION command 801
continuation characters 631 SET AUTHENTICATION command 865
copying 634 SET CLIENTACTDURATION command 583
defining 627 SET CONFIGMANAGER command 703, 706
deleting 636 SET CONFIGREFRESH command 719
EXIT statement 632 SET CONTEXTMESSAGING command 808
GOTO statement 633 SET CROSSDEFINE command 695, 698
IF clause 632 SET DBREPORTMODE command 671
querying 635 SET DRMCHECKLABEL command 1009
renaming 636 SET DRMCOPYSTGPOOL command 1006
routing commands in 727 SET DRMCOURIERNAME command 1009
running 636 SET DRMDBBACKUPEXPIREDAYS command 1009
running commands in parallel 630 SET DRMFILEPROCESS command 1009
running commands serially 630 SET DRMINSTPREFIX command 1006
substitution variables 631 SET DRMNOTMOUNTABLE command 1009
updating 633, 634 SET DRMPLANPOSTFIX command 1006

Index 1161
SET DRMPLANPREFIX command 1006 SnapMirror to Tape 237
SET DRMPRIMSTGPOOL command 1006 snapshot, using in backup 9, 11, 876
SET DRMRPFEXPIREDAYS 1022 using in directory-level backups 236
SET DRMVAULTNAME command 1009 SNMP
SET EVENTRETENTION command 575, 626 agent 815
SET INVALIDPWLIMIT command 863 communications 815
SET LICENSEAUDITPERIOD command 597 configuring 818
SET MAXCMDRETRIES command 582 enabled as a receiver 807, 815
SET MAXSCHEDSESSIONS command 579 heartbeat monitor 807, 815
SET PASSEXP command 860 manager 815
SET QUERYSCHEDPERIOD command 581 subagent 815
SET RANDOMIZE command 580 software support
SET REGISTRATION command 421 describing problem for IBM Software Support 1107
SET RETRYPERIOD command 582 determining business impact for IBM Software
SET SCHEDMODES command 577 Support 1107
SET SERVERHLADDRESS command 695, 698 submitting a problem 1108
SET SERVERLLADDRESS command 695, 698 Software Support
SET SERVERNAME command 613, 694, 695, 698 contacting 1106
SET SERVERPASSWORD 694, 695, 698 Sony WORM media (AIT50 and AIT100) 133
SET SUBFILE 552 source server 734
SET SUMMARYRETENTION 798 space
set up storage agent 838 directories associated with FILE-type device classes 400
SETOPT command 615 space requirements 940
setting space-managed file 481
clients to use subfile backup 553 special file names 76
compression 422 spoke servers
library mode 74 removing 591
password 860 restoring to preconfigured state 592
time interval for checking in volumes 183 SQL 794
setting data deduplication options 318 SQL activity summary table 798
settings 587 SQL SELECT * FROM PVUESTIMATE_DETAILS 601
shared access, nodes 442 ssl 832, 838
shared dynamic serialization, description of 502, 508 configuration 838
shared file system 70 SSL (Secure Sockets Layer)
shared library 102, 119 certificate
shared static serialization, description of 502, 508 adding CA-signed 837
SHRED DATA command 540 adding to key database 836
shredding communication using 831
BACKUP STGPOOL command 541 digital certificate file protection 881
COPY ACTIVEDATA command 541 SSL communication 832
DEFINE STGPOOL command 541 SSLTCPADMINPORT
DELETE FILESPACE, command 541 server option 835
DELETE VOLUME, command 541 SSLTCPPORT
deleting empty volumes 415 server option 835
deleting volumes with data 415 standard management class, copying 501
description 539 standard storage management policies, using 477
enforcing 541 start time, randomizing for a schedule 580
EXPIRE INVENTORY command 541 startup
EXPORT NODE command 541, 742 server 604
EXPORT SERVER command 541, 742 startup window, description of 580
GENERATE BACKUPSET command 541, 543 static serialization, description of 502, 508
MOVE DATA command 402, 541 status monitoring 587
setting up 540 status of a volume in an automated library 46, 127
UPDATE STGPOOL command 541 stopping the server 607
SHREDDING server option 540 storage agent 49
simultaneous-write operations to primary and copy storage storage area network (SAN)
pools client access to devices 49
drives 353, 354 device changes, detecting 125
inheritance model 339 LAN-free data movement 49
mount points 352 NDMP operations 50, 203
storage pools 355 policy for clients using LAN-free data movement 525
SnapLock sharing a library among servers 47, 102, 119
data protection, ensuring 521 storage agent role 49
event-based retention 520 storage devices 85, 177, 179
reclamation 517 storage hierarchy 19
retention periods 517 copying active backup data 241
WORM FILE volumes, setting up 521 defining in reverse order 249, 261

1162 IBM Tivoli Storage Manager for Windows: Administrator's Guide

storage hierarchy (continued) storage pool (continued)
establishing 260 updating for disk, example 250, 262
example 243 using cache on disk 245, 284
for LAN-free data movement 261 validation of data 890
how the server stores files in 262 viewing information about 382
next storage pool volumes, restoring 909, 910, 911
definition 261 storage privilege class
deleting 414 description 845
migration to 271, 392 granting 849
restrictions 261 reducing 849
staging data on disk for tape storage 270 revoking 850
storage management policies storage volume
description of 24, 482 auditing 887
managing 475 contents 388
tailoring 496 formatting random access 69, 255
using standard 477 information about 385
storage occupancy, querying 397 labeling sequential access 129, 255
storage pool monitoring use 384
3592, special considerations for 183 overview 44
active-data pool 241 preparing sequential access 129, 255
amount of space used 398 StorageTek devices 195
auditing a volume 887 stub file 481
comparing primary and copy types 412 subfile backups
copy 241 deleting 554
creating a hierarchy 260 description of 552
data format 206, 245, 249 example of 552
defining 245 expiring 554
defining a copy storage pool 410 managing 553
defining for disk, example 249, 261 restoring 554
defining for NDMP operations 217 subordinate storage pool 261
defining for tape, example 249, 261 subscriber, deleting 724
deleting 413 subscription
description of 240 defining 717, 719
destination in copy group 502, 508 deleting 720
determining access mode 245, 410 scenario 718
determining maximum file size 245 subset node replication 944
determining whether to use collocation 245, 357, 410 substitution variables, using 631
disk 21 support contract 1107
duplicate, using to restore 913 support information 1103
enabling cache for disk 245, 284 support subscription 1107
estimating space for archived files on disk 381 supported devices 36
estimating space for backed up files on disk 380 system catalog tables 794
estimating space for disk 380 system privilege class
estimating space for sequential 381 revoking 850
estimating space in multiple 260
incomplete, using to restore 913
increasing sizes 21
LTO Ultrium, special considerations for 191
T
table of contents 233
managing 239
generating for a backup set 549
monitoring 382
managing 209, 234
moving files 401
tape
moving files between 402
capacity 199
multiple, using to restore 913
compatibility between drives 159
next storage pool
exporting data 755
definition 261
finding for client node 391
deleting 414
monitoring life 387
migration to 271, 392
number of times mounted 387
overview 43
planning for exporting data 754
policy use 502, 508
recording format 180
primary 240
rotation 55, 146
protecting 883, 885, 886, 887, 898
scratch, determining use 245, 257, 410
querying 382
setting mount retention period 182
renaming 409
tape drive, replacing 159
restoring 907, 908, 912, 913, 916
tape failover 994
search-and-selection order for active files 243
target replication server
simultaneous-write function 331
configuring 966
updating 245
target server 734

Index 1163
TCP keepalive type, device
enabling 211 3590 179
overview 210 4MM 177, 179
specifying connection idle time 211 8MM 177, 179
TCP/IP 432 CENTERA 42
connect server to database 646 DISK 177
IPv4 432 DLT 177, 179
IPv6 432 ECARTRIDGE 179
TECUTF8EVENT option 814 FILE 177
temporary disk space 654, 655 GENERICTAPE 177, 179
temporary space 654, 655 LTO 179, 191
test replication 977 multiple in a single library 52
text editor REMOVABLEFILE 177
to work with client 427 SERVER 177, 179, 734, 736
threshold VOLSAFE 195
migration, for storage pool WORM 177, 179
random access 273 typographic conventions xv
sequential access 280
reclamation 245, 367, 411
throughput capability 977
THROUGHPUTDATATHRESHOLD server option 470
U
Ultrium, LTO device type
THROUGHPUTTIMETHRESHOLD server option 470
device class, defining and updating 191
tiered data deduplication 317
encryption 154, 193, 536
tiering 317
WORM 133, 195
time interval, setting for checking in volumes 183
unavailable access mode
timeout
description 259
client session 470
marked with PERMANENT parameter 145
Tivoli Directory Server
uncertain, schedule status 574, 626
configure for TLS 840
Unicode
configure for TLS on the CLI 841
automatically renaming file space 456
Tivoli Enterprise Console 811
client platforms supported 453
setting up as a receiver 814
clients and existing backup sets 462
Tivoli event console 807, 811
deciding which clients need enabled file spaces 454
Tivoli Storage Manager
description of 453
server network 687
displaying Unicode-enabled file spaces 462
Tivoli Storage Manager definitions 789
example of migration process 460
Tivoli Storage Manager device driver 80, 81
file space identifier (FSID) 462
Tivoli Storage Manager for Space Management 501
how clients are affected by migration 459
description 481
how file spaces are automatically renamed 457
files, destination for 501
migrating client file spaces 455
migration of client files
options for automatically renaming file spaces 456
description 481
Unicode versions
eligibility 496
planning for 458
policy for, setting 496, 501
unified logon
premigration 481
enable 866
recall of migrated files 482
unified logon for Windows NT 865
reconciliation between client and server 482
uninstalling 81
selective migration 481
UNIQUETDPTECEVENTS option 812
setting policy for 501
UNIQUETECEVENTS option 812
simultaneous-write function, version support for 333
UNLOCK ADMIN command 852
space-managed file, definition 481
UNLOCK NODE command 440
stub file 481
UNLOCK PROFILE command 711, 712
TLS (Transport Layer Security)
unplanned shutdown 607
specifying communication ports 835
unreadable files 896, 897
TLS communication 832
UPDATE ADMIN command 850
transactions, database 643, 685
UPDATE ARCHIVE command 556
transparent recall 482
UPDATE BACKUPSET command 550
Transport Layer Security (TLS) 833
UPDATE CLIENTOPT command 467
specifying communication ports 835
UPDATE CLOPTSET command 467
troubleshooting
UPDATE COPYGROUP command 502, 508
device configuration 123
UPDATE DEVCLASS command 179
errors in database with external media manager 150
UPDATE DOMAIN command 500
tsmdlst 76
UPDATE LIBVOLUME command 46
tsmdlst utility 76
UPDATE MGMTCLASS command 501
TXNBYTELIMIT client option 263
UPDATE NODE command 434, 460, 465
TXNGROUPMAX server option 262
UPDATE POLICYSET command 500
UPDATE RECOVERYMEDIA command 1017

1164 IBM Tivoli Storage Manager for Windows: Administrator's Guide

UPDATE SCHEDULE command 621 volumes (continued)
UPDATE SCRIPT command 633, 634 automated library inventory 47, 128
UPDATE SERVER command 700, 701 capacity, compression effect 200
UPDATE VOLUME command 256 checking out 136
upgrading tape drives 159 contents, querying 388
URL defining to storage pools 256
Operations Center 589 delaying reuse 378, 887
URL for client node 420 deleting 415, 416, 615
user exit 807 detailed report 390
user exit program 825, 1099 determining which are mounted 754
user ID, administrative disk storage 256
creating automatically 448 disk storage pool, auditing 894
description of 420 errors, read and write 386
preventing automatic creation of 448 estimated capacity 386
user-exit declarations 823, 1097 finding for client node 391
help in dsmc loop session 467
labeling using commands 148
V location 387
managing 135
validate
monitoring life 387
node data 536
monitoring movement of data 405
VALIDATE LANFREE command 123
monitoring use 384
VALIDATE POLICYSET command 510
mount retention time 182
validating data
moving files between 401
during a client session 535
number of times mounted 387
for storage pool volumes 890
off-site, limiting number to be reclaimed 377
for virtual volumes 732
offsite, limiting number to be reclaimed 245
logical block protection 155
overview 46
performance considerations for nodes 536
pending status 387
performance considerations for storage pools 893
private 46
variable, accounting log 803
querying contents 388
VARY command 70
querying for general information 385
varying volumes on or off line 70
random access storage pools 240, 255, 257
VERDELETED parameter 477, 504
reclamation 372
VEREXISTS parameter 477, 504
reuse delay 378, 887
verify
scratch 46
cluster configuration 1000
scratch, using 257
versions data deleted, description of 477, 504
sequential 256
versions data exists, description of 477, 504
sequential storage pools 129, 255
virtual file space mapping, command 235
setting access mode 259
virtual tape libraries 116, 118
standard report 389
configuring 116
status, in automated library 46, 127
managing 116
status, information on 386
virtual tape library 38, 116, 117
updating 256
configuring 117
using private 46, 127
storage capacity 116
varying on and off 70
virtual volume
WORM scratch category 96
performance expectations 735
VTL 117
virtual volumes, server-to-server
deduplication 732
reclaiming 374
using to store data 732 W
VIRTUALMOUNTPOINT client option 452 web administrative interface
Vital Cartridge Records (VCR), corrupted condition 62 description 17
VMware host environment Web administrative interface
hardware scans 450, 603 limitation of browser for script definitions 627
VOLSAFE device class 195 Web backup-archive client
volume capacity 180 granting authority to 447
volume history 878 remote access overview 445
deleting information from 615 URL 420, 445
volume history file 72, 878 web server
volume reuse 72 starting 593
volumes stopping 593
access preemption 612 what's new
allocating space for disk 69, 255 best practice xx
assigning to storage pool 255 Windows Active Directory
auditing 137, 887 configuring 842
auditing considerations 887 Windows cluster configuration 992

Index 1165
Windows clustered environment 992
Windows device driver 81
Windows Server 2008 82
Windows unified logon 865
wizard
client configuration 427
cluster configuration 1001
setup 427
workstation, registering 424
WORM devices and media
DLT WORM 133
IBM 3592 133
LTO WORM 133
Oracle StorageTek T10000B drives 134
Oracle StorageTek T10000C drives 134
Quantum LTO3 133
reclamation of optical media 373
Sony AIT50 and AIT100 133
special considerations for WORM media 133
VolSafe
considerations for media 133
defining VOLSAFE device classes 195
WORM FILE and SnapLock 516
WORM parameter 195
WORM scratch category for volumes in 349X library 96
writing data simultaneously to primary and copy storage
pools
use during client storage operations 331

1166 IBM Tivoli Storage Manager for Windows: Administrator's Guide




Product Number: 5608-E01

5608-E02
5608-E03

Printed in USA