VERITAS NetBackup

Advanced Client 5.1

System Administrators Guide

UNIX and Windows

N12348C
Disclaimer
The information contained in this publication is subject to change without notice. VERITAS Software
Corporation makes no warranty of any kind with regard to this manual, including, but not limited to,
the implied warranties of merchantability and fitness for a particular purpose. VERITAS Software
Corporation shall not be liable for errors contained herein or for incidental or consequential damages
in connection with the furnishing, performance, or use of this manual.

VERITAS Legal Notice

Copyright 2002 - 2004 VERITAS Software Corporation. All rights reserved. VERITAS, the VERITAS

logo, and all other VERITAS product names and slogans are trademarks or registered trademarks of

VERITAS Software Corporation. VERITAS, NetBackup, the VERITAS logo, Reg. U.S. Pat. & Tm. Off.

Other product names and/or slogans mentioned herein may be trademarks or registered trademarks

of their respective companies.

VERITAS Software Corporation

350 Ellis Street

Mountain View, CA 94043

USA

Phone 6505278000 Fax 6505272908

www.veritas.com

Third-Party Copyrights
ACE 5.2A: ACE(TM) is copyrighted by Douglas C.Schmidt and his research group at Washington University and University of California, Irvine,
Copyright (c) 1993-2002, all rights reserved.
IBM XML for C++ (XML4C) 3.5.1: Copyright (c) 1999,2000,2001 Compaq Computer Corporation; Copyright (c) 1999,2000,2001 Hewlett-Packard
Company; Copyright (c) 1999,2000,2001 IBM Corporation; Copyright (c) 1999,2000,2001 Hummingbird Communications Ltd.; Copyright (c)
1999,2000,2001 Silicon Graphics, Inc.; Copyright (c) 1999,2000,2001 Sun Microsystems, Inc.; Copyright (c) 1999,2000,2001 The Open Group; All
rights reserved.
Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to
deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, and/or sell copies
of the Software, and to permit persons to whom the Software is furnished to do so, provided that the above copyright notice(s) and this permission
notice appear in all copies of the Software and that both the above copyright notice(s) and this permission notice appear in supporting
documentation.
This product includes software developed by the Apache Software Foundation (http://www.apache.org/).
JacORB 1.4.1: The licensed software is covered by the GNU Library General Public License, Version 2, June 1991.
Open SSL 0.9.6: This product includes software developed by the OpenSSL Project * for use in the OpenSSL Toolkit. (http://www.openssl.org/)
TAO (ACE ORB) 1.2a: TAO(TM) is copyrighted by Douglas C. Schmidt and his research group at Washington University and University of
California, Irvine, Copyright (c) 1993-2002, all rights reserved.
Contents

Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xiii

What is In This Manual? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xiii

Related Manuals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .xv

Getting Help . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .xv

Glossary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .xvii

Conventions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .xvii

Advanced Client Information on the Web . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xix

NDMP Information on the Web . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xx

Chapter 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1

Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2

Features Included in Advanced Client . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2

Snapshots . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2

Offhost Backup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2

Instant Recovery . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4

FlashBackup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4

BLI . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4

Advanced Client and NDMP . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4

Snapshot Basics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4

Copy-on-Write . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5

Mirror . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6

Snapshot methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7

Local Backup of Snapshot . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7

Offhost Backup Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9

iii
 File/Volume Mapping Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9

Offhost Backup Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10

Alternate Client Backup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10

Data Sharing Between Clients . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11

NetBackup Media Server (UNIX Only) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14

Third-Party Copy Device (UNIX Only) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15

Network Attached Storage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16

Offhost Backup Without a SAN (UNIX Only) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16

Offhost Backup with Multi-Ported Array (UNIX Only) . . . . . . . . . . . . . . . . . . . . . . . . . 17

Features and Required Software . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18

Snapshot Methods and Requirements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21

Requirements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24

Restrictions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24

Terminology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26

Chapter 2. Installation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33

Prerequisites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33

Installing Advanced Client On UNIX . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34

Loading From Media . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34

Distributing Advanced Client Software to UNIX Clients . . . . . . . . . . . . . . . . . . . . . 35

Installing Advanced Client On Windows . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37

Distributing Client Software in Mixed-Platform Environments . . . . . . . . . . . . . . . . . . 38

Creating Log Directories . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39

Upgrading from Earlier Releases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39

Conversion from 4.5 FP3 Feature Pack . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39

Conversion from 4.5 GA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39

Uninstalling NetBackup Advanced Client or Earlier Packages . . . . . . . . . . . . . . . . . . . 40

Server Uninstall - UNIX . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40

Client Uninstall - UNIX . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41

Server/Client Uninstall - Windows . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42

iv NetBackup Advanced Client System Administrators Guide

 Chapter 3. SAN Configuration for Advanced Client . . . . . . . . . . . . . . . . . . . . . . . . . . 43

SAN Configuration Diagram . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44

Supported Peripherals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45

Media Server/Third-Party Copy Requirements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45

Diagram for NetBackup Media Server . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46

Diagram for Third-Party Copy Device . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47

Diagram for Third-Party Copy Device - Remote . . . . . . . . . . . . . . . . . . . . . . . . . . 48

Configuration Flowcharts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49

Verify NetBackup Access to SAN Devices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53

Solaris only: Example for sg.links, sg.conf, and st.conf files . . . . . . . . . . . . . . . . . . . 57

Device Checklist . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58

Solaris only: Configure HBA Drivers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59

Create Backup Configuration Files . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60

The 3pc.conf and mover.conf Files: An Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . 60

3pc.conf Description . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61

Example 3pc.conf file . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61

Determining Requirements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64

What bptpcinfo Automatically Provides . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65

What the Backup Methods Require . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65

mover.conf Description . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65

Types of entries in mover.conf . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65

For sites that have one third-party copy device . . . . . . . . . . . . . . . . . . . . . . . . . . . 67

For sites that have multiple third-party copy devices . . . . . . . . . . . . . . . . . . . . . 69

SCSI Reserve/Release . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69

Keywords in Mover File . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70

Naming the Mover File . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73

Selection Priority for mover.conf files . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74

Create the 3pc.conf File . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75

Create the mover.conf File . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77

Contents v

 Chapter 4. Policy Configuration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83

Notes and Prerequisites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84

Configuring an Advanced Client Policy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85

Automatic Snapshot Selection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89

Selecting the Snapshot Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90

Configuring Backup Scripts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95

Configuring Alternate Client Backup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96

Basic Requirements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96

Available Backup Methods and Snapshot Methods . . . . . . . . . . . . . . . . . . . . . . . . . 97

Example configurations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98

Before running the alternate client backup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99

Configuration Tips . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99

Maximum Pathname Length . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99

Snapshot Tips . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99

Backup Policy Configuration Wizard . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100

Multiple Data Streams . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100

Chapter 5. FlashBackup Configuration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101

FlashBackup Capabilities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102

Restrictions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102

Configuring a FlashBackup Policy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103

Configuring FlashBackup in the Earlier Manner (UNIX only) . . . . . . . . . . . . . . . . . . 106

Snapshot Selection in Earlier Manner . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106

Cache Partition in Earlier Manner . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106

Using Multiple Data Streams with Pre-5.0 Clients . . . . . . . . . . . . . . . . . . . . . . . . . . 108

Chapter 6. NAS Snapshot Configuration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111

NAS Snapshot Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112

Notes and Prerequisites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113

Setting up a Policy for Snapshots . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113

vi NetBackup Advanced Client System Administrators Guide

 Chapter 7. Instant Recovery Configuration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117

Instant Recovery Capabilities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 118

Requirements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 118

Restrictions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 118

Instant Recovery Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119

Snapshot and Backup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120

Maintaining the NetBackup Catalogs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120

VxFS Storage Checkpoints . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120

VxVM Split Mirrors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121

Configuring a Policy for Instant Recovery . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122

Configuring Third Mirror Split Volume Rotation . . . . . . . . . . . . . . . . . . . . . . . . . . . 125

Configuring VxVM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126

Creating a Snapshot Mirror . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126

Creating an Instant Snapshot . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127

Creating space-optimized snapshots . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127

Creating full-sized snapshots . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128

Using the VxVM Graphical User Interface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129

Using VERITAS Enterprise Administrator (VxVM 3.5) . . . . . . . . . . . . . . . . . . . 129

Instant Recovery for Databases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131

Chapter 8. Snapshot Configuration Notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133

nbu_snap . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134

Cache device . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134

Sizing the Cache Partition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135

How to Enter the Cache . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 136

VxFS_Checkpoint . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138

VxFS Multi-Device System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138

Storage Checkpoint Disk Usage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138

Checkpoint Retention Schedules . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139

Block-Level Restore . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139

Contents vii

 VxFS_Snapshot . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 140

vxvm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141

Creating a Snapshot Mirror of the Source . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141

VxVM Instant Snapshots . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 142

NetBackup Snapshot Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 143

Space-Optimized Instant Snapshots . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 143

FlashSnap . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 144

Testing Volumes for FlashSnap . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 144

VVR . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147

Set Up Volume Replication . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147

Name Registration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147

Primary/Secondary Disk Group and Volume Names . . . . . . . . . . . . . . . . . . . . . . 147

Test the Replication Setup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 148

NAS_Snapshot . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149

Array-Specific Snapshot Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 150

Configuration Checklist . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 150

Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151

The Snapshot Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151

Client Data Must Be Mirrored . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153

Disk Terms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153

Disk Configuration Requirements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 154

Access to Disk Arrays . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 154

Connection to Disk Array: SCSI and Fibre Channel . . . . . . . . . . . . . . . . . . . . . . . . 154

Multiple Connectivity to EMC Array: Common Serial Number mode . . . . . . . . 156

Configuring Primary and Secondary Disks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 156

EMC Symmetrix . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 157

Hitachi and HP Arrays . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 158

Volume Manager Configuration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 168

Disk Label . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 168

Disk Types . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 168

viii NetBackup Advanced Client System Administrators Guide

 Disk Group Clones . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 170

When Secondary Disks are Split and Synched . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 171

Best Practices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 172

NetBackup Access to Arrays . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 172

Resynchronizing Disks At End of Backup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 172

Hardware-Level Disk Restore . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 172

Volume Manager Disk Groups . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 172

Volume Manager with Dynamic Multipathing (DMP) . . . . . . . . . . . . . . . . . . . . . . 173

Backups Concurrently Accessing Same Disk (no VxVM) . . . . . . . . . . . . . . . . . . . . 173

Backups Concurrently Accessing VxVM Volumes . . . . . . . . . . . . . . . . . . . . . . . . . . 174

Concurrent Access to Same VxVM Volume . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 174

Concurrent Access to Volume Manager Volumes on Same Disks . . . . . . . . . . 174

Avoiding Concurrent Access Conflicts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 175

Chapter 9. Notes on Media Server and Third-Party Copy . . . . . . . . . . . . . . . . . . . . 177

Disk Requirements for Media Server/ Third-Party Copy . . . . . . . . . . . . . . . . . . . . . . . 178

ALL_LOCAL_DRIVES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 178

Storage Units . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 178

Multiplexing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 178

Raw Partition Backups . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 178

Chapter 10. Backup and Restore Procedures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 181

Performing a Backup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 182

Automatic Backup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 182

Manual Backup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 182

User-Directed Backup and Archive . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 182

Performing a Restore . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 183

Restores from a FlashBackup Backup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 183

Restores in Clustered File System (VxFS on UNIX Only) . . . . . . . . . . . . . . . . . . . . 184

Instant Recovery Restore . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 184

Instant Recovery: Block-Level Restore (UNIX Clients Only) . . . . . . . . . . . . . . . 185

Contents ix

 Instant Recovery: File Promotion (UNIX Clients Only) . . . . . . . . . . . . . . . . . . . 185

Instant Recovery: Snapshot Rollback . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 186

Configurations for Restore . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 192

Restoring from a Disk Snapshot . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 193

Chapter 11. Troubleshooting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 199

Gathering Information and Checking Logs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 200

Logging Directories for UNIX Platforms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 200

Logging Folders for Windows platforms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 201

Contacting VERITAS Customer Support . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 203

Important Notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 203

Particular Issues . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 205

Installation Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 205

FlashBackup and Status code 13 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 205

Removing a Snapshot . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 206

Removing a VxVM Volume Clone . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 211

Appendix A. Managing Snapshots from the Command Line . . . . . . . . . . . . . . . . . 213

Reasons for Using the Command Line . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 214

Notes on Using the Command Line . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 214

Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 215

Appendix B. Processing Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 219

Processing Before and After the Snapshot . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 220

Quiescing the System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 220

Quiescing the Database Application . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 221

Quiescing the Stack . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 222

FlashBackup: Local and Offhost Diagrams . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 223

Local Client . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 223

NetBackup Media Server . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 224

Third-Party Copy Device . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 225

x NetBackup Advanced Client System Administrators Guide

 Process Diagrams: Backup and Restore . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 226

Alternate Client Backup: Split-Mirror/VVR . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 226

NetBackup Media Server: Copy-On-Write Process . . . . . . . . . . . . . . . . . . . . . . . . . 228

NetBackup Media Server: Mirror Process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 230

Third-Party Copy Device: Copy-On-Write Process . . . . . . . . . . . . . . . . . . . . . . . . . 232

Third-Party Copy Device: Mirror Process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 234

Instant Recovery: Backup Process Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 236

Instant Recovery: Backup to disk only . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 238

Instant Recovery: Backup to tape . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 239

Instant Recovery: Normal Restore . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 240

Instant Recovery: Browsing from Snapshot . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 242

Instant Recovery: File Promotion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 243

Instant Recovery: Rollback . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 244

FlashBackup: Backup Process Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 245

FlashBackup: Restore Process Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 247

How Copy-on-Write Works . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 249

Copy-on-write process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 250

Backing up a copy-on-write snapshot . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 251

Appendix C. Managing nbu_snap (Solaris) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 253

Cache for nbu_snap . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 254

Determining Cache Size . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 254

Terminating nbu_snap . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 254

nbu_snap Commands . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 255

snapon . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 255

snaplist . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 255

snapcachelist . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 257

snapstat . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 258

snapoff . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 258

Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 259

Contents xi

xii NetBackup Advanced Client System Administrators Guide

Preface

This guide explains how to install, configure, and use VERITAS NetBackup Advanced
Client. This guide is intended for the NetBackup system administrator and assumes a
thorough working knowledge of UNIX or Windows, and of NetBackup administration.

Note Advanced Client combines the features formerly provided by several products:
Core Frozen Image Services, Extended Frozen Image Services (Array Integration
Option), Offhost and SAN Data Movement Services, FlashBackup, Persistent
Frozen Image, and Oracle BLI Agent. This Advanced Client manual replaces the
following three guides: NetBackup ServerFree Agent, NetBackup Persistent Frozen
Image, and NetBackup FlashBackup.

What is In This Manual?

Chapters in This Manual

Chapter Description

Introduction An overview of the products capabilities, with

background information on snapshots and offhost
backup.

Installation Explains how to install and de-install NetBackup

Advanced Client.

SAN Configuration for Advanced Provides assistance in setting up the hardware.

Client

Policy Configuration Explains how to configure NetBackup policies/clients

for snapshots and offhost backup.

FlashBackup Configuration Explains how to configure FlashBackup policies.

xiii
What is In This Manual?

Chapters in This Manual

Chapter Description

NAS Snapshot Configuration Explains how to configure policies to create snapshots of

client data on NDMP (NAS) hosts.

Instant Recovery Configuration Explains how to prepare for and configure an Instant
Recovery policy.

Snapshot Configuration Notes Provides configuration assistance for the Advanced

Client snapshot methods.

Notes on Media Server and Provides notes on offhost (server-free) backup.

Third-Party Copy

Backup and Restore Procedures Briefly explains how to back up and restore files.

Troubleshooting Helps in resolving problems when using NetBackup

Advanced Client. See also the NetBackup Troubleshooting
Guide.

Managing Snapshots from the Explains basic procedures for creating and managing
Command Line snapshot images using the NetBackup command-line
interface.

Processing Background Presents background information on snapshot

processing and backup/restore processing.

Managing nbu_snap (Solaris) Provides information about snapctl.

xiv NetBackup Advanced Client System Administrators Guide

 Related Manuals

Related Manuals
NetBackup Release Notes

Describes supported platforms and provides operating notes not found in the
manuals or in the online help.
NetBackup Installation Guide
Explains how to install NetBackup.
NetBackup System Administrators Guide, Volumes I and II
Explains how to configure and manage NetBackup.
NetBackup Media Manager Device Configuration Guide for UNIX and Windows
Explains how to add device drivers and perform other system level configuration for
storage devices that are supported by NetBackup and Media Manager.
NetBackup Troubleshooting Guide
Explains NetBackup Advanced Client error codes.

Getting Help
VERITAS offers you a variety of support options.

Accessing the VERITAS Technical Support Web Site

The VERITAS Support Web site allows you to:
obtain updated information about NetBackup, including system requirements,
supported platforms, and supported peripherals
contact the VERITAS Technical Support staff and post questions to them
get the latest patches, upgrades, and utilities
view the NetBackup Frequently Asked Questions (FAQ) page
search the knowledge base for answers to technical support questions
receive automatic notice of product updates
find out about NetBackup training
read current white papers related to NetBackup
The address for the VERITAS Technical Support Web site follows:
http://support.veritas.com

Preface xv
Getting Help

Subscribing to VERITAS Email Notification Service

Subscribe to the VERITAS Email notification service to be informed of software alerts,
newly published documentation, Beta programs, and other services.
Go to http://support.veritas.com. Select a product and click E-mail Notifications on the
right side of the page. Your customer profile ensures you receive the latest VERITAS
technical information pertaining to your specific interests.

Accessing VERITAS Telephone Support

Telephone support for NetBackup is only available with a valid support contract. To
contact VERITAS for technical support, dial the appropriate phone number listed on the
Technical Support Guide included in the product box and have your product license
information ready for quick navigation to the proper support group.

To locate the telephone support directory on the VERITAS web site

1. Open http://support.veritas.com in your web browser.

2. Click the Phone Support icon. A page that contains VERITAS support numbers from
around the world appears.

Accessing VERITAS E-mail Support

To contact support using E-mail on the VERITAS web site

1. Open http://support.veritas.com in your web browser.

2. Click the E-mail Support icon. A brief electronic form will appear and prompt you to:
Select a language of your preference
Select a product and a platform
Associate your message to an existing technical support case
Provide additional contact and product information, and your message

3. Click Send Message.

Contacting VERITAS Licensing

For license information call 1-800-634-4747 option 3, fax 1-650-527-0952, or e-mail
[email protected].

xvi NetBackup Advanced Client System Administrators Guide

 Glossary

Glossary
If you encounter unfamiliar terminology, consult the NetBackup online glossary. The
glossary contains terms and definitions for NetBackup and all additional NetBackup
options and agents.
The NetBackup online glossary is included in the NetBackup help file.

To access the NetBackup online glossary

1. In the NetBackup Administration Console, click Help > Help Topics.

2. Click the Contents tab.

3. Click Glossary of NetBackup Terms.

Use the scroll function to navigate through the glossary.

Conventions
The following conventions apply throughout the documentation set.

Product-Specific Conventions
The following term is used in the NetBackup 5.1 documentation to increase readability
while maintaining technical accuracy.
Microsoft Windows, Windows
Terms used to describe a specific product or operating system developed by
Microsoft, Inc. Some examples you may encounter in NetBackup documentation are,
Windows servers, Windows 2000, Windows Server 2003, Windows clients, Windows
platforms, or Windows GUI.
When Windows or Windows servers is used in the documentation, it refers to all of
the currently supported Windows operating systems. When a specific Windows
product is identified in the documentation, only that particular product is valid in that
instance.
For a complete list of Windows operating systems and platforms that NetBackup
supports, refer to the NetBackup Release Notes for UNIX and Windows or go to the VERITAS
support web site at http://www.support.veritas.com.

Preface xvii
Conventions

Typographical Conventions
Here are the typographical conventions used throughout the manuals:

Conventions

Convention Description

GUI Font Used to depict graphical user interface (GUI) objects, such as fields,
listboxes, menu commands, and so on. For example: Enter your
password in the Password field.

Italics Used for placeholder text, book titles, new terms, or emphasis. Replace
placeholder text with your specific text. For example: Replace filename
with the name of your file. Do not use file names that contain spaces.
This font is also used to highlight NetBackup server-specific or operating
system-specific differences. For example: This step is only applicable for
NetBackup Enterprise Server.

Code Used to show what commands you need to type, to identify pathnames
where files are located, and to distinguish system or application text that
is displayed to you or that is part of a code example.

Key+Key Used to show that you must hold down the first key while pressing the
second key. For example: Ctrl+S means hold down the Ctrl key while
you press S.

You should use the appropriate conventions for your platform. For example, when
specifying a path, use backslashes on Microsoft Windows and slashes on UNIX.
Significant differences between the platforms are noted in the text.
Tips, notes, and cautions are used to emphasize information. The following samples
describe when each is used.

Tip Used for nice-to-know information, like a shortcut.

Note Used for important information that you should know, but that shouldnt cause any
damage to your data or your system if you choose to ignore it.

Caution Used for information that will prevent a problem. Ignore a caution at your own
risk.

Command Usage
The following conventions are frequently used in the synopsis of command usage.

xviii NetBackup Advanced Client System Administrators Guide

 Advanced Client Information on the Web

brackets [ ]
The enclosed command line component is optional.
Vertical bar or pipe (|)
Separates optional arguments from which the user can choose. For example, when a
command has the following format:
command arg1|arg2

In this example, the user can use either the arg1 or arg2 variable.

Navigating Multiple Menu Levels

When navigating multiple menu levels, a greater-than sign (>) is used to indicate a
continued action.
The following example shows how the > is used to condense a series of menu selections
into one step:

Select Start > Programs > VERITAS NetBackup > NetBackup Administration
Console.
The corresponding actions could be described in more steps as follows:

1. Click Start in the task bar.

2. Move your cursor to Programs.

3. Move your cursor to the right and highlight VERITAS NetBackup.

4. Move your cursor to the right. First highlight and then click NetBackup
Administration Console.

Advanced Client Information on the Web

The VERITAS support web site includes a compilation of notes on Advanced Client.
These notes include up-to-date lists of supported operating systems and peripherals, and
assistance for some of the configuration procedures found in this manual. To locate this
information, do the following:

1. Go to www.support.veritas.com.

2. In the Knowledge Base Search field, enter the following:

AdvancedClient_ConfigNotes

Preface xix
NDMP Information on the Web

3. Click Search.
The full title of the document is:
VERITAS NetBackup Advanced Client Configuration and Compatibility

NDMP Information on the Web

The VERITAS support web site has up-to-date information on NDMP supported
operating systems and NAS vendors. It also contains configuration and troubleshooting
help for particular NAS systems. To locate this information, do the following:
1. Go to www.support.veritas.com.

2. In the Knowledge Base Search field, enter the following:

NAS Appliance
The full title of the document is:
NetBackup for NDMP Supported OS and NAS Appliance Information

xx NetBackup Advanced Client System Administrators Guide

Introduction 1

This chapter describes NetBackup Advanced Client and contains the following topics.

Overview

Snapshot Basics

Local Backup of Snapshot

Offhost Backup Overview

Offhost Backup Methods

Offhost Backup Without a SAN (UNIX Only)

Offhost Backup with Multi-Ported Array (UNIX Only)

Features and Required Software

Requirements

Terminology

Note For help with first-time setup of Advanced Client, see the NetBackup Advanced Client
Quick Start Guide.

1
Overview

Overview
Advanced Client combines the features of snapshot backup, FlashBackup, BLI Agent,
offhost backup, and Instant Recovery. It supports clients on UNIX and Windows
platforms, on either Fibre Channel networks (SANs) or traditional LANs.

Features Included in Advanced Client

Starting with NetBackup 5.0, Advanced Client combined the features formerly provided
by the following NetBackup add-on products: Core Frozen Image Services, Extended
Frozen Image Services (Array Integration Option), Offhost and SAN Data Movement
Services, FlashBackup, Persistent Frozen Image, and Oracle BLI Agent.

Snapshots
A snapshot is a disk image of the clients data made almost instantaneously. NetBackup
backs up the data from the snapshot image, not directly from the clients primary data.
This allows client operations and user access to continue without interruption during the
backup.
A snapshot image is required for all features of Advanced Client. A number of methods
are provided for creating snapshots. You can select the snapshot method manually from
the Policy dialog of the NetBackup Administration Console, or allow NetBackup to select
the method for you.
Refer to Snapshot Basics on page 4 for background on snapshot technology.

Offhost Backup
Another major component of NetBackup Advanced Client is support for offhost backup.
Offhost backup shifts the burden of backup processing onto a separate backup agent,
greatly reducing the impact on the clients computing resources ordinarily caused by a
local backup. The client supplies a relatively small amount of mapping information, but
the backup agent does the bulk of the work by sending the clients actual data to the
storage device.
See the following network diagram showing a backup agent.

2 NetBackup Advanced Client System Administrators Guide

 Overview

Backup Agent for Offhost Backup

NetBackup master
server

LAN / WAN

NetBackup
SCSI
client Backup agent
Local client
data storage

Fibre Channel/SAN Backup agent

Robot on
SAN
Disks of client
data on SAN

The backup agent can be any of the following:

an additional (alternate) client
a NetBackup media server or a third-party copy device that implements the SCSI
Extended Copy command.
a NAS host (Network Attached Storage)
Note that many types of devices are designed to act as third-party copy devices, such as
routers, bridges, robotic libraries, and disk arrays. The backup agent can direct the data to
SCSI-attached storage or to storage on the SAN.

Chapter 1, Introduction 3
Snapshot Basics

Instant Recovery
This feature makes backups available for instant recovery from disk. Instant Recovery
combines snapshot technologythe image is created without interrupting user access to
datawith the ability to do rapid snapshot-based restores. The image is retained on disk
and can also be backed up to tape or other storage.
Instant Recovery makes possible three additional variations of restore: block-level restore,
file promotion, and snapshot rollback. For a description, refer to Instant Recovery
Restore on page 184.

FlashBackup
FlashBackup is a policy type that combines the speed of raw-partition backups with the
ability to restore individual files.

BLI
Block level incremental backup extends the capabilities of NetBackup to back up only
changed data blocks of Oracle database files. Refer to the NetBackup for Oracle System
Administrators Guide for more details.

Advanced Client and NDMP

Using the NDMP V4 snapshot extension, NetBackup Advanced Client can make
policy-based snapshots of data on a Network Attached Storage (NAS) host. The snapshot
is stored on the same NAS device that contains the primary client data. From the
snapshot, you can restore individual files or roll back an entire volume or file system, by
means of Instant Recovery.

Note NetBackup for NDMP add-on software is required, and the NAS vendor must
support snapshots.

Snapshot Basics
Large active databases or file systems that must be available around-the-clock are difficult
to back up without incurring a penalty. Often, the penalty takes one of two forms:
The entire database is taken offline or the file system is unmounted, to allow time for
the backup, resulting in suspension of service and inconvenience to users.

4 NetBackup Advanced Client System Administrators Guide

 Snapshot Basics

The copy is made very quickly but produces an incomplete version of the data, some
transactions having failed to complete.
A solution to this problem is to create a snapshot of the data. This means capturing the data
at a particular instant, without causing significant client downtime. The resulting capture
or snapshot can be backed up without affecting the performance or availability of the file
system or database. Without a complete, up-to-date snapshot of the data, a correct backup
cannot be made.
When a backup is managed by a backup agent on a Fibre Channel network, the data to
back up must be contained in a snapshot. The backup agent can only access the data by
means of the raw physical disk. Once the data is captured as a snapshot, the NetBackup
client maps the logical representation of the data to its absolute physical disk address.
These disk addresses are sent to the backup agent over the LAN and the data is then read
from the appropriate disk by the backup agent. (This process is explained in greater detail
under Offhost Backup Overview later in this chapter.)
Two types of snapshots are available, both supported by NetBackup: copy-on-write, and
mirror.

Copy-on-Write
A copy-on-write type of snapshot is a detailed account of data as it existed at a certain
moment. A copy-on-write is not really a copy of the data, but a specialized record of it.
A copy-on-write snapshot is created in available space in the clients file system or in a
designated raw partition, not as a complete copy of the client data on a separate or mirror
disk. The snapshot is then backed up to storage as specified in the backup policy. Users
can access their data without interruption, as though no backup is taking place. The file
system is paused just long enough to assemble a transactionally consistent record.
For a description of the copy-on-write process, see How Copy-on-Write Works on
page 249. Note that copy-on-write snapshots are called Storage Checkpoints in VxFS.

Benefits of copy-on-write:
Consumes less disk space: no need for secondary disks containing complete copies of
source data.
Relatively easy to configure (no need to set up mirror disks).

Chapter 1, Introduction 5
Snapshot Basics

Mirror
Unlike a copy-on-write, a mirror is a complete data copy stored on a separate disk,
physically independent of the original. Every change or write to the data on the primary
disk is also made to the copy on the secondary disk. This creates a mirror image of the
source data.

NetBackup
client

All data written to source

is also written to mirror

Source image Mirror image

(on primary disk) (on secondary)

As in a copy-on-write, transactions are allowed to finish and new I/O on the primary disk
is briefly halted. When the mirror image is brought up-to-date with the source (made
identical to it), the mirror is split from the primary, meaning that new changes can be
made to the primary but not to the mirror. At this point the mirror can be backed up (see
next diagram).

NetBackup
client
Mirror is split from primary:
further writes to primary
are not made to mirror.

Source image Mirror image

(on primary disk) (on secondary)

If the mirror will be used again it must be brought up-to-date with the primary volume
(resynchronized). During resynchronization, the changes made to the primary
volumewhile the mirror was splitare written to the mirror.
Since mirroring requires an exact, complete copy of the primary on a separate device
(equal in size to the disk being mirrored), it consumes more disk space than a
copy-on-write.

6 NetBackup Advanced Client System Administrators Guide

 Local Backup of Snapshot

Benefits of mirror:
Has less impact on the performance of the application or database host being backed
up (NetBackup client), because there is no need to run the copy-on-write mechanism.
Allows faster backups: the backup process reads data from a separate disk (mirror)
operating independently of the primary disk that holds the clients source data. This
means that, unlike the copy-on-write, there is no need to share disk I/O with other
processes or applications. Apart from NetBackup, no other applications have access to
the mirror disk. During a copy-on-write, the source data can be accessed by other
applications as well as by the copy-on-write mechanism.

Note If additional disk drives are available and volumes have been configured with the
VERITAS Volume Manager, a mirror snapshot method is usually a good choice.

Snapshot methods
Advanced Client supports a number of methods for creating a snapshot image. You can
select the method or let NetBackup select it based on your environment. The snapshot
methods are described in the Policy Configuration chapter.

Local Backup of Snapshot

A snapshot can be backed up to any NetBackup storage device. A fibre channel network
or SAN is not required. The following diagram shows a network configuration sufficient
for a local backup of a snapshot. The network configuration is identical to that for normal
NetBackup (no snapshot).

Chapter 1, Introduction 7
Local Backup of Snapshot

Snapshot Backup on Local Network (No Fibre channel/SAN Required)

NetBackup master
server
1
LAN / WAN

2 3
SCSI SCSI

4
Clients disks storage
NetBackup client NetBackup
media server

1. Client backup is initiated by master server, which tells the NetBackup client to create the
snapshot data on the disk.
2. Client sends the data to the media server.
3. Media server processes the backup and reads the client data.
4. Media server writes data to local storage.

8 NetBackup Advanced Client System Administrators Guide

 Offhost Backup Overview

Offhost Backup Overview

One of the principal goals of NetBackup Advanced Client is to move I/O processing off
the primary NetBackup client (application host) to a backup agent. Advanced Client
supports several methods of doing this:
Alternate client backup: the backup is performed by another client (in a split mirror
or data replication configuration).
NetBackup Media Server: the backup is performed by a NetBackup media server
(supported for UNIX clients only).
Third-Party Copy Device: the backup is performed by an independent backup agent
that uses the Extended Copy command (supported for UNIX clients only).
Network Attached Storage: the backup is performed by an NDMP (NAS) host.
For a list of supported third-party copy devices and other peripherals, refer to Advanced
Client Information on the Web on page xix, or to the NetBackup Release Notes.
Note that the disk(s) containing the clients data must be shared, that is, visible to both
the primary NetBackup client and the backup agent. The disk(s) can be connected locally,
or through a SAN.
The backup agent reads data from the snapshot and writes the data to a storage device.
If the backup agent is a media server or third-party copy device, it is unaware of logical
organizations of data such as file systems and volume managers, and can access the data
only from the physical disk address location. In order for NetBackup to perform this type
of backup, it must first translate the logical representation of the data to its physical disk
addresses. This logical-to-physical translation process is referred to as mapping the data.
During the backup, the mapping information is transmitted to the media server.

File/Volume Mapping Methods

The mapping methods used by NetBackup Advanced Client are installed as part of the
NetBackup Advanced Client product. When a backup is initiated, the correct mapping
method is automatically selected by NetBackup, depending on whether the backup data
is configured over physical devices, logical volumes, or file systems.

Chapter 1, Introduction 9
Offhost Backup Methods

Offhost Backup Methods

NetBackup Advanced Client supports several forms of offhost backup, as explained in the
following sections.

Alternate Client Backup

With this feature, all backup processing is off-loaded to another client. Offloading the
work to another or alternate client saves computing resources on the original client. The
snapshot is created on the alternate client, and the backup has zero impact on the original
client.
The following diagram is a basic view of alternate client backup. A NetBackup master
server is connected by means of a local or wide-area network to two clients and a media
server. The primary or original NetBackup client contains the data to be backed up, and
the alternate NetBackup client (which could be at a remote site) has a copy of that data.
The media server can be accessed by the alternate client. This means the media server can
back up the alternate client as a local host.

Alternate Client Backup: Backup is performed on alternate client

NetBackup
master server

LAN / WAN

Data sharing
(mirrors or
replication)
NetBackup primary NetBackup alternate NetBackup
client client media server

Backup processing is done on

alternate client, not on primary
client. Media server can back up
this client as a local host.

10 NetBackup Advanced Client System Administrators Guide

 Offhost Backup Methods

Data Sharing Between Clients

For alternate client backup, the original (primary) and alternate clients must share storage.
Two types of sharing are supported: split mirror, and data replication.

Split mirror
In this configuration, the alternate client has access to mirror disks containing the primary
clients data. Before the backup, the mirror is split from the primary disk and the snapshot
is made on the mirror disk. The alternate client has access to the mirror disk, and the
media server can back up the alternate client. After the backup, the mirror can be
optionally resynchronized with the primary disk.

Note The mirror disk need not be visible to the primary client, only to the alternate client.

Split Mirror, Conceptual View: Primary client and alternate client share data through mirroring.

NetBackup master
server

LAN / WAN

NetBackup primary alternate media

client client server

Shared disks; Snapshot is captured on mirror disk;

mirror split from alternate client has access to this
primary mirror. Media server can back up
alternate client as a local host.
Storage
primary mirror
disk disk

The above configuration is supported by the FlashSnap, TimeFinder, ShadowImage, and

BusinessCopy snapshot methods.

Chapter 1, Introduction 11
Offhost Backup Methods

Split Mirror implementation: Alternate client performs backup for multiple primary clients.

Multiple clients can share a single alternate

backup client of the same platform type

LAN / WAN

Solaris

Solaris Alternate client/

clients media server,
with SSO

Shared
Fibre Channel/SAN Storage

Alternate client/
HP media server,
clients with SSO

HP

12 NetBackup Advanced Client System Administrators Guide

 Offhost Backup Methods

Data Replication (UNIX Only)

In this configuration, a disk on the alternate client keeps a replicated copy of the clients
data. When the backup starts, the mirror on the alternate client is split from its primary
disk, and the snapshot is made on the mirror. At this point, the media server can back up
the alternate client. After the backup, the volume and file system on the mirror are
unmounted and the mirror is resynchronized with its primary disk.

Replication: Primary client and alternate client share data through replication

NetBackup master
server

LAN / WAN

primary alternate media

replication
client client server

client Mirror split

disks from primary

The NetBackup clients

data is replicated on an Storage
alternate clients disks. primary mirror
disk disk

Snapshot is made on the mirror disk on the

alternate client. Media server can back up
alternate client as a local host.

The above configuration is supported by the VVR snapshot method for UNIX clients only,
and requires the VERITAS Volume Manager (VxVM version 3.2 or later) with the VVR
license.

Chapter 1, Introduction 13
Offhost Backup Methods

NetBackup Media Server (UNIX Only)

In this offhost backup method, the NetBackup media server performs the data movement.
This method is not supported for Windows clients.
The figure NetBackup Media Server shows the basic components used in the
NetBackup Media Server method. A NetBackup media server handles the backup
processing and sends the backup data over Fibre Channel to the storage device.

Note If you have a multi-ported SCSI disk array, a fibre channel SAN is not required. See
Offhost Backup Without a SAN (UNIX Only) on page 16.

NetBackup Media Server

NetBackup
master
server

1 LAN / WAN

client media
server

3 4

Fibre Channel/SAN

Robot on
SAN
Disks of client
data on SAN

1. On LAN, client backup is initiated by master server, which tells the NetBackup client to
map the snapshot data on the disk.
2. On LAN, client sends the mapping information to the media server.
3. Media server processes the backup and reads client data over the SAN, from the
addresses specified by the client.
4. Media server writes data across the SAN to storage.

14 NetBackup Advanced Client System Administrators Guide

 Offhost Backup Methods

Third-Party Copy Device (UNIX Only)

In this backup method, a third-party copy device performs the I/O processing of the
backup (data movement). This method is not supported for Windows clients.
Choose this option if processing time on the NetBackup client is critical and off-loading
backup processing to a third-party copy device may save time.
Figure Third-Party Copy shows the basic components used in the Third-Party Copy
Device method. A third-party copy device handles the backup data movement.

Third-Party Copy

NetBackup
master
server
1 LAN / WAN

media
client
server

Fibre Channel/SAN

4 5
Disks of client Third-party robot on
data on SAN copy device SAN
SCSI

1. On LAN, client backup is initiated by master server, which tells the client to map the
snapshot data.
2. On LAN, client sends the mapping information to the media server.
3. Media server sends third-party copy commands to the third-party copy device over the SAN.
4. Third-party copy device reads the client data from either SAN-attached or SCSI-attached disk.
5. Third-party copy device writes data to SAN-attached or SCSI-attached storage.

Chapter 1, Introduction 15
Offhost Backup Without a SAN (UNIX Only)

Network Attached Storage

In this backup method, an NDMP (NAS) host makes a snapshot of client data and stores
the snapshot on the same NAS-attached device that contains the original (primary) data.
The snapshot can be restored by means of Instant Recovery.

Offhost Backup Without a SAN (UNIX Only)

Offhost backup does not require a fibre channel SAN. You can configure a multi-ported
SCSI disk array on a LAN or WAN (as shown below) to support a NetBackup Media
Server offhost backup. The NetBackup media server performs the data movement. This
approach is not supported for Windows clients.
NetBackup Media Server with Multi-Ported Disk Array (No SAN)

NetBackup master
server

1 LAN / WAN

2
3 4

SCSI
SCSI

client media server Storage

Disk array

1. Client backup is initiated by master server, which tells the NetBackup client to map the
snapshot data on the disk.
2. Client sends the mapping information to the media server.
3. Media server processes the backup and reads client data from the addresses specified
by the client.
4. Media server writes data to storage.

16 NetBackup Advanced Client System Administrators Guide

 Offhost Backup with Multi-Ported Array (UNIX Only)

Offhost Backup with Multi-Ported Array (UNIX Only)

A multi-ported disk array can be combined with fibre channel to support either a
NetBackup Media Server or Third-Party Copy Device backup. The diagram below shows
the following placement options for the disk array:
On the LAN using SCSI connections to the NetBackup client and media server
On the SAN behind a third-party copy device
Directly attached to the SAN

NetBackup master
server

LAN / WAN

media
server
Client
SCSI SCSI

Disk array Storage

Fibre Channel/SAN

Third-party Storage
copy device on SAN

SCSI

Disk array
on SAN

Disk array behind

third-party copy device

Chapter 1, Introduction 17
Features and Required Software

Features and Required Software

The following table shows the types of backup you can configure with Advanced Client,
and the corresponding NetBackup features and requirements. The table in the next section
shows the software and hardware requirements for each snapshot method.

Note Advanced Client supports the following policy types: DB2, FlashBackup,
FlashBackup-Windows, MS-Exchange, MS-SQL-Server, MS-Windows-NT, Oracle,
and Standard.

Advanced Client Features and Requirements

Type of Backup NetBackup Advanced NetBackup Software Required Other Software

Client Feature Required
Client and Advanced Client not NetBackup 5.1.
storage devices required.
on local network,
and snapshot not
required.
Windows client Advanced Client not NetBackup 5.1.
and storage required.
devices on local Snapshot method:
network, open VSS, VSP (included in
file backup base NetBackup).
required.
Client and Snapshot method: Advanced Client Requirements
storage devices nbu_snap, Additional NetBackup database depend on the
on local network, VxFS_Snapshot, software needed for backing up snapshot method
snapshot VxFS_Checkpoint, clients in a database. (see next table).
required. vxvm, VVR. For For Oracle clients,
particular arrays: Oracle8i or later must
Wider selection
TimeFinder, be installed.
of snapshots
ShadowImage,
needed for UNIX
BusinessCopy. See
or Windows
next table.
clients.
Optional policy
attribute:
Retain snapshots for
instant recovery.

18 NetBackup Advanced Client System Administrators Guide

 Features and Required Software

Advanced Client Features and Requirements

Type of Backup NetBackup Advanced NetBackup Software Required Other Software

Client Feature Required
Local client Snapshot method: Advanced Client Requirements
shares data with FlashSnap, VVR, Additional NetBackup database depend on the
alternate client TimeFinder, software needed for backing up snapshot method
(split-mirror or ShadowImage, clients in a database. (see next table).
replication). BusinessCopy. See For Oracle clients,
next table. Oracle8i or later must
Optional policy be installed.
attribute:
Retain snapshots for
instant recovery.
Client and Snapshot method: Advanced Client Requirements
storage devices nbu_snap, Additional NetBackup database depend on the
on SAN; VxFS_Checkpoint, software needed for backing up snapshot method
third-party copy vxvm. For particular clients in a database. (see next table).
device available. arrays: TimeFinder, For Oracle clients,
ShadowImage, Oracle8i or later must
NOTE: Not
BusinessCopy. See be installed.
supported on
next table.
Windows clients.
Data mover:
Third-Party Copy
Device.
Optional policy
attribute:
Retain snapshots for
instant recovery.
Client and Snapshot method: Advanced Client Requirements
storage devices nbu_snap, Additional NetBackup database depend on the
on SAN; VxFS_Checkpoint, software needed for backing up snapshot method
third-party copy vxvm. For particular clients in a database. (see next table).
device not arrays: TimeFinder, For Oracle clients,
available. ShadowImage, Oracle8i or later must
BusinessCopy. See be installed.
NOTE: Not
next table.
supported on
Windows clients. Data mover:
NetBackup Media
Server.
Optional policy
attribute:
Retain snapshots for
instant recovery.

Chapter 1, Introduction 19

Features and Required Software

Advanced Client Features and Requirements

Type of Backup NetBackup Advanced NetBackup Software Required Other Software

Client Feature Required
Client data Snapshot method: Advanced Client See next table under
stored on NDMP NAS_Snapshot. NetBackup for NDMP NAS_Snapshot.
(NAS) host Data mover: Network
Attached Storage.
Additional required
policy attributes:
Retain snapshots for
instant recovery; and
Instant recovery
backups to disk only
on Schedule -
Attributes tab

20 NetBackup Advanced Client System Administrators Guide

 Features and Required Software

Snapshot Methods and Requirements

The following table describes each snapshot method and its requirements.

Note Advanced Client snapshot methods and offhost backup methods perform mapping
of the underlying file system and volume structure. This mapping has been verified
for the I/O system components listed in this table under Data Type Supported.
The use of other components in the I/O system, such as other volume managers or
storage replicators, may result in an unreliable backup. Such configurations are not
supported. For an updated list of supported configurations, see Advanced Client
Information on the Web on page xix.

Snapshot Methods and Requirements

Snapshot Type Data Type Supported Requirements and Where to Obtain

Method (for Backup Restrictions Required Software
Selections list)

nbu_snap Copy-on-write File systems: Advanced Client www.veritas.com

(Solaris only) UFS (Solaris) Does not support
clustered file systems
VxFS
Raw partitions:
VxVM volumes
raw disks2

VxFS_Checkpoint Copy-on-write File systems: Advanced Client VxFS 3.4 is part of

(Solaris or HP) (Storage VxFS VxFS 3.4 or later (for VERITAS Database
Checkpoint) HP, VxFS 3.5), with Edition for Oracle
Online JFS (HP)
Storage Checkpoint 2.2:
Supports Instant license. www.veritas.com
Recovery
Note: Does not support
raw-partition type
backups (whether
FlashBackup or Standard
policy).

VxFS_Snapshot Copy-on-write File systems: Advanced Client www.veritas.com

(Solaris or HP) UFS (Solaris) VxFS 3.4 or later (for
HP, VxFS 3.5)
(For FlashBackup Online JFS (HP)
policies only) VxFS
Raw partitions:
VxVM volumes
raw disks2

Chapter 1, Introduction 21
Features and Required Software

Snapshot Methods and Requirements

Snapshot Type Data Type Supported Requirements and Where to Obtain

Method (for Backup Restrictions Required Software
Selections list)

vxvm Snapshot File systems: Advanced Client www.veritas.com

(Solaris, HP, mirror NTFS (Windows) VxVM 3.1 or later1
Windows UFS (Solaris) Note: vxvm cannot be
2000/2003) used with VxVM volumes
VxFS
Supports Instant configured as RAID-5.
Online JFS (HP)
Recovery Note: The snapshot source
Raw partitions: must be configured with a
VxVM volumes VxVM 3.1 or later1
snapshot mirror (see
Creating a Snapshot
Mirror of the Source on
page 141).

FlashSnap Snapshot File systems: Advanced Client www.veritas.com

(Solaris, HP, mirror for NTFS (Windows) VxVM 3.2 or later with
Windows alternate client additional FlashSnap
UFS (Solaris)
2000/2003) license. (For Windows,
VxFS VxVM 3.1 or later1.)
Online JFS (HP)
Raw partitions:
VxVM volumes

VVR Snapshot File systems: Advanced Client www.veritas.com

(Solaris, HP) mirror for UFS (Solaris) VxVM 3.2 or later, with
alternate client additional VVR
Supports Instant VxFS
license.
Recovery Online JFS (HP)
Raw partitions:
VxVM volumes

TimeFinder Mirror File systems: Advanced Client www.veritas.com

(Solaris, HP) UFS (Solaris) EMC TimeFinder Contact EMC for
Symapi Symapi
VxFS
Note: TimeFinder
Online JFS (HP)
snapshot method is for
Raw partitions: EMC Symmetrix disk
VxVM volumes arrays only.
raw disks2

22 NetBackup Advanced Client System Administrators Guide

 Features and Required Software

Snapshot Methods and Requirements

Snapshot Type Data Type Supported Requirements and Where to Obtain

Method (for Backup Restrictions Required Software
Selections list)

ShadowImage Mirror File systems: Advanced Client www.veritas.com

(Solaris, HP) UFS (Solaris) ShadowImage Contact Hitachi
(HOMRCF) Data Systems for
VxFS
Note: ShadowImage Hitachi Raid
Online JFS (HP)
snapshot method is for Manager
Raw partitions: Hitachi Data Systems disk
VxVM volumes arrays only.
raw disks2

BusinessCopy Mirror File systems: Advanced Client www.veritas.com

(Solaris, HP) UFS (Solaris) Business Copy Services Contact Hewlett
- ShadowImage Packard for
VxFS
(HOMRCF) Business Copy
Online JFS (HP)
Note: BusinessCopy Services
Raw partitions: snapshot method is for HP
VxVM volumes XP series disk arrays only.
raw disks2

NAS_Snapshot NAS vendor File systems: Advanced Client www.veritas.com

specific NAS vendor specific NetBackup for NDMP

1. For Windows, VxVM 3.1 or later is required, with all the latest VxVM service packs and
updates.
2. Supported raw disks are SCSI (local or fibre channel attached), with sd, dad, and ssd
drivers (Solaris) or sdisk drivers (HP).

Note The VSP and VSS snapshot types, for Windows Open File Backup, are included
with the base NetBackup product, not Advanced Client.

Chapter 1, Introduction 23
Requirements

Requirements
NetBackup Advanced Client requires the following components:
A master server with NetBackup Advanced Client server software installed.
Clients running Solaris 7, 8, or 9, HP 11.00 or 11i, or Windows 2000 or 2003, with
NetBackup Advanced Client software installed.

Note Certain operating system and device patches (such as for the host bus adapter) may
be required for both servers and clients. To obtain the latest information, refer to
Advanced Client Information on the Web on page xix.

Please note the following additional requirements:

For the VxFS_Checkpoint snapshot method, all clients must have VxFS 3.4 or later
with the Storage Checkpoints feature.
To use Advanced Client to back up a VxFS file system, the VxFS file system on the
client has to be patched with the correct dynamic linked libraries.
For the vxvm snapshot method, all clients must have VxVM 3.1 or later.
For the FlashSnap and VVR snapshot methods, all clients must have VxVM 3.2 or
later. Each method requires its own add-on license to VxVM.
For the TimeFinder, ShadowImage, or BusinessCopy snapshot methods, assistance
may be required from the disk array vendor. Refer to Array-Specific Snapshot
Methods on page 150.
To use the snapshot and offhost backup features of NetBackup Advanced Client with
a NetBackup Oracle policy, UNIX clients must have Oracle8i or later installed.
To use the snapshot and offhost backup feature of NetBackup Advanced Client for HP
clients, the HP client must be using the Online JFS file system, not the default JFS.

Restrictions
For a complete list of supported peripherals, and for other operational notes, refer to the
NetBackup Release Notes, or to Advanced Client Information on the Web on page xix.
Note the following restrictions:
Advanced Client does not support the ALL_LOCAL_DRIVES entry in the policys
Backup Selections list.
Although the NetBackup 5.1 Administration Console can be used to administer a
NetBackup 4.5 server, please note:

24 NetBackup Advanced Client System Administrators Guide

 Requirements

The Advanced Snapshot Options button is disabled and snapshot configuration

is done by means of the policy Clients tab > Snapshot Backups button. Use
alternate client is also disabled.
The Windows version of the Administration Console cannot configure snapshot
methods, and cannot configure Instant Recovery with offhost backup.
For NetBackup Media Server or Third-Party Copy Device backup method, the disk
containing the clients data (the files to back up) must be either a SCSI or Fibre
Channel device.
For offhost backup using a data mover with the nbu_snap, VxFS_Checkpoint, or
vxvm snapshot methods, the NetBackup media server must be able to access all the
disks that make up the snapshot image. The disk(s) can be connected to a SAN. For
each of these snapshot methods, note the following:
nbu_snap: media server requires access to the active disk and the cache disk.
VxFS_Checkpoint: media server requires access to the primary or active disk.
vxvm: access requirements depend on layout of the volume group. Media server
must be able to access all disks that make up the snap mirror volume.
For the TimeFinder, ShadowImage, or BusinessCopy snapshot methods (when using
the NetBackup Media Server or Third-Party Copy Device backup methods), the
NetBackup clients must be able to access the mirror (secondary) disk containing the
snapshot of the clients data. The NetBackup clients must also be able to access the
primary disk. The NetBackup media server only needs access to the mirror
(secondary) disk.
For the TimeFinder, ShadowImage, or BusinessCopy snapshot methods, a Volume
Manager disk group must consist of disks that are all made by the same vendor.
For the NetBackup Media Server or Third-Party Copy Device backup method, the
disk must be able to return its SCSI serial number in response to a serial-number
inquiry (serialization), or the disk must support SCSI Inquiry Page Code 83.
Multiplexing is not supported for Third-Party Copy Device offhost backups.
For alternate client backup, the user and group identification numbers (UIDs and
GIDs) for the files to be backed up must be available to both hosts (the primary client
and the alternate backup client).
Inline Tape Copies (called Multiple Copies in Vault) is not supported for Third-Party
Copy Device offhost backups.
For media servers running AIX (4.3.3 and higher), note the following:
Clients must be Solaris or HP-UX.
Supports third-party copy device backup only.
Requires the use of tape LUNs to send the Extended Copy commands for backup.

Chapter 1, Introduction 25
Terminology

The tape must be behind a third-party-copy-capable FC-to-SCSI router, and the

router must be able to intercept Extended Copy commands sent to the tape LUNs.
The mover.conf file must have a tape path defined, not a controller path.

Terminology
This section introduces terms used with NetBackup Advanced Client. For explanations of
other NetBackup terms, consult the NetBackup online glossary. For instructions, see the
Glossary section in the preface.

Alternate Client Backup

A NetBackup client that performs backup processing on behalf of another client.

Backup agent (see also Third-Party Copy Device)

A general term for the host that manages the backup on behalf of the NetBackup client.
This is either another client, the NetBackup media server, a third-party copy device, or a
NAS filer.

BCV

The mirror disk in an EMC primary-mirror array configuration (see mirror). BCV stands
for Business Continuance Volume.

Bridge

In a SAN network, a bridge connects SCSI devices to Fibre Channel. A third-party copy
device can be implemented as part of a bridge or as part of other devices. Note that not all
bridges function as third-party copy devices.

BusinessCopy

One of several snapshot methods included in Advanced Client. BusinessCopy is for

making snapshots of client data on HP disk arrays.

Copy manager (see Third-Party Copy Device)

Copy-on-Write

In NetBackup Advanced Client, one of two types of supported snapshots (see also mirror).
Unlike a mirror, a copy-on-write does not create a separate copy of the clients data. It
creates a block-by-block account that describes which blocks in the client data have
changed and which have not, from the instant the copy-on-write was activated. This
account is used by the backup application to create the backup copy.

Data movement

A copy operation as performed by a third-party copy device or NetBackup media server.

26 NetBackup Advanced Client System Administrators Guide

 Terminology

Data mover

The host or entity that manages the backup on behalf of the NetBackup client. This is
either the NetBackup media server, a third-party copy device, or a NAS filer.

Disk group

A configuration of disks to create a primary-mirror association, using commands unique

to the disks vendor. See mirror and volume group.

Extent

A contiguous set of disk blocks allocated for a file and represented by three values: device
identifier, starting block address (offset in the device) and length (number of contiguous
blocks). The mapping methods in Advanced Client determine the list of extents and send
the list to the backup agent.

FastResync

Formerly known as Fast Mirror Resynchronization or FMR, FastResync performs quick

and efficient resynchronization of mirrors. NetBackups Instant Recovery feature utilizes
FastResync to create and maintain a point-in-time copy of a production volume.

Fibre channel

A type of high-speed network composed of either optical or copper cable and employing
the Fibre Channel protocol. NetBackup Advanced Client supports both arbitrated loop
and switched fabric (switched fibre channel) environments.

File system

Has two different meanings.

When referring to a product, such as the UFS (Sun Solaris) or VxFS (VERITAS) file
system, it refers to the management and allocation schemes on which the entire file
tree is structured.
When referring to a particular component in a file tree, file system means a directory
(with any subdirectories and files) that is attached to the UNIX file tree by means of
the mount command. When a file system is selected as an entry in the NetBackup
Backup Selections list, this definition applies.

Instant Recovery

A restore feature of a disk snapshot of a client file system or volume. Client data can be
rapidly restored from this disk image, even after a system reboot.

Mapping

The process of converting a file or raw device (in the file system or Volume Manager) to
absolute physical disk addresses or extents for use by backup agents on the network.
NetBackup Advanced Client uses the VxMS library to perform file mapping.

Chapter 1, Introduction 27
Terminology

Mapping methods

A set of routines for converting logical file addresses to absolute physical disk addresses
or extents. NetBackup Advanced Client includes support for file-mapping and
volume-mapping methods.

Mirror

A disk that maintains an exact copy or duplicate of another disk. A mirror disk is
often called a secondary, and the disk that it copies is called the primary. All writes to
the primary disk are also made to the mirror (or secondary) disk.
A type of snapshot captured on a mirror disk. At an appropriate moment, all further
writes to the primary disk are held back from the mirror, thus causing the mirror to be
split from the primary. As a result of the split, the mirror becomes a snapshot of the
primary. The snapshot can then be backed up.

NetBackup Media Server method

An offhost backup method in which data movement is performed by a NetBackup media

server.

Offhost backup

The off-loading of backup processing to a separate backup agent executing on another

host. NetBackup Advanced Client provides the following offhost backup options:
alternate client backup, NetBackup Media Server, Third-Party Copy Device, and Network
Attached Storage.

Primary disk

In a primary-mirror array configuration, the primary is the disk on which client data is
stored, and which is directly accessed by client applications. An exact duplicate of the
primary disk is the mirror.

Raw partition

A single section of a raw physical disk device occupying a range of disk sectors, without a
file system or other hierarchical organization scheme (thus, a raw stream of disk
sectors). This is different from a block device, over which the file system is mounted.

Recovery Manager (RMAN)

Oracle's backup and recovery program. RMAN performs backup and restore by making
requests to a NetBackup shared library.

RMAN Proxy Copy

An extension to the Oracle8i Media Management API which enables media management
software such as NetBackup to perform data transfer directly.

SAN (Storage Area Network)

28 NetBackup Advanced Client System Administrators Guide

 Terminology

A Fibre Channel-based network connecting servers and storage devices. The storage
devices are not attached to servers but to the network itself, and are visible to all servers
on the network.

Secondary disk
See mirror.

ShadowImage

One of several snapshot methods included in Advanced Client. ShadowImage is for

making snapshots of client data on Hitachi disk arrays.

Snapshot

A stable disk copy of the data prior to backup. A snapshot is created very rapidly, causing
minimal impact on other applications. There are two basic types: copy-on-write and mirror.

Snapshot method

A set of routines for creating a snapshot. You can select the method, or let NetBackup
select it when the backup is started (auto method).

Snapshot mirror

A disk mirror created by the VERITAS Volume Manager (VxVM). This is an exact copy of
a primary volume at a particular moment, reproduced on a physically separate device.

Snapshot source

The entity (file system, raw partition, or logical volume) to which a snapshot method is
applied. NetBackup automatically selects the snapshot source based on the entries in the
policys Backup Selections list.

Snapshot Volume

A mirror that has been split from the primary volume or device, and made available to

users. Snapshot volumes are created by the VERITAS Volume Manager (VxVM) as a

point-in-time copy of the primary volume. Subsequent changes made to the primary

volume are recorded in the Data Change Log and can be used later to resynchronize with

the primary volume by means of VxVM FastResync. Only the changes made while the

snapshot volume was detached from the primary would be applied to the snapshot

volume to make it identical to the primary volume again.

Standard device

Refers to the primary disk in an EMC primary-mirror disk array (see primary disk).

Chapter 1, Introduction 29
Terminology

Storage Checkpoint

Provides a consistent and stable view of a file system image and keeps track of modified
data blocks since the last checkpoint. Unlike a mirror, a Storage Checkpoint does not
create a separate copy of the primary or original data. It creates a block-by-block account
that describes which blocks in the original data have changed and which have not, from
the instant the checkpoint was activated.
A Storage Checkpoint stores its information in available space on the primary file system,
not on a separate or designated device. (Also, the ls command does not list Storage
Checkpoint disk usage; you must use the fsckptadm list command instead.)

Third-Party Copy Device

This term has two meanings:

A backup agent on the SAN that operates on behalf of backup applications. The
third-party copy device receives backup data from a disk attached to Fibre Channel and
sends it to a storage device, using the SCSI Extended Copy command. The third-party
copy device is sometimes called a copy manager, third-party copy engine, or data
mover. In SAN hardware configurations, a third-party copy device can be
implemented as part of a bridge, router, or storage device. The third-party copy
device may or may not be the device to which the storage units are connected.
An offhost backup method in NetBackup Advanced Client that allows backups to be
made by means of a backup agent on the SAN.

TimeFinder

One of several snapshot methods included in Advanced Client. TimeFinder is for making
snapshots of client data on EMC Symmetrix disk arrays.

UFS file system

This is the UNIX File System (UFS), which is the default file system type on Sun Solaris.
The UFS file system was formerly the Berkeley Fast File System.

VxMS (VERITAS Federated Mapping Services)

A library of routines (methods) used by NetBackup Advanced Client to obtain the

physical addresses of logical disk objects such as files and volumes.

Volume

A virtual device configured over raw physical disk devices (not to be confused with a
NetBackup Media Manager volume). Consists of a block and character device.
If a snapshot source exists over a volume, NetBackup automatically uses a volume
mapping method to map the volume to physical device addresses. Any of the Advanced
Client snapshot methods can be used when backing up client data configured over volumes.
For NetBackup, volumes must be created by means of the VERITAS Volume Manager
(VxVM).

30 NetBackup Advanced Client System Administrators Guide

 Terminology

Volume group

A logical grouping of disks, created with the VERITAS Volume Manager, to allow more
efficient use of disk space.

VxFS

The VERITAS extent-based File System (VxFS), designed for high performance and large
volumes of data.

VxVM

The VERITAS Volume Manager (VxVM), which provides logical volume management
that can be used in SAN environments.

Chapter 1, Introduction 31
Terminology

32 NetBackup Advanced Client System Administrators Guide

Installation 2

This chapter explains how to install NetBackup Advanced Client software on UNIX and
Windows platforms.

Prerequisites
NetBackup Enterprise server 5.1 or later must be installed on the master/media
servers. For performing local backups, the master/media server can be running any
supported UNIX or Windows platform. For NetBackup Media Server or Third-Party
Copy Device backups, the NetBackup media server must be installed on Solaris 7, 8,
or 9, or HP-UX 11.00 or 11i.
For a detailed list of platform versions supported by NetBackup Advanced Client,
refer to the NetBackup Release Notes, or to Advanced Client Information on the Web on
page xix.
NetBackup 5.1 or later client software must be installed on clients running Solaris 7, 8,
or 9, HP-UX 11.00 or 11i, or Windows 2000 or 2003.
For Instant Recovery using the VxFS_Checkpoint method, the VxFS File System with
the Storage Checkpoints feature must be installed on clients.

33
Installing Advanced Client On UNIX

Installing Advanced Client On UNIX

Loading From Media

Note If you are installing in a cluster environment, you must freeze the active node before
you begin the installation process so that migrations do not occur during
installation. For information about freezing a service group, see the clustering
section in the NetBackup High Availability System Administrators Guide for the cluster
software you are running.

1. Log in as root on the NetBackup master/media server.

2. In a separate window, make sure a valid license key for NetBackup Advanced Client
has been installed. To do this, enter the following command to list and add keys:
/usr/openv/netbackup/bin/admincmd/get_license_key

3. Insert the options CD-ROM containing the Advanced Client software in the drive.

4. Change your working directory to the CD-ROM directory:

cd /cd_rom_directory
Where cd_rom_directory is the path to the directory where you can access the
CD-ROM. On some platforms, it may be necessary to mount this directory.

5. To install Advanced Client software on the NetBackup master server, execute the
following:
./install

Since other NetBackup products are included on the CD-ROM, a menu appears.

6. Select NetBackup Add-On Product Software.

A second menu appears. Select Advanced Client.

7. Enter q to quit selecting options. When asked if the list is correct, answer y.
NetBackup Advanced Client software is installed in
/usr/openv/netbackup/vfms/hardware/os/version/
Where:
hardware is Solaris, HP9000-700, or HP9000-800
os is Solaris7, Solaris8, Solaris9, or HP-UX11.00 or HP-UX11.11

34 NetBackup Advanced Client System Administrators Guide

 Installing Advanced Client On UNIX

version is a six digit number representing the NetBackup version

8. In a clustered environment, the above steps must be done on each node in the cluster.

Note If you are installing in a cluster environment, unfreeze the active node after the
installation completes. For information about unfreezing a service group, see the
clustering section in the NetBackup High Availability System Administrators Guide for the
cluster software you are running

Note The NetBackup 5.1 client software must be installed on the clients before
performing the next procedure. For instructions, refer to the NetBackup Installation
Guide for UNIX.

Distributing Advanced Client Software to UNIX Clients

Note If the master server is a Solaris or HP system and the master server is also a client,
you do not have to distribute Advanced Client software to the master. This
distribution is done automatically when you load the software from the media.
However, you must distribute the software as explained below to all other UNIX
clients that will be using Advanced Client.

Note If installing in a clustered environment, you must do the following on the active
node of the cluster.

You should also perform this procedure if you are doing either of the following:

Installing a software upgrade (patch)

Reinstalling Advanced Client software on a Solaris client

Execute the following as the root user on the NetBackup master server.

1. Check whether bprd is running by executing:

/usr/openv/netbackup/bin/bpps
If only one bprd shows up in the bpps output, there are no active backups or restores.
In that case, terminate the bprd daemon by executing:
/usr/openv/netbackup/bin/admincmd/bprdreq -terminate

If more than one bprd appears, wait until the backups and/or restores are complete
and then run the /usr/openv/netbackup/bin/bpps command again. When only
one bprd shows up, terminate the bprd daemon.

Chapter 2, Installation 35
Installing Advanced Client On UNIX

2. You can distribute the Advanced Client software to Solaris and HP clients in either of
two ways:

Note If you are distributing the Advanced Client software to clients located in a cluster,
specify the host names of the individual nodes (not virtual names) in the list of
clients.

a. Distribute the software to all currently defined clients by executing the following
command:
/usr/openv/netbackup/bin/update_clients -Install_ADC

b. Distribute the software to specific clients.

Create a file that lists the specific clients. For each client, enter a line in this file
containing the following three parameters:
hardware_type operating_system clientname

For example:
Solaris Solaris7 othersparc

or
Solaris Solaris8 othersparc

or
Solaris Solaris9 othersparc

or
HP9000-800 HP-UX11.11 myhp

In each backup policy, clients must be configured with their correct OS type.
For example, an HP client running HP-UX11.11 must be configured as such.
Execute the following command (all on one line):
/usr/openv/netbackup/bin/update_clients -Install_ADC -ClientList file

Where file is the name of the file that you created in the previous step.

Note Advanced Client software cannot be distributed to clients by means of the

NetBackup Administration Console.

3. Start the NetBackup daemon as the root user on the master server by executing:
/usr/openv/netbackup/bin/initbprd

36 NetBackup Advanced Client System Administrators Guide

 Installing Advanced Client On Windows

Installing Advanced Client On Windows

For Windows, NetBackup Advanced Client software is automatically installed with the
core NetBackup server and client product. For the NetBackup installation procedure, see
the NetBackup Installation Guide for Windows.
You must add a valid license key for Advanced Client, as follows.

On each Windows 2000 or 2003 server:

Note If you are installing in a cluster environment, you must freeze the active node before
you begin the installation process so that migrations do not occur during
installation. For information about freezing a service group, see the clustering
section in the NetBackup High Availability System Administrators Guide for the cluster
software you are running.

1. Log in.

2. Add a license key for Advanced Client:

a. From the NetBackup Administration window, choose Help.

b. From the Help menu, select License Keys ....

The NetBackup License Keys window appears. Existing keys are listed in the
lower part of the window.

c. To register a new key, click the star icon to open the Add a New License Key
dialog. Type the new license key in the New license key field and click Add.
The new license key appears in the lower part of the dialog box.

3. In a clustered environment, the above steps must be done on each node in the cluster.

Note If you are installing in a cluster environment, unfreeze the active node after the
installation completes. For information about unfreezing a service group, see the
clustering section in the NetBackup High Availability System Administrators Guide for the
cluster software you are running.

Chapter 2, Installation 37
Distributing Client Software in Mixed-Platform Environments

Distributing Client Software in Mixed-Platform

Environments
NetBackup Advanced Client software can be pushed from server to client in an
all-UNIX environment only: server and clients must be running UNIX. Please see
Distributing Advanced Client Software to UNIX Clients on page 35.
For Windows PCs, Advanced Client software is automatically installed with the base
NetBackup software, for server and clients. Please see the NetBackup Installation Guide for
Windows, and Installing Advanced Client On Windows on page 37 of this Advanced
Client guide.
For mixed environments, please note the following:
If you have a Windows server with UNIX clients (no UNIX server), you must set up a
NetBackup UNIX server to distribute the software to your clients.
If you have a UNIX server with Windows clients, you must install the client software
on the Windows client machines individually, from the NetBackup CD-ROM. Please
see Installing Advanced Client On Windows on page 37.

38 NetBackup Advanced Client System Administrators Guide

 Creating Log Directories

Creating Log Directories

During backup and restore, Advanced Client messages are written to several log
directories on the NetBackup server and client. You must create these directories
manually. Refer to Gathering Information and Checking Logs on page 200 for help
creating log directories.

Upgrading from Earlier Releases

During Advanced Client installation, NetBackup automatically updates existing policies
from the format used in earlier releases to that used by Advanced Client.

Conversion from 4.5 FP3 Feature Pack

In NetBackup 4.5 Feature Pack, snapshot (frozen image) configuration was done per client
per policy. Only one frozen image method could be designated for a particular client in a
particular policy. Another client, however, could be included in the policy with the same
frozen image method or a different method.
When Advanced Client is installed, 4.5 feature pack policies are changed as follows:
Policies are converted to one snapshot method per policy. Clients configured with the
same snapshot method are grouped together into a new policy, such that only one
snapshot method is used in that policy. Existing policies that use only one snapshot
method are left as they are.
Numbers are added to the end of the new policy names:
old_policy_name_number
The original policy is marked inactive and set to a snapshot method of auto.

Conversion from 4.5 GA

Note At NetBackup 4.5 GA, a frozen image source had to be specified in order to create
a policy for snapshots. Starting with the 4.5 FP3 feature pack, NetBackup
automatically selects the snapshot source (frozen image source) based on your
Backup Selections list entries.

In NetBackup 4.5 GA and 4.5 maintenance packs, snapshot (frozen image) configuration
was client based: a single frozen image configuration file applied to all policies backing up
the client. There was no limitation on the number of frozen image methods used within a

Chapter 2, Installation 39
Uninstalling NetBackup Advanced Client or Earlier Packages

policy to back up a client. This meant that a NetBackup client could be configured with
multiple snapshot methods within the same policy, one method for each snapshot source
designated on the 4.5 Frozen Image Client Configuration display.
When Advanced Client is installed, these 4.5 GA policies are changed as follows:
The snapshot (frozen image) methods specified for the clients in a policy are replaced
with the auto method. The auto method means that NetBackup selects an
appropriate snapshot method at the start of the backup.

To see how your existing policies were configured

Before or after installing Advanced Client, you can view the configuration details of your
policies, for each client, as they existed prior to Advanced Client installation. Go to:
/usr/openv/netbackup/db/client/client_name

and list (ls) the contents of the client_name directory. The configuration information is
contained in files named fi_info.

Uninstalling NetBackup Advanced Client or Earlier

Packages

Server Uninstall - UNIX

On the master server where you initially loaded the NetBackup Advanced Client or other
software, do the following.

Note The following procedure results in total removal of the Advanced Client software.

Note If you are uninstalling in a cluster environment, you must first freeze the active
node so that migrations do not occur during the uninstall. For information about
freezing a service group, see the clustering section in the NetBackup High Availability
System Administrators Guide for the cluster software you are running.

On Solaris:

1. Check the Activity Monitor in the NetBackup Administration Console to make sure
no NetBackup Advanced Client (or pre-5.0 ServerFree Agent or FlashBackup)
backups are active or running (the State field should read Done).

2. To remove the NetBackup Advanced Client package, execute the following:

40 NetBackup Advanced Client System Administrators Guide

 Uninstalling NetBackup Advanced Client or Earlier Packages

pkgrm VRTSnbadc

3. To remove ServerFree Agent packages installed prior to 5.0, execute the following:
For the Offhost and SAN Data Movement Services product:
pkgrm VRTSnbodm
For the Extended Frozen Image Services product:
pkgrm VRTSnbefi
For the Core Frozen Image Services product:
pkgrm VRTSnbfis
For the FlashBackup product:
pkgrm VRTSnbfsh

4. See the note on cluster environments, below.

On HP:

1. Check the Activity Monitor in the NetBackup Administration Console to make sure
no NetBackup Advanced Client (or pre-5.0 ServerFree Agent or FlashBackup)
backups are active or running (the State field should read Done).

2. Execute the following:

/usr/openv/netbackup/bin/install_adc -d
rm -rf /usr/openv/netbackup/vfms

Note If you are uninstalling in a cluster environment, unfreeze the active node after the
uninstall. For information about unfreezing a service group, see the clustering
section in the NetBackup High Availability System Administrators Guide for the cluster
software you are running.

Client Uninstall - UNIX

1. On the master server, check the Activity Monitor in the NetBackup Administration
Console to make sure no NetBackup Advanced Client backups are active or running
for the client (the State field should read Done).

2. Execute the following command to deinstall the NetBackup Advanced Client

software on the client:
/usr/openv/netbackup/bin/install_adc -d

Chapter 2, Installation 41
Uninstalling NetBackup Advanced Client or Earlier Packages

Server/Client Uninstall - Windows

1. On the master server, check the Activity Monitor in the NetBackup Administration
Console to make sure no NetBackup Advanced Client backups are active or running
for the client (the Job State field should read Done).

2. Perform the basic uninstall procedure described in the NetBackup Installation Guide for
Windows.

42 NetBackup Advanced Client System Administrators Guide

SAN Configuration for Advanced Client
3
Due to the complex and rapidly changing nature of SAN configuration, this chapter
describes SAN issues pertaining to NetBackup Advanced Client only. Please note the
following assumptions:
You have considerable technical expertise in both SAN and NetBackup
configuration.
Your hardware environment is already configured and functional, including
switches, hubs, optional bridges or third-party copy devices, robots, tape drives,
and disk arrays.

Note This chapter applies to the NetBackup Media Server and Third-Party Copy Device
backup methods only. If your backup policies are not using either of these methods,
you may skip this chapter.

This chapter includes the following topics:

SAN Configuration Diagram

Media Server/Third-Party Copy Requirements

Configuration Flowcharts

Verify NetBackup Access to SAN Devices

Solaris only: Configure HBA Drivers

Create Backup Configuration Files

43
SAN Configuration Diagram

SAN Configuration Diagram

The following diagram shows the devices and configuration files described by the
procedures in this chapter. This diagram shows devices configured behind a third-party
copy device (bridge) as well as directly attached to the SAN.

Backup configuration NetBackup NetBackup

files on media server: media server client

3pc.conf file: LAN / WAN

Contains tape and
client disk info used by
third-party copy device.
mover.conf file: HBA HBA
Identifies the
third-party copy device.

Fibre Channel SAN

Third-party
copy device

SCSI

Tape on
SAN
Disk array
on SAN
Tape behind
Disk array behind third-party
third-party copy copy device
device

44 NetBackup Advanced Client System Administrators Guide

 Media Server/Third-Party Copy Requirements

Supported Peripherals
A complete list of Advanced Client supported peripherals can be found on the VERITAS
support web site. For instructions, refer to Advanced Client Information on the Web on
page xix.

Media Server/Third-Party Copy Requirements

The configuration requirements for NetBackup Media Server and Third-Party Copy
Device backup methods are as follows:
The information needed for the NetBackup Media Server method is obtained by
means of the bptpcinfo command described later in this chapter.
The information needed for the Third-Party Copy Device method requires the
bptpcinfo and bpmoverinfo commands. Depending on your devices, you may
need the following:
The VERITAS SANPoint Control product for locating the world-wide port name
and luns. If you have SANPoint Control, you can use the bpSALinfo command
to look up world-wide name and lun information.
The instructions provided with your HBA and bridge/router/third-party copy
device. The VERITAS support web site also contains information to help you
configure devices (see Advanced Client Information on the Web on page xix for
instructions).

Chapter 3, SAN Configuration for Advanced Client 45

Media Server/Third-Party Copy Requirements

Diagram for NetBackup Media Server

In this backup method, the NetBackup media server handles the backup processing and
sends the backup data over Fibre Channel to the storage device.

Note If you have a multi-ported SCSI disk array, a fibre channel SAN is not required. See
Offhost Backup Without a SAN (UNIX Only) on page 16.

NetBackup Media Server

NetBackup
master server

LAN / WAN
3pc.conf file on media server:
Contains client disk information.
NetBackup NetBackup Use bptpcinfo command to
client media server create the 3pc.conf file.

Robot on
Fibre Channel SAN
SAN

Third-party
copy device
(passive*)
SCSI

Client disks on SAN

*passive means the third-party copy device allows

media server access to the disks/tapes but does not
perform SCSI Extended Copy commands.

46 NetBackup Advanced Client System Administrators Guide

 Media Server/Third-Party Copy Requirements

Diagram for Third-Party Copy Device

In this backup method, a third-party copy device performs the I/O processing of the
backup (data movement).The 3pc.conf file describes the disks to be backed up and the
tape devices for data storage. Be sure to use the flowcharts and accompanying
instructions provided under Configuration Flowcharts on page 49.

Third-Party Copy

3pc.conf file on media server:

Contains tape and client disk
information.
Use bptpcinfo command to create
the 3pc.conf file.
NetBackup
master server Use bpSALinfo command to obtain
world-wide port names from
LAN / WAN SANPoint Control.

mover.conf file on media server:

NetBackup NetBackup Identifies the third-party copy
client media server device(s).
Use bpmoverinfo to create the
mover.conf file.

Robot on
Fibre Channel SAN
SAN

Third-party
copy device
(active*)
SCSI

Client disks on SAN

*active means the third-party copy device performs SCSI Extended

Copy commands to move the data.

Chapter 3, SAN Configuration for Advanced Client 47

Media Server/Third-Party Copy Requirements

Diagram for Third-Party Copy Device - Remote

In this configuration, the NetBackup media server and the disks containing the client data
are on different fibre-channel networks (due to zoning or LUN-masking). The media
server can communicate with the NetBackup client by means of the LAN, but does not
have access to the clients disks located on a different fibre channel network (or zone). In
this case, the 3pc.conf file must be modified with the bptpcinfo command using the -x
client_name option, to include information about the client disks.

Note If all devices support identification descriptors (E4 target), NetBackup

automatically creates a complete 3pc.conf file and you do not need to run the
bptpcinfo command.

Third-Party Copy: Remote

NetBackup NetBackup On media server, use bptpcinfo

client media server command with -x client_name
option to obtain info about disks that
LAN / WAN
reside in zone 2. This allows
third-party copy device in zone 1 to
back up disks residing in zone 2.

Fibre Channel Zone 1

Switch 1

Fibre Channel
Zone 2 Third-party
Switch 2 copy device
(active)
SCSI
Client disks in
zone 1

Client Disks in
zone 2 Client Disks
in zone 1

48 NetBackup Advanced Client System Administrators Guide

 Configuration Flowcharts

Configuration Flowcharts
The following four charts show the process for setting up configuration files for Media
Server or Third-Party Copy backup. Instructions are included later in this chapter.

Chart I: Verify SAN Device Visibility

For instructions, see Verify NetBackup Access to SAN Devices on page 53.

Run bptpcinfo -a -o -

N
All OS device paths Go to Chart II.
visible?
Y
Run ioscan -nf on HP
or sgscan on Solaris

See Verify passthru device path

All passthru N
visibility on page 54, or refer to the
device paths NetBackup Media Manager Device
visible? Configuration Guide.
Y

Start filling out

NetBackup N Device Checklist
Media Server on page 58.
method?

Y SeeAdvanced Client
Have N Information on the Web on
Done. No more SANPoint page xix for help obtaining
device configuration Control? world-wide port names.
is required. Y

Enter world-wide names in the

Device Checklist on page 58.

Go to Chart III (Solaris) or Chart IV (HP).

Chapter 3, SAN Configuration for Advanced Client 49

Configuration Flowcharts

Chart II: Verify OS Device Paths Visibility

For instructions, see Making OS device paths visible on page 53.

N
Tape devices On Solaris: Correct
visible? the st.conf file

Disks N On Solaris: Correct

visible? the sd.conf file

SeeAdvanced Client
Third-party N Information on the Web on
copy device page xix for help enabling the
visible? third-party copy device.

SeeAdvanced Client
Devices behind N Information on the Web on
third-party copy page xix for help discovering
device visible? LUNs and world-wide names.

Set third-party copy device

mapping mode to FIXED.

Perform a reconfigure
reboot. Go to Chart I.

50 NetBackup Advanced Client System Administrators Guide

 Configuration Flowcharts

Chart III: Solaris only: Configure HBA Drivers

For instructions, see Solaris only: Configure HBA
Drivers on page 59.

Storage target Y
IDs bound to world- Done. Go to Chart IV.
wide name?

Refer to Device Checklist on

page 58 for SCSI target IDs (in device
pathname) and world-wide names.

Update HBA configuration file with target

world-wide port name bindings.
(SeeAdvanced Client Information on the
Web on page xix for help.)

Reboot.

Go to Chart I.

Chapter 3, SAN Configuration for Advanced Client 51

Configuration Flowcharts

Chart IV: Create the Backup Configuration Files

For instructions, see Create Backup Configuration Files on
page 60.

Using
Third-Party N Done. No further
Copy Device configuration needed.
method?

Identification
Y
descriptors (i=)
available for all
devices?

Run bptpcinfo and bpSALinfo

to create the 3pc.conf file.

If you do not have SANPoint Control, take

world-wide port name from Device
Checklist and enter (w=) in 3pc.conf file.

Run bpmoverinfo to
create the mover.conf file.

Done.

52 NetBackup Advanced Client System Administrators Guide

 Verify NetBackup Access to SAN Devices

Verify NetBackup Access to SAN Devices

Note It is assumed that NetBackup and all device drivers are installed, and that devices
are properly connected and powered up.

Verify OS device path visibility

1. On the media server, run the bptpcinfo command.

The following sends the output to the screen, using -o - (note the space before the final
hyphen).
/usr/openv/netbackup/bin/bptpcinfo -a -o -

The following sends the output to a file:

/usr/openv/netbackup/bin/bptpcinfo -a -o output_file_name

2. Examine the bptpcinfo output to see if your OS device paths are listed. If all devices
are listed, go to step 8 for HP or to step 9 for Solaris.

Making OS device paths visible

3. For Solaris: If your tape devices are not listed in the bptpcinfo output, make sure
you have target and LUN values for each tape device in the st.conf file.

4. For Solaris: If your disks are not listed in the bptpcinfo output, make sure you have
target and LUN values for each disk in the sd.conf file.

5. If the devices behind the bridge (or third-party copy device) are not listed in the
bptpcinfo output, or if the third-party copy device is not enabled for third-party
copy data movement, consult the VERITAS support website for assistance (see
Advanced Client Information on the Web on page xix).

6. On the bridge or third-party copy device, set the address mapping mode to FIXED.
This prevents the addresses from changing when the devices are reset. For help
configuring supported third-party copy devices, go www.support.veritas.com.
(See Advanced Client Information on the Web on page xix.)

7. Enter the following to reboot the operating system on the media server:
Solaris:
reboot -- -r

HP:
reboot

Chapter 3, SAN Configuration for Advanced Client 53

Verify NetBackup Access to SAN Devices

Verify passthru device path visibility

8. For HP: Enter the following to list all passthru devices:

ioscan -nf

a. If all devices now appear, enter the following to regenerate HP special files:
insf -e
Then go to step 10 on page 55.

b. If some devices do not appear in the ioscan output, check hardware connections
to the devices that are not appearing. Then repeat step 8.

Note On HP, there is a limit of eight devices per target. For instance, if you have a JBOD
disk array consisting of ten disks, and the array is connected to a bridge, it may be
that only the first eight disks in the array are accessible.

9. For Solaris:

a. Perform an sgscan to list all passthru devices. Check for proper output and
recognition of devices.
Here is sample output from sgscan:
/dev/sg/c0t6l1: Disk (/dev/rdsk/c1t6d1): "SEAGATE ST39175LW"

/dev/sg/c0t6l4: Tape (/dev/rmt/2): "QUANTUM DLT7000"

/dev/sg/c0t6l5: Changer: "HP C6280-7000"

b. If tape devices still do not show up, make sure you have entries for all SCSI target
and LUN combinations in the sg.links and sg.conf files. Refer to the Media
Manager Device Configuration Guide, Chapter 2, under Understanding the SCSI
Passthru Drivers.
If tape devices are fibre attached, make sure you have entries for the tape
devices in the above files.
If tape devices are behind a bridge (or third-party copy device), make sure
you have entries for the tape devices AND for the bridge/third-party copy
device.
For an example, refer to Solaris only: Example for sg.links, sg.conf, and st.conf
files on page 57.

54 NetBackup Advanced Client System Administrators Guide

 Verify NetBackup Access to SAN Devices

If you are unsure how to acquire the SCSI target and LUN values for your
configuration, see Advanced Client Information on the Web on page xix for
help with particular devices. For instance, if your tape drives are configured
behind a bridge, router or other fibre-channel device, you may need to telnet into
the device to determine the target ID and LUN for each tape drive.

c. When finished updating the sg.links, sg.conf, and st.conf files, remove
the old sg configuration:
rm /kernel/drv/sg.conf

rem_drv sg

d. Run the /usr/openv/volmgr/bin/driver/sg.install script to copy the

files into the correct locations.

e. Copy the sg.links and sg.conf files (in /usr/openv/volmgr/bin

/driver) to another location, for future reference. Whenever NetBackup is
re-installed, these files in /usr/openv/volmgr/bin/driver are overwritten.

10. Run the bptpcinfo command again to see which devices are now visible to the
media server. Repeat at step 2 if any of your SAN devices are not showing up in the
bptpcinfo command output.

11. If the offhost backup method is NetBackup Media Server, no more device
configuration is required. You can skip the rest of this chapter.

12. When all devices are listed in the bptpcinfo command output, use that information
to fill in the device pathname (p=), serial number (s=), and LUN (l=) in the Device
Checklist on page 58 for each device.

Determine the world-wide name for each device

There are two options:
Use VERITAS SANPoint Control (step 13).
Otherwise, refer to Advanced Client Information on the Web on page xix for help
determining the world-wide name of your devices. Then go to step 14.

13. You can use VERITAS SANPoint Control to determine the world-wide port names for
the devices.

a. Start the SANPoint Control console by entering the following:

/opt/VRTSspc/bin/spc &

Chapter 3, SAN Configuration for Advanced Client 55

Verify NetBackup Access to SAN Devices

b. Click on the Storage tab in the left pane, then click on a device in the left pane
(you may have to drill down in the tree).

c. Click the Connectivity tab to find the world-wide port name of the device (Port
WWN). Repeat these steps for each device.

14. Update the Device Checklist on page 58 with the world-wide port names of your
devices.

Note It is important to record this information! It will be needed again, to complete the
configuration.

15. For Solaris: continue with Solaris only: Configure HBA Drivers on page 59. For HP,
continue with Create Backup Configuration Files on page 60.

56 NetBackup Advanced Client System Administrators Guide

 Verify NetBackup Access to SAN Devices

Solaris only: Example for sg.links, sg.conf, and st.conf files

The following is an example for step 9 on page 54. For the devices in this example, fibre
channel LUNs 0, 1, 4, and 5 are needed for target (Loop ID) 6. In this example, LUN 0 is
the third-party copy device, LUN 1 is the tape controller, and LUNs 4 and 5 are the tape
drives.
Add entries in the /usr/openv/volmgr/bin/driver/sg.links file so that the
necessary /dev/sg/* nodes are created.
Note that the target and LUNs in the address part of the sg.links entries are
hexadecimal, but are decimal in the sg/c\N0tmln part of the entries. Also, make
sure there are tabs between the columns, not spaces.
type=ddi_pseudo;name=sg;addr=6,0; sg/c\N0t6l0

type=ddi_pseudo;name=sg;addr=6,1; sg/c\N0t6l1

type=ddi_pseudo;name=sg;addr=6,4; sg/c\N0t6l4

type=ddi_pseudo;name=sg;addr=6,5; sg/c\N0t6l5

Add additional target and LUN entries to the

/usr/openv/volmgr/bin/driver/sg.conf file.
name="sg" class="scsi" target=6 lun=0;

name="sg" class="scsi" target=6 lun=1;

name="sg" class="scsi" target=6 lun=4;

name="sg" class="scsi" target=6 lun=5;

In the /kernel/drv/st.conf file, do the following:

Add (or un-comment) the appropriate drive entries in the tape-config-list
section.
tape-config-list =

"DEC TZ89", "DEC DLT", "DLT7k-data";

Then add (un-comment) the matching data-property-name entry:

DLT7k-data = 1,0x38,0,0x39639,4,0x82,0x83,0x84,0x85,2;

For each tape drive, add a name entry to the st.conf file.
Here is an example name entry:
name="st" class="scsi" target=6 lun=4;

name="st" class="scsi" target=6 lun=5;

Make sure you have entries for all target and bus combinations for each device.

Chapter 3, SAN Configuration for Advanced Client 57

Verify NetBackup Access to SAN Devices

Device Checklist
Use this checklist or one like it to record information about each of your SAN devices.
Some of this information is provided by the bptpcinfo command (such as device
pathname and serial number), and some has to be obtained by other means as explained
in these procedures. It is vital that the information be recorded accurately.

Type of Device pathname used by UNIX Serial number LUN World-wide port name (w=)
Device host (p=) (s=) (l=)
(disk or
tape)

58 NetBackup Advanced Client System Administrators Guide

 Solaris only: Configure HBA Drivers

Solaris only: Configure HBA Drivers

Fibre channel devices should be bound to specific SCSI target IDs by modifying the driver
configuration files for your host bus adapter (HBA). This binding ensures that the host
HBA and the third-party copy device are in agreement as to the target and LUN values for
each device. The binding also ensures that the target ID does not change after a system
reboot or after a fibre-channel reconfiguration. If the target ID changes, the backup
configuration files (3pc.conf, mover.conf) will also be incorrect and will have to be
recreated.
The binding process is unique to each vendor and product. For assistance, refer to the
documentation provided for your HBA, or to the VERITAS support website. (See
Advanced Client Information on the Web on page xix.) The binding requires the
fibre-channel world-wide port name.

Note Each time a device is added or removed, the binding must be updated to reflect the
new configuration.

If storage device SCSI target IDs are bound to world-wide port names in your HBA
configuration file, skip this section and go to Create Backup Configuration Files on
page 60.

To configure HBA drivers on the media server:

1. If storage device target IDs are not already bound to world-wide port names, refer to
your Device Checklist on page 58 (filled out in the previous procedure) for the
world-wide names. Use the world-wide names to make the binding for each device.

2. Update your HBA configuration file by binding all SCSI device target IDs to their
associated world-wide port name.
For assistance with your particular HBA file, see Advanced Client Information on
the Web on page xix.

3. Reboot the media server (reboot -- -r).

4. To ensure device visibility, repeat the steps described under Verify NetBackup
Access to SAN Devices on page 53.
When you are finished, the bptpcinfo command should list device pathnames and
serial numbers for all of your devices. Update the Device Checklist with those
values if needed.

5. Continue with Create Backup Configuration Files.

Chapter 3, SAN Configuration for Advanced Client 59

Create Backup Configuration Files

Create Backup Configuration Files

This section pertains to the Third-Party Copy Device method only. If you are using any other
backup method, you can skip the rest of this chapter.
For the Third-Party Copy Device method, you must create the following file on the media
server:
/usr/openv/volmgr/database/mover.conf
If SANPoint Control does not support your environment and some devices (such as disks,
tapes, or third-party copy devices) do NOT support identification descriptors (E4 target),
you must add world-wide port name information for those devices in this file on the
media server:
/usr/openv/volmgr/database/3pc.conf

Note At the start of the backup, NetBackup creates a 3pc.conf file if one does not exist. If
all devices support identification descriptors, you do not need to create or edit the
3pc.conf file. You can skip to mover.conf Description on page 65 and to Create
the mover.conf File on page 77.

The 3pc.conf and mover.conf Files: An Overview

The NetBackup media server needs certain information about the devices available on the
SAN in order to coordinate the backup. This information is provided in two files:
3pc.conf:
Identifies the client disks on the SAN that can be backed up, and the robotic
libraries/tape drives on which NetBackup can store the data. The NetBackup media
server uses this information to access client disks when performing the backup. It also
uses this information to generate the SCSI Extended Copy commands required by
third-party copy devices.
mover.conf:
Identifies the third-party copy devices. These are devices that can execute the SCSI
Extended Copy commands. A variety of devices can be designed to operate SCSI
Extended Copy, such as routers, bridges, robotic libraries, and disk arrays. The
mover.conf file is needed for the Third-Party Copy Device backup method only, not
for the NetBackup Media Server method.

60 NetBackup Advanced Client System Administrators Guide

 Create Backup Configuration Files

3pc.conf Description
In the 3pc.conf file, each SAN device needs a one-line entry containing several kinds of
values. The values required depend on several factors (explained below). Typically, these
include (but are not limited to) the device ID, host-specific device path, and serial number.
One or more of the following are also required: the identification descriptor, logical unit
number (LUN) and world-wide port name. See Determining Requirements on page 64.
Some of this information will be automatically discovered and filled in by the bptpcinfo
command, as described under What bptpcinfo Automatically Provides on page 65. The
procedure for using the bptpcinfo command is under Create the 3pc.conf File on
page 75.

Example 3pc.conf file

Below is an example 3pc.conf file, followed by descriptions of each field. This file is
located in /usr/openv/volmgr/database.

Example 3pc.conf

# devid [a=wwpn:lun] [c=client] [p=devpath] [P=clientpath] [s=sn] [l=lun] [w=wwpn] [W=wwpn] [i=iddesc]
0 p=/dev/rdsk/c0t0d0s2 s=FUJITSU:MAB3091SSUN9.0G:01K52665 l=0
1 p=/dev/rdsk/c0t10d0s2 s=FUJITSU:MAG3091LSUN9.0G:00446161 l=0
2 p=/dev/rdsk/c4t0d0s2 s=HP:OPEN-3:30436000000 l=0 a=500060E80276E401:0
3 p=/dev/rdsk/c4t1d0s2 s=FUJITSU:MAN3367MSUN36G:01X37938 l=0 a=100000E00221C153:0
4 p=/dev/rdsk/c4t3d0s2 s=HITACHI:OPEN-3-CM:20461000000 l=0 i=10350060E800000000000004FED00000000 a=50060E80034FED00:0
5 p=/dev/rdsk/c4t14d0s2 s=HITACHI:OPEN-9:60159003900 l=0 w=500060e802eaff12
6 p=/dev/rdsk/c4t0d1s2 s=HP:OPEN-3:30436000100 l=1 a=500060E80276E401:1 a=1111222233334444:0
7 p=/dev/rdsk/c4t0d2s2 s=HP:OPEN-3:30436000200 l=2 a=500060E80276E401:2
8 p=/dev/rdsk/c4t0d3s2 s=HP:OPEN-3:30436000300 l=3 a=500060E80276E401:3
9 p=/dev/rdsk/c4t0d4s2 s=HP:OPEN-3-CM:30436005100 l=4 a=500060E80276E401:4
10 p=/dev/rdsk/c4t0d5s2 s=HP:OPEN-3:30436002600 l=5 a=500060E80276E401:5
11 p=/dev/rdsk/c4t0d6s2 s=HP:OPEN-3:30436002700 l=6 a=500060E80276E401:6
12 p=/dev/rdsk/c4t0d7s2 s=HP:OPEN-3:30436002800 l=7 a=500060E80276E401:7
13 p=/dev/rdsk/c4t0d8s2 s=HP:OPEN-3:30436002900 l=8 a=500060E80276E401:8
14 p=/dev/rdsk/c4t1d1s2 s=FUJITSU:MAN3367MSUN36G:01X37958 l=1 a=100000E00221C153:1
15 p=/dev/rdsk/c4t1d2s2 s=FUJITSU:MAN3367MSUN36G:01X38423 l=2 a=100000E00221C153:2
16 p=/dev/rdsk/c4t1d3s2 s=FUJITSU:MAN3367MSUN36G:01X38525 l=3 a=100000E00221C153:3
17 p=/dev/rdsk/c4t1d4s2 s=FUJITSU:MAN3367MSUN36G:01X37951 l=4 a=100000E00221C153:4
18 p=/dev/rdsk/c4t1d5s2 s=FUJITSU:MAN3367MSUN36G:01X39217 l=5 a=100000E00221C153:5
19 p=/dev/rdsk/c4t3d1s2 s=HITACHI:OPEN-3-SUN:20461000300 l=1 i=10350060E800000000000004FED00000003 a=50060E80034FED00:1
20 p=/dev/rdsk/c4t3d2s2 s=HITACHI:OPEN-3-SUN:20461000400 l=2 i=10350060E800000000000004FED00000004 a=50060E80034FED00:2

The 3pc.conf file can contain the following types of entries (keyword, if any, is in
parentheses):

Chapter 3, SAN Configuration for Advanced Client 61

Create Backup Configuration Files

# (comment line, or ignore-rest-of-line)

A pound sign (#) can be used to comment out whatever occurs to the right of the pound
sign to the end of the line. You can use this to introduce a comment, or to block out an
existing device-entry or portion of a line. For instance, if you remove a device from your
configuration temporarily, you could comment out the corresponding line for that device
in the 3pc.conf file, then remove the # later when the device is added back to the
configuration.

device ID (devid)
A unique NetBackup number for the device. In the 3pc.conf file, the device ID numbers
need not be in sequential order, but must be unique.

address (a=wwpn:lun)
The world-wide port name and lun as provided by the bpSALinfo command (see step 3
on page 76 for information on this command). For a device that has multiple FC ports,
there can be multiple a= entries.

client name (c=client)

The name of the remote client provided by the bptpcinfo command with the -x option (see
step 2 on page 75 for more information on this command).

device path (p=devpath)

The path to the Fibre Channel device. This entry must be specific to the media server on
which the 3pc.conf file resides.

client path (P=clientpath)

The path to the remote client provided by the bptpcinfo command with the -x option.

serial number (s=sn)

The serial number of the device, of the form:
Vendor ID:Product ID:device serial number

Note The disk devices must support SCSI serial-number inquiries or page code 83
inquiries. If a page code inquiry returns an identification descriptor (i=) for a disk,
the serial number is not required.

62 NetBackup Advanced Client System Administrators Guide

 Create Backup Configuration Files

lun (l=lun)
The devices logical unit number. The LUN allows NetBackup to identify devices that are
attached by SCSI connection to the third-party copy device, bridge, or other SAN device,
or that are directly attached to the fibre channel.

world-wide port name (w=wwpn)

Note w=wwpn is allowed for backwards compatibility with 4.5 clients. If you run the
bpSALinfo or bptpcinfo command, the w=wwpn entry will be converted to
a=wwpn:lun.

The devices fibre channel world-wide port name, which identifies the device on the SAN.
This is a 16-digit identifier, consisting of an 8-digit manufacturer name, and an 8-digit
device name (numeric).
The following is an example message showing a world-wide name for a device, written to
the /var/adm/messages log on the server. Note there are two versions of the
world-wide name: the node wwn and port wwn. For Advanced Client, use the port wwn.

Example of World-Wide Name in /var/adm/messages log

On some devices, the world-wide port name can be found on the back of the device or in
the boot-time messages written to the /var/adm/messages log on the NetBackup
media server.

alternate world-wide port name (W=wwpn)

Note W=wwpn is allowed for backwards compatibility with 4.5 clients. If you run the
bpSALinfo or bptpcinfo command, the W=wwpn entry will be converted to
a=wwpn:lun.

Chapter 3, SAN Configuration for Advanced Client 63

Create Backup Configuration Files

A disk in a disk array can be assigned to multiple Fibre Channel ports. This is done, for
instance, for load balancing or redundancy, to allow other devices on the SAN to access
the same disk through different ports. The two ports allow NetBackup to select the port by
which the storage device will access the disk.
In such a configuration, while w= specifies the first world-wide port name, W= specifies
the second world-wide port name for the disk. (Note the uppercase W in W=wwpn.)

identification descriptor (i=iddesc)

When available, this value (up to 43 characters) identifies the device on the SAN. If

available, the identification descriptor is automatically included in the 3pc.conf file

when you run the bptpcinfo command. See Determining Requirements for more

information on this value.

Determining Requirements

The following determines which values are required in the 3pc.conf file.

identification descriptor
The identification descriptor is optional, and is not supported by all vendors. (To produce
this descriptor, the device must support a page code inquiry with a type 2 or 3 descriptor
of less than 20 bytes.) The NetBackup bptpcinfo command (explained below) will
detect the devices identification descriptor and place it in the 3pc.conf file if the
identification descriptor is available.
Even when this descriptor is available, some third-party copy devices do not support its
use.

Note If an identification descriptor is available and the third-party copy device supports
it, the descriptor is used to identify the device on the SAN; in this case, there is no
need for the LUN or world-wide name. To determine whether your third-party
copy device supports identification descriptors, go to www.support.veritas.com
(see Advanced Client Information on the Web on page xix).

world-wide port name

If an identification descriptor is not available or the third-party copy device does not
support identification descriptors, the devices world-wide port name must be included in
the 3pc.conf file.

64 NetBackup Advanced Client System Administrators Guide

 Create Backup Configuration Files

What bptpcinfo Automatically Provides

The NetBackup bptpcinfo command detects some or all of the device information
needed for the backup and places that information in the 3pc.conf file, as follows:
The bptpcinfo command provides the device path, serial number, identification
descriptor (if available), and the LUN.
The bptpcinfo command does not provide the world-wide name. If you have
SANPoint Control, use bpSALinfo to obtain world-wide names.

What the Backup Methods Require

For the NetBackup Media Server backup method, the bptpcinfo command provides all
the information you need (no manual editing is required).
The Third-Party Copy Device method requires more information for each disk. In some
instances, the bptpcinfo command cannot gather all the information required.

mover.conf Description
The /usr/openv/volmgr/database/mover.conf file identifies the third-party copy
devices that NetBackup can use for the Third-Party Copy Device backup method. This file
is needed for the Third-Party Copy Device backup method only.
You can use the bpmoverinfo command to create the mover.conf file (see Create the
mover.conf File on page 77). In most cases, the bpmoverinfo command makes the
appropriate entry in the mover.conf file and no further configuration is needed. The
next few sections describe the types of entries that can be made in mover.conf, in case
manual configuration is needed.

Types of entries in mover.conf

Depending on your device configuration, the mover.conf file can consist of the following:
the passthru driver device path (sg path on Solaris or sctl path for HP)
the /dev/rmt/device_name of the third-party-copy-capable tape drive(s), or the
TAPE keyword
the /dev/rdsk/device_name of the third-party-copy-capable disk drive(s), or the
DISK keyword

Chapter 3, SAN Configuration for Advanced Client 65

Create Backup Configuration Files

Please note the following:

passthru driver device path

You must enter a /dev/sg (or /dev/sctl) passthru path in the mover.conf file when the
storage device for the backup is not behind (SCSI-connected to) the third-party copy
device that will handle the backup.

When passthru path is required in mover.conf file

Fibre Channel SAN

/dev/sg/c0t16l0

Tape 1 Third-party
copy device

In this example, to use the third-party copy device to send the backup to Tape 1, the mover.conf
file must include the passthru device path of the third-party copy device: /dev/sg/c0t16l0.

/dev/rmt/device_name of the tape

/dev/rdsk/device_name of the disk
You can enter the /dev/rmt/device_name of a tape or the /dev/rdsk/device_name of a
disk in the mover.conf file when the tape or disk is behind a third-party copy device, or
when the tape or disk has built-in third-party copy functionality. The third-party copy
device in front of (or inside) the tape or disk drive is used for the backup.

Note The /dev/rmt/device_name path will be used if it matches the drive path that
NetBackup selects for the backup. As a rule, this is not a problem, since the
bpmoverinfo command detects all available third-party copy devices (and any
tape devices behind them) and enters them in the mover.conf file. See Create the
mover.conf File on page 77.

66 NetBackup Advanced Client System Administrators Guide

 Create Backup Configuration Files

/dev/rmt/device_name in mover.conf file

Fibre Channel SAN

Third-party Third-party
copy device copy device

SCSI
Tape 1 Tape 3 /dev/rmt/3cbn

Tape 2 /dev/rmt/2cbn

In this example, to use a third-party copy device to send the backup to Tape 2 or to Tape3, the
mover.conf file can specify the device_name of the tape drive: /dev/rmt/2cbn or /dev/rmt/3cbn.
To use Tape 1, the mover.conf file would need the passthru device path of a third-party copy
device.

Note To use a tape unit as a third-party copy device (such as Tape 3), a SCSI passthru
device path must have been configured for that tape device.

TAPE and DISK keywords

For convenience, you can specify the TAPE keyword or DISK keyword in mover.conf
instead of listing individual tape or disk device_name paths. In the above example, the
TAPE keyword could be used to send the backup to either Tape 2 or Tape 3.
See Keywords in Mover File on page 70 for more information on these keywords.

For sites that have one third-party copy device

The mover.conf file can consist of one line specifying the device by means of its sg
driver device path (Solaris) or sctl device path (HP).
For example, on Solaris:
/dev/sg/c6t1l0

On HP:
/dev/sctl/c6t1l0

That is all you need in the mover.conf file. In most cases, you can use the
bpmoverinfo command to provide this entry.

Chapter 3, SAN Configuration for Advanced Client 67

Create Backup Configuration Files

You can use the following command to make sure the sg driver device path is correct.
This command is also useful for creating the mover.conf.policy_name or
mover.conf.storage_unit_name version of the mover.conf file (see Naming the
Mover File on page 73).
/usr/openv/volmgr/bin/sgscan

Here is some sample sgscan output showing third-party copy devices (see notes
following):
/dev/sg/c0t16l0: Mover: "ADIC Scalar SNC"/

/dev/sg/c0t12l3: Mover: "ATTO FB 4500H"

/dev/sg/c0t15l0: Mover: "CNSi FS2620"

/dev/sg/c0t17l0: Mover: "PATHLGHTSAN Gateway"

/dev/sg/c0t11l3: Mover: "Crossrds4250 Router"

/dev/sg/c0t2l10: Mover: "Crossrds8000 Router"

/dev/sg/c0t23l2: Mover: "OVERLANDNEO VIA FCO"

/dev/sg/c4t0l0: Changer: "SPECTRA 215"

Notes:
CNSi indicates a Chaparral device.
The number of entries returned for Crossroads depends on how many
controller LUNS have been configured on that device. The mover.conf file
must contain the /dev/sg path for each controller LUN that is configured on
the Crossroads.
The Spectra Logic tape library does not have separate controller LUNs for the
third-party functionality. For this reason, the sgscan output lists the library as
a Changer rather than as a Mover.
For an Hitachi, HP, or Sun disk array, you must check the HBA binding to
obtain the SCSI target number for the arrays ECopy target port, and use that
SCSI target number to identify the correct /dev/sg path in the sgscan output.
An alternative: the mover.conf file can consist of one line specifying the device by
means of the /dev/rmt/device_file_name or /dev/rdsk/device_file_name, where
device_file_name specifies the actual file name of the tape or disk. Note that the tape
device must be the same as the device that NetBackup selects for the backup, and the
disk must be one that is involved in the backup.
Instead of the /dev/rmt/device_file_name or /dev/rdsk/device_file_name path,
you can use the TAPE or DISK keyword. For more information, refer to Keywords in
Mover File on page 70.

68 NetBackup Advanced Client System Administrators Guide

 Create Backup Configuration Files

For sites that have multiple third-party copy devices

In the mover.conf file, if you want to specify one of the third-party copy devices and
prevent the others from being used, specify the device by means of its sg or sctl
driver device path (such as /dev/sg/c6t1l0 on Solaris or /dev/sctl/c6t1l0 on
HP), or specify its /dev/rmt/device_file_name or /dev/rdsk/device_file_name.
See the list of example /dev/sg paths on the previous page.
If you want to use any available tape or disk drive that is third-party-copy capable or
connected to a third-party copy device, specify the TAPE or DISK keyword.
If you want to limit the third-party copy device to that of a particular vendor or type,
while including a variety of devices in the file, use the END keyword. First enter the
device(s) you want to use, followed by END, then specify any other devices you might
want to use at another time. For more information, refer to Keywords in Mover File
on page 70.

SCSI Reserve/Release
For backups that use the third-party copy device method, SCSI reserve/release may be
required to prevent unintended sharing of tape devices by multiple hosts on the SAN.
With SCSI reserve/release, either the media server or the third-party copy device acquires
exclusive control over the tape drive, thus preventing other jobs from accessing the tape
during the backup.
The bptm process logs all SCSI reserve/release commands. For background information
on SCSI reserve/release, refer to the NetBackup Shared Storage Option System Administrators
Guide.

To use SCSI reserve/release

Note
If your mover.conf file contains only /dev/rmt/device_path entries or the TAPE
keyword, SCSI reserve/release will be used for the backup. No further
configuration is needed for SCSI reserve/release.

SCSI reserve/release is configured by means of the mover.conf file. The type of entry to
make in mover.conf depends on the type of tape device and the network connection it is
using, as follows:
If the tape device is a Fibre Channel device (not connected behind a router or bridge)
and does not have third-party copy functionality:
Specify the passthru path of the third-party copy device followed by the
i=reserve_value. For example:
/dev/sg/c6t1l0 i=2000001086100d5e

Chapter 3, SAN Configuration for Advanced Client 69

Create Backup Configuration Files

where 2000001086100d5e is a 16-digit user-supplied value. See i=reserve_value on

page 71 for more information on this value.
The third party copy device must be the tape device itself, or the tape drive must
support a special kind of SCSI reserve/release called third-party reservation;
otherwise, SCSI reserve/release will not be used. As of this writing, the only
Advanced Client-supported device that supports third-party reservation is the
ADIC/Pathlight Gateway. See i=reserve_value on page 71 for more information on
the reserve_value required by the ADIC/Pathlight.

A if third-party reservation is supported by the tape drive as you have

If the tape device does not have third-party copy functionality and does not support
the i=reserve_value:
Specify the passthru path of the third-party copy device followed by the hr keyword.
For example:
/dev/sg/c6t1l0 hr

The hr keyword tells NetBackup to use SCSI reserve/release. If the hr keyword is

omitted, SCSI reserve/release is not used.
If the tape is behind the third-party copy device or has its own third-party copy
functionality:
Specify the tape device path or the TAPE keyword. For example:
/dev/rmt/2cbn

or
TAPE

Keywords in Mover File

The following keywords can be included in the mover file:

DISK
For a third-party copy backup, attempt to use a disk involved with the current backup if
that disk has third-party copy functionality or is behind (SCSI-connected to) a third-party
copy device. This allows better concurrent backup processing, so that two or more backup
jobs can execute simultaneously.

Note The disks SCSI passthru driver device path must be included in the mover.conf file.

70 NetBackup Advanced Client System Administrators Guide

 Create Backup Configuration Files

TAPE
For a third-party copy backup, attempt to use the current tape device selected for the
backup if that device has third-party copy functionality or is behind (SCSI-connected to) a
third-party copy device. This has two advantages:
There is no need to specify a device path or passthru driver device path. Instead of
having to enter /dev/rmt/ paths for a number of tape devices, you can use the TAPE
keyword as shorthand for all of them.
Allows better concurrent backup processing, so that two or more backup jobs can
execute simultaneously.

END
Stop searching the mover.conf file for third-party copy devices for the current
third-party copy backup.
If there are two or more third-party copy devices in the mover.conf file, NetBackup tries
them sequentially, starting with the first one listed in the file, until one is found that can
successfully move the data. END means do not look further in the current mover file and
do not look in any other mover files, even if the last device tried was unsuccessful. Note
that if no successful device is found before END is reached, the backup fails.
The END keyword limits the search for a third-party copy device in a mover.conf file
that contains entries for more than one device. This can save you the trouble of deleting
device entries and re-entering them later.
For example, if the mover.conf file contains the following:
/dev/sg/c6t4l0

END

/dev/sg/c6t4l2

/dev/sg/c6t4l3

NetBackup will try to use device /dev/sg/c6t4l0 and will not try the other devices.

The following optional keywords can be added to each entry in mover.conf

i=reserve_value
Use SCSI reserve/release for third-party reservation, if supported by the tape device or by
the third-party copy device to which the tape device is connected. The reserve_value is a
world-wide port name or fibre channel port identifier, as follows.
For the ADIC/Pathlight Gateway, the reserve_value is the world-wide port name of
the ADIC/Pathlight.

Chapter 3, SAN Configuration for Advanced Client 71

Create Backup Configuration Files

For devices made by other vendors, the reserve_value may be the fibre channel port
identifier (destination ID) of the third-party copy device, with two leading zeros. For
example, if the fibre channel port identifier is 231DE4, the reserve_value is 00231DE4.
Please contact the vendor of the device for specifications.

hr
Hold the tape reservation (SCSI reserve/release) when a third-party copy device that is
not a tape device is designated by means of a passthru device path (/dev/sg/ on Solaris
or /dev/sctl/ on HP). If you do not specify the hr keyword, the default is to drop or
omit the reservation.

dr
Omit the use of SCSI reserve/release when a tape device is designated by the TAPE
keyword or its tape device path (such as /dev/rmt/2cbn). If you do not specify the dr
keyword, the default is to hold the reservation.
For example:
/dev/rmt/2cbn

/dev/rmt/3cbn

TAPE dr

In this example, if neither of the specified /dev/rmt devices can use SCSI reserve/release,
NetBackup will try a tape device without the reserve.

to
If the third-party copy device needs additional time to respond to a backup request, you
can increase the timeout value by specifying to followed by the limit in seconds. The
default is 300 seconds (5 minutes). Additional time may be needed, for instance, if the
third-party copy device is running in debug mode.
The following example resets the timeout for third-party copy device /dev/rmt/2cbn to
600 seconds:
/dev/rmt/2cbn to 600

In this example, NetBackup will allow the third-party copy device (accessible through
/dev/rmt/2cbn) ten minutes to respond to a backup request. If the device does not
respond within 10 minutes, NetBackup will try the next third-party copy device listed in
the mover file. If no other devices are listed, the backup will fail.

72 NetBackup Advanced Client System Administrators Guide

 Create Backup Configuration Files

A Note on Keywords for SCSI Reserve/Release

The same path (passthru or /dev/rmt/device_name path) can be specified several times
with different keywords or no keywords. NetBackup tries each path in succession
(whether or not they specify the same path), attempting to use SCSI reserve/release or
not, as specified.
Example:
/dev/sg/c6t1l0 i=4873968475898744

/dev/sg/c6t1l0 hr

/dev/sg/c6t1l0

In this example, NetBackup will try to use the third-party copy device specified by
/dev/sg/c6t1l0 and will attempt to use reserve/release by means of the
i=reserve_value. If unsuccessful, NetBackup will try to use the same third-party copy
device and reserve/release by means of the hr keyword (hold the reserve). If unsuccessful,
NetBackup will use the third-party copy device without the reserve.

Naming the Mover File

In addition to the standard mover.conf file name, there are two other options for
naming the mover file:
Per Policy
/usr/openv/volmgr/database/mover.conf.policy_name
where policy_name is the name of a NetBackup policy. All backups for this policy will use
the third-party copy device specified in this mover.conf.policy_name file.
For a disk that has third-party copy device capability, use the mover.conf.policy_name
to specify the disk as the third-party copy device for the policy that backs up that disk.
Per Storage Unit
/usr/openv/volmgr/database/mover.conf.storage_unit_name
where storage_unit_name is the name of a storage unit. This allows a third-party copy
device to use a particular storage device by means of a storage unit name. Here is an
example mover.conf file name of the storage_unit_name type:
mover.conf.nut-4mm-robot-tl4-0

where nut-4mm-robot-tl4-0 was selected as the storage unit in the policy.

Note
The storage_unit_name in this file name must exactly match the name of the storage
unit as it appears in the Policy storage unit field of the Change Policy dialog.

Chapter 3, SAN Configuration for Advanced Client 73

Create Backup Configuration Files

Selection Priority for mover.conf files

NetBackup looks for an appropriate mover.conf file in the following order:

1. mover.conf.policy_name

2. mover.conf.storage_unit_name

3. mover.conf

74 NetBackup Advanced Client System Administrators Guide

 Create Backup Configuration Files

Create the 3pc.conf File

The /usr/openv/volmgr/database/3pc.conf file contains a list of all disk and
tape devices on the SAN that NetBackup Advanced Client can use. NetBackup
automatically creates this file at the start of the backup. In certain circumstances,
however, you must create this file manually.

Note
You must create a 3pc.conf file if you are using the Third-Party Copy Device
backup method AND some of your devices do not support identification
descriptors (E4 target). Otherwise, you can skip to Create the mover.conf File on
page 77.

1. Create a 3pc.conf file as follows when no backups are in progress.

On the media server, run the bptpcinfo command:
/usr/openv/netbackup/bin/bptpcinfo -a
If a 3pc.conf file already exists in /usr/openv/volmgr/database, you are
asked if you want to overwrite it. You can use the -o output_file_name
option to send the output to a temporary file. However, the correct output must
be included in the /usr/openv/volmgr/database/3pc.conf file before
backups can succeed.
If a storage device is currently involved in a backup, the bptpcinfo command
cannot gather information on that device and skips to the next device. If the
3pc.conf file contains no entry for a storage device on your network, use the
verbose mode (-v) of the bptpcinfo command to determine if the device was
busy (see the NetBackup Commands for UNIX guide for more information on
bptpcinfo).

2.
If the media server does not have access to all disks (due to zoning or LUN-masking
issues), run the following command on the media server:
/usr/openv/netbackup/bin/bptpcinfo -x client_name

where client_name is the name of a NetBackup client on the fibre channel network
where the third-party copy device is located. The 3pc.conf file will be updated with
information about the disks on this network, allowing the media server to see those
disks. This information may have to be edited by adding the world-wide name
(wwn=) of each device, as explained in the next two steps.

Chapter 3, SAN Configuration for Advanced Client 75

Create Backup Configuration Files

Note that the entries added by the -x option do not include p=devpath. Instead, they
have c=client and P=clientpath. In the following example, lines 21 and 22 were
added by the -x option:

16 p=/dev/rdsk/c4t1d3s2 s=FUJITSU:MAN3367MSUN36G:01X38525 l=3 a=100000E00221C153:3

17 p=/dev/rdsk/c4t1d4s2 s=FUJITSU:MAN3367MSUN36G:01X37951 l=4 a=100000E00221C153:4

18 p=/dev/rdsk/c4t1d5s2 s=FUJITSU:MAN3367MSUN36G:01X39217 l=5 a=100000E00221C153:5

19 p=/dev/rdsk/c4t3d1s2 s=HITACHI:OPEN-3-SUN:20461000300 l=1 i=10350060E800000000000004FED00000003 a=50060E80034FED00:1

20 p=/dev/rdsk/c4t3d2s2 s=HITACHI:OPEN-3-SUN:20461000400 l=2 i=10350060E800000000000004FED00000004 a=50060E80034FED00:2

21 s=SEAGATE:ST39204LCSUN9.0G:3BV0B09V0000104413ZF l=4 c=magner P=/dev/rdsk/c4t28d4s2

22 s=FUJITSU:MAG3182LSUN18G:01533134 l=8 c=magner P=/dev/rdsk/c4t28d8s2

3.
If you have SANPoint Control, you can use it to add world-wide name and lun
information to the 3pc.conf file, by entering the following command on the media
server:
/usr/openv/netbackup/bin/admincmd/bpSALinfo -S SPC_server

where -S SPC_server identifies the host where SANPoint Control is running.

bpSALinfo adds world-wide name and lun values to device entries in the 3pc.conf
file. For additional command options, refer to the man page for bpSALinfo, or to the
NetBackup Commands for UNIX guide.
If bpSALinfo successfully updates the 3pc.conf file, no further editing of the 3pc.conf
file is required. You can skip to Create the mover.conf File on page 77.

4.
If you do not have SANPoint Control or it does not support your environment, edit
the 3pc.conf file as follows:
For each storage device listed in the 3pc.conf file, you may need to provide
world-wide port names, depending on what NetBackup was able to discover about
the device and what the third-party copy device supports.
These are the editing tasks:
In the 3pc.conf file, if each device that will be backed up with Advanced Client
has an identification descriptor (i=), and if the third-party copy device supports
the use of identification descriptors, the 3pc.conf file is complete. No editing is
needed; skip the rest of this section and continue with Create the mover.conf
File on page 77.
If the 3pc.conf file does not have an identification descriptor for each device (or
the descriptor is not supported by the third-party copy device), enter the
world-wide port name (w=) for each device. (Obtain the world-wide port name
from your Device Checklist on page 58.)

76 NetBackup Advanced Client System Administrators Guide

 Create Backup Configuration Files

Create the mover.conf File

This section describes how to create a list of the available third-party copy devices in a
mover.conf file.

Note This is required for the Third-Party Copy Device method only.

1.
On the NetBackup media server, enter the following command:
/usr/openv/netbackup/bin/admincmd/bpmoverinfo
This creates the following file:
/usr/openv/volmgr/database/mover.conf

The bpmoverinfo command discovers any third-party copy devices available on the
SAN and lists them in the mover.conf file. Any tape drives with third-party copy
capability are listed first.
For a description of the bpmoverinfo command, refer to the NetBackup Commands for
UNIX guide.

Note
For bpmoverinfo to correctly list third-party copy devices in the mover.conf file,
the third-party copy devices must already have passthru paths defined. For an
example, see Solaris only: Example for sg.links, sg.conf, and st.conf files on
page 57.

2.
If you need to control the circumstances under which a third-party copy device is
used, create a separate mover.conf file for a policy or storage unit:
/usr/openv/volmgr/database/mover.conf.policy_name
or
/usr/openv/volmgr/database/mover.conf.storage_unit_name
For information on these naming formats and possible mover file entries, refer to
mover.conf Description on page 65 and Naming the Mover File on page 73.
Following are example storage environments and mover.conf files.

Chapter 3, SAN Configuration for Advanced Client 77

Create Backup Configuration Files

Example mover.conf file for a site with one third-party copy device

NetBackup NetBackup On media server

client media server mover.conf file contains this
third-party copy device:
LAN / WAN /dev/sg/c6t1l0

Fibre Channel SAN

Third-party
copy device
/dev/sg/c6t1l0

robot0 robot1

Client disks on SAN

storage unit name: storage unit name:

TLD_robot0 TLD_robot1

In the above example, backups will use third-party copy device /dev/sg/c6t1l0
specified in the mover.conf file. The backup uses the storage unit (TLD_robot0 or
TLD_robot1) specified for the policy on the Change Policy dialog.
See the next figure for an example configuration involving a disk array with third-party
copy device capability.

78 NetBackup Advanced Client System Administrators Guide

 Create Backup Configuration Files

Example mover.conf.policy_name file for site with third-party copy capability in disk array

NetBackup NetBackup On media server:

client media server mover.conf.array_1 file contains
this third-party copy device:
LAN / WAN /dev/sg/c6t1l0

Fibre Channel SAN

Disk array with

third-party copy.
Array contains the robot0
client data to back up.

/dev/rdsk/c6t1d0sx storage unit name:

TLD_robot0

In this example, policy array_1 is configured to back up the client data contained on the
disk array. The backup uses storage unit TLD_robot0 to store the data.
All backups configured in this policy will use the disk array as the third-party copy
device. The mover.conf.array_1 file specifies that array.

Note The client data must reside in the array that is used as the third-party copy device.

See the next figure for an example configuration with two third-party copy devices, where
both devices can use the same robot.

Chapter 3, SAN Configuration for Advanced Client 79

Create Backup Configuration Files

Example mover.conf.storage_unit_name files for two third-party copy devices

NetBackup NetBackup On media server:

client media server file name: mover.conf.TLD_robot0
contains: /dev/sg/c6t1l0
LAN / WAN
file name: mover.conf.TLD_robot00
contains: /dev/sg/c6t4l0
file name: mover.conf.TLD_robot1
contains: /dev/sg/c6t4l0

Fibre Channel SAN

Client disks
Third-party Third-party
on SAN
copy device-1 copy device-2
/dev/sg/c6t1l0 /dev/sg/c6t4l0

robot0 robot1

storage unit names: storage unit name:

TLD_robot0 TLD_robot1
TLD_robot00

The above example shows two robots (robot0 and robot1). Robot0 has been assigned
two storage unit names, TLD_robot0 and TLD_robot00. Robot1 has been assigned
one storage unit name, TLD_robot1.
The above example also shows two third-party copy devices, device-1 with a SCSI
passthru device path of /dev/sg/c6t1l0, and device-2 with a SCSI passthru device
path of /dev/sg/c6t4l0.
To allow third-party copy device-1 to use robot0, create a file named
mover.conf.TLD_robot0. In the file, include the device path of device-1
(/dev/sg/c6t1l0).

80 NetBackup Advanced Client System Administrators Guide

 Create Backup Configuration Files

To allow third-party copy device-2 to use the same robot (robot0), create a file
named mover.conf.TLD_robot00. In the file, include the device path of
device-2 (/dev/sg/c6t4l0). Notice that the file name must refer to a different
storage unit, TLD_robot00, which is assigned to robot0.
To allow third-party copy device-2 to use robot1, create a file named
mover.conf.TLD_robot1 that includes the device path of device-2
(/dev/sg/c6t4l0).

Note
The storage_unit_name portion of the mover.conf.storage_unit_name file
name must exactly match the actual name of the storage unit. See under
Configuring an Advanced Client Policy on page 85 for an example Change Policy
dialog showing a storage unit name in the Policy storage unit field.

Chapter 3, SAN Configuration for Advanced Client 81

Create Backup Configuration Files

82 NetBackup Advanced Client System Administrators Guide

Policy Configuration 4

This chapter explains how to configure NetBackup Advanced Client policies.

The following topics are covered in this chapter:

Notes and Prerequisites

Configuring an Advanced Client Policy

Automatic Snapshot Selection

Selecting the Snapshot Method

Configuring Backup Scripts

Configuring Alternate Client Backup

Configuration Tips

83
Notes and Prerequisites

Notes and Prerequisites

Before proceeding with this chapter, note the following.
NetBackup Enterprise and the Advanced Client add-on product must be installed on
master server(s) and clients.
Although the NetBackup 5.1 Administration Console can be used to administer a
NetBackup 4.5 server, please note:
The Advanced Snapshot Options button is disabled and snapshot configuration
is done by means of the policy Clients tab > Snapshot Backups button. Use
alternate client is also disabled.
The Windows version of the Administration Console cannot configure snapshot
methods, and cannot configure Instant Recovery with offhost backup.
For the NetBackup Media Server and Third-Party Copy Device offhost backup
methods, a fibre channel network or multi-ported SCSI disk array must be configured.
Storage devices must be configured (you can use the Device Configuration wizard).
To use BLIB with Advanced Client (Perform block level incremental backups option
on the policy attributes tab), you must install the NetBackup for Oracle software.
Oracle is the only client data type supported by the block level incremental backup
feature of Advanced Client. Refer to the NetBackup for Oracle System Administrators
Guide for assistance in making block level incremental backups with Advanced Client.

Note If you choose the Perform block level incremental backups option on the policy
attributes tab, the other features of Advanced Client are not available and are
grayed out.

84 NetBackup Advanced Client System Administrators Guide

 Configuring an Advanced Client Policy

Configuring an Advanced Client Policy

This procedure explains how to configure Advanced Client features in a policy.

1. Start the NetBackup Administration Console as follows:

On UNIX, enter: /usr/openv/netbackup/bin/jnbSA &
On Windows, click Start > Programs > VERITAS NetBackup > NetBackup
Administration Console.

2. Click on Policies.

3. In the All Policies pane, double click on the name of the policy (or right-click to create
a new one).

Select the policy type.

Select appropriate storage

unit or storage unit group.

Requires NetBackup for Oracle

database agent software. See
the NetBackup for Oracle
System Administrators Guide.

Click Perform snapshot

backups.

If you checked Perform

Offhost backup, specify
method.

Chapter 4, Policy Configuration 85

Configuring an Advanced Client Policy

4. Select the policy type:

If client data is in a database, select the database type (DB2, Oracle,
MS-Exchange-Server, MS-SQL-Server).
To use FlashBackup, select FlashBackup for UNIX clients or
FlashBackup-Windows for Windows 2000/2003 clients. (See the FlashBackup
Configuration chapter for further information.)
For all other cases, select Standard for UNIX clients and MS-Windows-NT for
Windows clients.

5. Make sure Perform snapshot backups is selected.

6. Optional: To select the snapshot method manually, refer to Selecting the Snapshot
Method on page 90. Skip this step if you want NetBackup to select the snapshot
method for you. Automatic Snapshot Selection on page 89 describes how
NetBackup chooses a snapshot method.

7. To create a backup that enables Instant Recovery, select the Retain snapshots for
instant recovery attribute.
This attribute is required for the following types of restore: block-level restore, file
promotion, and image rollback. These are described under Instant Recovery Restore
on page 184. For help in creating a policy for instant recovery backups, refer to
Instant Recovery Configuration on page 117.

8. To reduce the processing load on the client, select Perform offhost backup.
This option may require additional configuration.

a. For a backup performed by an alternate client, select Use alternate client and
enter the name of the alternate client. Refer to Configuring Alternate Client
Backup on page 96 for more info.

b. For a backup performed by a data mover (not by a client), select Use data mover
and select the method:
NetBackup Media Server
Backup processing will be handled by a Solaris or HP NetBackup media server.
Third-Party Copy Device
Backup processing will be handled by a third-party copy device.
Network Attached Storage

86 NetBackup Advanced Client System Administrators Guide

 Configuring an Advanced Client Policy

Backup processing will be handled by an NDMP host (NAS filer), with the
NAS_Snapshot method. NetBackup for NDMP software is required. For help
configuring a policy for Network Attached Storage and NAS_Snapshot, refer to
the NAS Snapshot Configuration chapter.

c. Specify a policy storage unit or group of storage units in the Policy storage unit
pull-down menu.

Note Any_available is not supported for the following data mover types: NetBackup
Media Server and Third-Party Copy Device. Disk storage units are not supported
for Third-Party Copy Device.

Instead of using a particular storage unit, you can create a storage unit group that
designates devices configured on the SAN. Storage unit groups are described in
the NetBackup Media Manager System Administrators Guide.

d. Also note the following:

For the Third-Party Copy Device option, you must create a mover.conf file.
See Create the mover.conf File on page 77. For background information on
this file, see mover.conf Description on page 65.
If you do not have SANPoint Control, and one or more devices needed for the
backup do not support identification descriptors (SCSI E4 target descriptors),
you may need to edit the 3pc.conf file as explained in Create the 3pc.conf
File on page 75.

9. To save these settings, click Apply.

10. Use the Schedules tab to define a schedule, the Backup Selections tab to specify the
files to be backed up, and the Clients tab to specify the clients.
Please note the following:
Only one snapshot method can be configured per policy. If, for instance, you want
to select one snapshot method for clients a, b, and c, and a different method for
clients d, e, and f, create two policies for each group of clients and select one
method for each policy.
Advanced Client policies do not support the ALL_LOCAL_DRIVES entry in the
Backup Selections list.
If you use the Backup Policy Configuration wizard, see Backup Policy
Configuration Wizard on page 100.

Chapter 4, Policy Configuration 87

Configuring an Advanced Client Policy

For snapshot backups, the maximum pathname length is 1000 characters (as
opposed to 1023 characters for backups that do not use a snapshot method). Refer
to Maximum Pathname Length on page 99 for more information on this
restriction. The NetBackup System Administrators Guide describes other file-path
rules.

11. Click Close when done with Schedules and Clients tabs.
Advanced Client validation begins. An error message may be displayed, such as the
following:

The Details pane explains the problem. You can click No, resolve the problem, and
close the policy again, or click Yes to override the message.

88 NetBackup Advanced Client System Administrators Guide

 Automatic Snapshot Selection

Automatic Snapshot Selection

To have NetBackup select the snapshot method, click Perform snapshot backups on the
policy Attributes tab.
If this is a new policy, NetBackup will select a snapshot method when the backup
starts (by default, the snapshot method is set to auto).
If this is a copy of a policy that was configured for a particular snapshot method, click
the Advanced Snapshot Options button and set the snapshot method to auto.
NetBackup will select a snapshot method when the backup starts.

Use of the auto method does not guarantee that NetBackup can select a snapshot method
for the backup. NetBackup looks for a suitable method based on a number of factors:
The client platform and policy type.
The presence of up-to-date software licenses, such as VERITAS VxFS and VxVM.
How the client data is configured. For instance, whether a raw partition has been
specified for a copy-on-write cache (refer to How to Enter the Cache on page 136),
or whether the clients data is contained in VxVM volumes configured with one or
more snapshot mirrors.
NetBackup uses the first suitable method found.
For greater control, refer to Selecting the Snapshot Method on page 90.

Chapter 4, Policy Configuration 89

Selecting the Snapshot Method

Selecting the Snapshot Method

This section explains how to manually configure a snapshot method from the
Administration Console.

1. Start the NetBackup Administration Console:

On UNIX, enter: /usr/openv/netbackup/bin/jnbSA &
On Windows, click Start > Programs > VERITAS NetBackup > NetBackup
Administration Console.

2. Click on Policies. In the All Policies pane, double click on the name of the policy. The
Change Policy dialog appears.

3. Make sure that Perform snapshot backups is selected (for more information on this,
refer to the previous procedure).

4. Click Advanced Snapshot Options.

The Snapshot Options dialog appears. This is for selecting a snapshot method for the
policy.

Select snapshot method

Specify values as needed

(if any)

5. Select the snapshot method for the policy.

The available methods depend on how your clients are configured and which
attributes you selected on the Attributes tab.

90 NetBackup Advanced Client System Administrators Guide

 Selecting the Snapshot Method

Note Only one snapshot method can be configured per policy. If you want to select one
snapshot method for clients a, b, and c, and a different method for clients d, e, and f,
create two policies for each group of clients and select one method for each policy.

auto
NetBackup will select a snapshot method when the backup starts. See Automatic
Snapshot Selection on page 89.
BusinessCopy
For mirror snapshots with Hewlett Packard XP series disk arrays with BusinessCopy
Services. For clients on Solaris or HP.
FlashSnap
For mirror snapshots on alternate clients, with the VxVM FlashSnap feature. UNIX
clients must be at VxVM 3.2 or later; Windows clients must be at VxVM 3.1 or later,
with all the latest VxVM service packs and updates.
FlashSnap is based on the VxVM disk group split and join technology.
NAS_Snapshot
For snapshots of client data residing on an NDMP host. Requires NetBackup for
NDMP software. For help configuring a policy for NAS_Snapshot, refer to the NAS
Snapshot Configuration chapter.
nbu_snap
For copy-on-write snapshots of UFS or VERITAS VxFS file systems. For Solaris clients
only. nbu_snap is not supported in clustered file systems.
nbu_snap requires a designated cache; seeHow to Enter the Cache on page 136.
ShadowImage
For mirror snapshots with Hitachi Data Systems disk arrays with ShadowImage
(HOMRCF). For clients on Solaris or HP.
TimeFinder
For mirror snapshots with EMC Symmetrix Disk Arrays (with TimeFinder SYMAPI).
For clients on Solaris or HP.
VSP
(VERITAS Volume Snapshot Provider:) for snapshots of open and active files. For
clients on Windows 2000 or 2003 only.

Note You can use VSP without Advanced Client, as explained in the NetBackup System
Administrators Guide for Windows, Volume I. In some cases, however, such as when
the Busy File Timeout has expired, no snapshot is created and the backup job may

Chapter 4, Policy Configuration 91

Selecting the Snapshot Method

continue without backing up the busy file. If you use VSP with Advanced Client,
the backup will either successfully create a snapshot of all files, or the backup job
will fail with a status code 11.

VSS
For snapshots using the Shadow Copy Service of Windows 2003. For clients on
Windows 2003 only. The note under VSP also applies to VSS.
For configuration assistance, please refer to your Microsoft documentation.
VVR
For alternate client backups of a replicated VxVM volume. Requires VxVM 3.2 or later
with the VERITAS Volume Replicator license. For clients on Solaris or HP.
VxFS_Checkpoint
For copy-on-write snapshots (Solaris or HP). This method is not supported by the
FlashBackup policy type.
Requires the Storage Checkpoint feature of VxFS 3.4 or later. For HP, VxFS 3.5 or later
is required.

Note Note that VxFS_Checkpoint requires the NetBackup Advanced Client license and
the VERITAS File System license with the Storage Checkpoints feature. Without
both licenses, the copy-on-write snapshot (Storage Checkpoint) cannot be opened
and the backup fails.

VxFS_Snapshot
For snapshots of Solaris or HP clients on the local host (not offhost), for FlashBackup
policies only. This method requires VxFS 3.4 (Solaris) or 3.3 (HP) or later. This method
also requires a designated cache; see VxFS_Snapshot on page 140 for details. Note
that all files in the Backup Selections list must reside in the same file system.
vxvm
For any of the following types of snapshots with data configured over Volume
Manager volumes (clients on Solaris, HP, or Windows).
For third-mirror snapshots (VxVM 3.1 or later).
For full-sized instant snapshots (VxVM 4.0).
For space-optimized instant snapshots (VxVM 4.0).

Note For further notes relating to these methods, refer to the Snapshot Configuration
Notes chapter.

6. Specify required parameters (if any).

92 NetBackup Advanced Client System Administrators Guide

 Selecting the Snapshot Method

Cache device path: specify a raw partition for the cache by entering the cache
partitions full path name in the Value field. For example:
/dev/rdsk/c2t0d3s3
This setting overrides the cache specified on Host Properties > Clients > Client
Properties dialog > UNIX Client > Client Settings (see How to Enter the Cache on
page 136).
Do not specify wildcards (such as /dev/rdsk/c2*). See Cache device on page 134
for a complete list of requirements.

Caution The cache partitions contents will be overwritten by the nbu_snap process.

Keep snapshot after backup:

This option specifies whether or not the snapshot image is retained on the mirror disk
after the backup completes (default is No). Retaining the image (Yes) enables a quick
restore from the mirror in case data is deleted from the clients primary disk. The
image is retained on the mirror until the next time a backup is run using this policy.
No indicates that the image will be deleted from the mirror as soon as the backup
completes.

Note If the client is rebooted, snapshots that have been kept must be remounted before
they can be used for a restore. You can use the bpfis command to discover the
images (refer to the bpfis man page or the NetBackup Commands manual). This does
not apply to snapshots for Instant Recovery: NetBackup automatically remounts
them as needed.

Note If the snapshot is made on an EMC, Hitachi, or HP disk array, and you want to use
hardware-level restore, read the Caution under Hardware-Level Disk Restore on
page 172.

Maximum Snapshots (Instant Recovery only)

Indicates the number of snapshots to keep when Instant Recovery is enabled. When
the maximum number of snapshots is reached, the next snapshot causes the oldest to
be deleted. The default is 1. If the number of mirrors configured for the snapshots is
less than the Maximum Snapshots value, the number of mirrors determines how
many snapshots are kept.
The best setting depends on several factors: how often the primary filesystem
changes, how much new data is being added to it, and the number of snapshots being
created (how often the policy is executed).

Chapter 4, Policy Configuration 93

Selecting the Snapshot Method

Caution If you specify a number that is smaller than the existing number of snapshots,
NetBackup deletes the older snapshots until the number of snapshots equals
that specified for Maximum Snapshots.

Synchronize mirror before backup:

At start of the backup, this option determines whether or not the primary and mirror
devices are automatically synchronized (if they were not already synchronized)
before the backup begins. Default is No. Specify Yes to have unsynchronized devices
synchronized before the backup begins. No means that unsynchronized devices will
not be synchronized before the backup starts. In this case (No), if the primary and
mirror devices are not synchronized, the backup will fail.
Resynchronize mirror in background:
At end of the backup, this option determines whether or not the backup job will
complete before the resynchronize operation has finished (default is No). Yes means
that a resynchronize request will be issued, but the backup will not wait for the resync
operation to complete. No means that the backup job cannot reach completion until
the resynchronize operation has finished.
Choosing Yes may allow more efficient use of backup resources. For example, if two
backup jobs need the same tape drive, the second job can start even though the
resynchronize in background operation issued by the first job has not completed.
IBC send timeout (seconds):
For the VVR snapshot method, this value determines how long NetBackup waits for
the next end-of-transaction marker to be automatically removed from the replication
data stream. If the marker is not removed, replication cannot continue. If this timeout
is exceeded before the current marker is removed, the backup will fail.
IBC receive timeout (seconds):
For the VVR snapshot method, this value determines how long NetBackup waits for
the next end-of-transaction marker to be received in the replication data stream. For
instance, a process may fail to send the marker, or the network connection may be
down or overloaded. If this timeout is exceeded before the next marker is received,
the backup will fail.

7. When finished, click OK.

94 NetBackup Advanced Client System Administrators Guide

 Configuring Backup Scripts

Configuring Backup Scripts

For backups using a snapshot method, you can run scripts before and after the snapshot
by adding directives to the Backup Selections list, as follows.

1. Start the NetBackup Administration Console by entering the following:

On UNIX: /usr/openv/netbackup/bin/jnbSA &
On Windows: Start > Programs > VERITAS NetBackup > NetBackup
Administration Console

2. Click on Policies. In the All Policies pane, double click on the name of the policy. The
Change Policy dialog appears.

3. Click the Files tab.

4. Add the following directive to the start of the Backup Selections list:
METHOD=USER_DEFINED

5. Add one or both of the following directive(s), as required.

DB_BEGIN_BACKUP_CMD=your_begin_script_path

DB_END_BACKUP_CMD=your_end_script_path

The arguments (arg) are optional.

For example:

In the above example, the script shutdown_db.ksh is run before the backup, and
restart_db.ksh is run after the snapshot is created.

Chapter 4, Policy Configuration 95

Configuring Alternate Client Backup

Configuring Alternate Client Backup

NetBackup Advanced Client supports three methods of doing offhost backup: alternate
client backup, NetBackup Media Server, and Third-Party Copy Device. This section
discusses alternate client backup only.
Note that alternate client backup does not require a SAN, a third-party copy device, or
Fibre Channel equipment. There is no need to create a 3pc.conf or mover.conf file, or to do
any of the procedures described in the SAN Configuration for Advanced Client chapter.
Alternate client backup is offhost in that all backup processing is off-loaded to another
client. Off-loading the work to another client saves computing resources on the original
client. The snapshot is captured on the alternate client, and the backup has zero impact on
the original client.
For an introduction to the basic configurations supported for alternate client backup (split
mirror and data replication), refer to Alternate Client Backup on page 10.

Basic Requirements
Before configuring a policy for alternate client backup, make sure the following have been
done:
For the FlashSnap and VVR snapshot methods, VxVM 3.2 or later (for UNIX) or
VxVM 3.1 or later (Windows) must be installed and volumes configured over the
primary hosts disks. The VxVM FlashSnap or VVR license must also be installed.
The user and group identification numbers (UIDs and GIDs) associated with the files
to be backed up must be available to both hosts (the primary client and the alternate
backup client).
The primary and alternate clients must be running the same operating system,
volume manager, and file system. For each of these I/O system components, the
alternate client must be at the same level as the primary client, or higher level.
Following are the supported configurations:

If primary client is: Alternate client must be:

Solaris 7, 8, or 9 Solaris, at same level as primary client or higher

HP 11i HP 11i

Windows 2000/2003 Windows 2000/2003, at same level as primary

client or higher

VxFS 3.4 or later (VxFS 3.3 for HP) VxFS, at same level as primary client or higher

96 NetBackup Advanced Client System Administrators Guide

 Configuring Alternate Client Backup

If primary client is: Alternate client must be:

VxVM 3.2 or later (UNIX) VxVM, at same level as primary client or higher.
VxVM 3.1 or later1 (Windows) Note: for VVR method, the alternate client must
be at exactly the same level as primary client.

1. For VxVM on Windows, use VxVM 3.1 or later with all the latest VxVM service packs
and updates.

Available Backup Methods and Snapshot Methods

When setting up a policy for alternate client backup, you can make the following
selections for policy type, offhost backup method, and snapshot method, depending on
your hardware configuration and NetBackup add-on product licenses:

Policy type:
Choose Standard,
FlashBackup,
FlashBackup-Windows,
MS-Windows-NT,
MS-Exchange-Server,
MS-SQL-Server, DB2, or
Oracle

Click Perform snapshot

backups

Click Perform offhost

backup
Click Use alternate
client, and select the
alternate client from the
drop-down list, or type it
in.
Click Advanced
Snapshot Options if you
want to choose the
snapshot method.

Chapter 4, Policy Configuration 97

Configuring Alternate Client Backup

Snapshot method:
You can select the auto
method, or the following:
FlashSnap (for a disk
group split configuration,
with VxVM 3.2 or later
with the FlashSnap
feature).
VVR (for a UNIX
replication host; requires
VxVM 3.2 or later with
VVR feature).
TimeFinder,
ShadowImage,
BusinessCopy (the
array-specific methods, UNIX only).

Example configurations

1. Client data is on an EMC disk array in split-mirror mode

To run the backup on an alternate client, choose Standard as the policy type, select
Perform snapshot backups, Perform offhost backup, Use alternate client, and select the
alternate client. On the Snapshot Options display, specify the TimeFinder snapshot
method.
If the data is in an Oracle database, select Oracle as the policy type.

2. Client data is replicated on a remote host

To run the backup on the replication host (alternate client), choose Standard as the policy
type, select Perform snapshot backups, Perform offhost backup, Use alternate client,
and select the alternate client (the replication host). On the Snapshot Options display,
specify the VVR snapshot method.

3. Client data is on a JBOD array in VxVM volumes with snapshot mirrors configured
To run the backup on the alternate client, choose Standard (for UNIX client) or
MS-Windows-NT (Windows client) as the policy type, select Perform snapshot backups,
Perform offhost backup, Use alternate client, and select the alternate client. On the
Snapshot Options display, specify the FlashSnap method.

98 NetBackup Advanced Client System Administrators Guide

 Configuration Tips

If the client data consists of many files, or if you need the ability to do individual file
restore from raw partition backups, select FlashBackup or FlashBackup-Windows as the
policy type.

Note Other combinations of policy type and snapshot method are possible, depending on
many factors, such as your hardware configuration, file system and volume
manager configuration, and installed NetBackup product licenses.

Before running the alternate client backup

Your volume configuration must be prepared and tested for the snapshot method you will
use (see the Snapshot Configuration Notes chapter).

Configuration Tips

Maximum Pathname Length

For snapshot backups, the maximum file list pathname length is 1000 characters (as
opposed to 1023 characters for backups that do not use a snapshot method). This is
because the snapshot is created on a new mount point which is derived from the original
Backup Selections list pathname. If this new mount point plus the original file path
exceeds 1023 characters, the backup fails with a status code 1, the requested operation
was partially successful.
Refer to the NetBackup System Administrators Guide for other NetBackup file-path rules.

Snapshot Tips
In the Backup Selections list, be sure to specify absolute path names. Refer to the
NetBackup System Administrators Guide for help specifying files in the Backup
Selections list.
If an entry in the Backup Selections list is a symbolic (soft) link to another file,
Advanced Client backs up the link, not the file to which the link points. This is
standard NetBackup behavior. To back up the actual data, include the file path to the
actual data in the Backup Selections list.
On the other hand, a raw partition can be specified in its usual symbolic-link form
(such as /dev/rdsk/c0t1d0s1): do not specify the actual device name that
/dev/rdsk/c0t1d0s1 is pointing to. For raw partitions, Advanced Client
automatically resolves the symbolic link to the actual device.

Chapter 4, Policy Configuration 99

Configuration Tips

The Cross mount points policy attribute is not available for policies that are
configured for snapshots. This means that NetBackup will not cross file system
boundaries during a backup of a snapshot. A backup of a high-level file system, such
as / (root), will not back up files residing in lower-level file systems unless those file
systems are also specified as separate entries in the Backup Selections list. To back up
/usr and /var, for instance, both /usr and /var must be included as separate
entries in the Backup Selections list.
On Windows, the \ must be entered in the Backup Selections list after the drive letter
(for example, C:\). For the correct format when using a FlashBackup-Windows policy,
see step 10 on page 105.
For more information on Cross mount points, refer to the NetBackup System
Administrators Guide.

Backup Policy Configuration Wizard

The Backup Policy Configuration Wizard helps you set up a basic NetBackup policy. If
you used this wizard to configure a policy for NetBackup, please note the following:
If you chose Back up all local drives, the wizard placed an entry called
ALL_LOCAL_DRIVES in the policy Backup Selections list. Advanced Client does not
support the ALL_LOCAL_DRIVES option. Be sure to remove the
ALL_LOCAL_DRIVES entry from the Backup Selections list.
The wizard automatically sets your storage device to Any_available. This is correct
for local backups. However, to use the Third-Party Copy Device or NetBackup Media
Server method, you must select a particular storage unitdo not select
Any_available.

Multiple Data Streams

Note Multiplexing is not supported for the Third-Party Copy Device method. (Keep in
mind that multiplexing is not the same as multiple data streams.)

For backups, make sure the following are set to allow the number of active streams to be
equal to or greater than the number of streams in the Backup Selections list:
Policy attribute: Limit jobs per policy
Schedule setting: Media multiplexing
Storage unit setting: Maximum multiplexing per drive
System configuration setting: Maximum jobs per client

100 NetBackup Advanced Client System Administrators Guide

FlashBackup Configuration 5

This chapter describes FlashBackup and explains how to configure FlashBackup policies.

The following topics are covered in this chapter:

FlashBackup Capabilities

Restrictions

Configuring a FlashBackup Policy

Configuring FlashBackup in the Earlier Manner (UNIX only)

101
FlashBackup Capabilities

FlashBackup Capabilities
FlashBackup is a policy type that combines the speed of raw-partition backups with the
ability to restore individual files. The features that distinguish FlashBackup from other
raw-partition backups and standard file system backups are these:
Increases backup performance over standard file-ordered backup methods. For
example, a FlashBackup of a file system completes in substantially less time than
other types of backup, if the file system contains a very large number of files and most
of the file system blocks are allocated.
Individual files can be restored from raw-partition backups.
Backs up the following file systems: VxFS (Solaris and HP), ufs (Solaris), Online JFS
(HP), and NTFS (Windows 2000/2003).
Supports multiple data streams, to further increase the performance of raw-partition
backups when there are multiple devices in the Backup Selections list.

Restrictions
FlashBackup policies do not support file systems managed by HSM.
FlashBackup does not support VxFS storage checkpoints used by the
VxFS_Checkpoint snapshot method.
FlashBackup supports the following I/O system components: ufs, VxFS, and
Windows NTFS file systems, VxVM volumes and LVM volumes, and raw disks. Other
components (such as non-VERITAS volume managers or storage replicators) are not
supported.
Note these restrictions for Windows clients:
FlashBackup-Windows policies do not support the backup of Windows
system-protected files (the System State, such as the Registry and Active
Directory).
FlashBackup-Windows policies do not support the backup of Windows OS
partitions that contain the Windows system files (usually C:).
FlashBackup-Windows policies do not support the backup of Windows System
database files (such as RSM Database and Terminal Services Database).
FlashBackup-Windows policies do not support exceptions to client exclude lists.

102 NetBackup Advanced Client System Administrators Guide

 Configuring a FlashBackup Policy

Configuring a FlashBackup Policy

1. Start the NetBackup Administration Console:
On UNIX, enter: /usr/openv/netbackup/bin/jnbSA &
On Windows, click Start > Programs > VERITAS NetBackup > NetBackup
Administration Console.

2. Click on Policies. Double click on the policy (or right-click to create a new one).

Select FlashBackup or
FlashBackup-Windows

Select storage unit or

group.

Click Perform
snapshot backups.
Pre-selected for
Windows, optional for
UNIX.

3. Select the Policy type: FlashBackup for UNIX clients, or FlashBackup-Windows for
Windows clients.
These policy types allow you to restore individual files from the raw-partition
backup.

4. Specify the storage unit.

Chapter 5, FlashBackup Configuration 103

Configuring a FlashBackup Policy

FlashBackup and FlashBackup-Windows policies support both tape storage units and
disk storage units.

5. Select a snapshot method in one of the following ways:

Click Perform snapshot backups on the Attributes tab.
If this is a new policy, NetBackup will select a snapshot method when the backup
starts.
If this is a copy of a policy that was configured for a particular snapshot method,
click the Advanced Snapshot Options button and set the snapshot method to
auto. NetBackup will select a snapshot method when the backup starts.
Click Perform snapshot backups, click the Advanced Snapshot Options button,
and then select a snapshot method. For more information, refer to Selecting the
Snapshot Method on page 90.

6. UNIX only: if you selected nbu_snap or VxFS_Snapshot as the snapshot method,

specify a raw partition as cache, in either of these ways:
Use the Host Properties node of the Administration Console to specify the default
cache device for a client. Click Host Properties > Clients, select the client, then
Actions > Properties, UNIX Client > Client Settings.
Use the Advanced Snapshot Options dialog (see How to Enter the Cache on
page 136) to specify the cache used for the client when backed up by this policy.

Note The partition must exist on all clients included in the policy.

7. To shift the backup I/O to an alternate client (for UNIX or Windows clients), or to a
NetBackup media server or third-party copy device (for UNIX clients only), select
Perform offhost backup. See step 8 on page 86 for more instructions on this option.
For FlashBackup, the Use data mover option is supported for UNIX clients only.

8. To reduce backup time when more than one raw partition is specified in the Backup
Selections list, select Allow multiple data streams.

9. Use the Schedules tab to create a schedule.

FlashBackup policies support full and incremental types only. User backup and
archive schedule types are not supported.

Note For FlashBackup and FlashBackup-Windows policies, a full backup backs up all
blocks in the disk or raw partition selected in the Backup Selections tab (see next
step). An incremental backup backs up only the blocks associated with the files that
were changed since the last full backup.

104 NetBackup Advanced Client System Administrators Guide

 Configuring a FlashBackup Policy

10. Use the Backup Selections tab to specify the drive letter or mounted volume
(Windows) or the raw disk partition (UNIX) containing the files to back up.
Windows examples:
\\.\E:

\\.\E:\mounted_volume

Note The drive must be designated exactly as shown above (E:\ is not correct). Backing
up the drive containing the Windows system files (usually the C drive) is not
supported.

Solaris examples:
/dev/rdsk/c1t0d0s6

/dev/vx/rdsk/volgrp1/vol1

HP-UX examples:
/dev/rdsk/c1t0d0

/dev/vx/rdsk/volgrp1/vol1

Note On UNIX: The Backup Selections tab must specify the raw (character) device
corresponding to the block device over which the file system is mounted. For
example, to back up /usr, mounted on /dev/dsk/c1t0d0s6, enter raw device
/dev/rdsk/c1t0d0s6. Note the r in /rdsk.

Wildcards (such as /dev/rdsk/c0*) are not permitted. Specifying the actual

device file name such as /devices/pci@1f,0/pci@1/scsi@3/sd@1,0:d,raw
is not supported.

Note Advanced Client policies do not support the ALL_LOCAL_DRIVES entry in the
Backup Selections list.

11. Use the Clients tab to select clients to back up.

Each client in the client list must contain all the raw partitions that are specified in the
Backup Selections tab. Exclude and include lists cannot be used to circumvent this
requirement.

Chapter 5, FlashBackup Configuration 105

Configuring FlashBackup in the Earlier Manner (UNIX only)

Configuring FlashBackup in the Earlier Manner (UNIX only)

Prior to NetBackup 5.0, FlashBackup was a separate product with two built-in snapshot
methods: snapctl driver (nbu_snap) for Solaris clients and a VxFS snap driver for HP-UX
clients. The means of configuring a pre-5.0 FlashBackup policy was different from that
described earlier in this chapter, as follows:
Unless FlashBackup was combined with ServerFree Agent at NetBackup 4.5, the
snapshot methods were preset to the snapctl driver for Solaris clients and VxFS snap
driver for HP-UX clients.
The raw partition to be used as cache for the snapctl driver and VxFS snap driver had
to be specified as a CACHE= entry in the policys Backup Selections list.
Using multiple data streams required adding other directives to the policys Backup
Selections list.
The following sections explain how to configure a FlashBackup policy in the pre-5.0
manner. Note that this means of configuration is provided for backwards compatibility
only and will be discontinued in a future release.

Snapshot Selection in Earlier Manner

To set up the policy to use the snapctl driver for Solaris clients or VxFS snap driver for
HP-UX, leave Perform snapshot backups unselected on the policy Attributes tab.
NetBackup will use nbu_snap (snapctl driver) for Solaris clients or VxFS_Snapshot for
HP-UX.

Cache Partition in Earlier Manner

The snapctl driver and VxFS snap driver are copy-on-write snapshot methods requiring a
cache partition. For FlashBackup clients prior to NetBackup 5.0, the cache partition was
specified in the policys Files tab as a CACHE = raw_partition entry. (The Files tab is now
called the Backup Selections tab.)
This section describes how to specify the CACHE entry. (Refer to How Copy-on-Write
Works on page 249 for a description of copy-on-write cache.)

Note CACHE entries are allowed only when the policys Perform snapshot backups
option is unselected. If Perform snapshot backups is selected, NetBackup will
attempt to back up the CACHE entry and the backup will fail.

On the policys Backup Selections tab, specify at least one cache device by means of
the CACHE directive. For example:

106 NetBackup Advanced Client System Administrators Guide

 Configuring FlashBackup in the Earlier Manner (UNIX only)

CACHE=/dev/rdsk/c0t0d0s1

This is the cache partition for storing any blocks that change in the source data while
the backup is in progress. CACHE= must precede the source data entry. See following
example.

Backup Selections list with CACHE entry

CACHE entry must come

before the source data
entry.

All entries must specify the

raw device, such as
/dev/rdsk/c0t0d0s1.

Do not use the actual file

name; must be the link
form of cxtxdxsx.

Please note:
Specify the raw device, such as /dev/rdsk/c1t0d0s6. Do not specify the block
device, such as /dev/dsk/c1t0d0s6.
Also, do not specify the actual device file name. The following, for example, is not
allowed:
/devices/pci@1f,0/pci@1/scsi@3/sd@1,0:d,raw

Wildcards (such as /dev/rdsk/c0*) are not allowed.

The CACHE entry must precede the entry for the source data you want to back up.
All entries in the Backup Selections list, including the source data to back up, must be
the full path name of a raw device in the form:
On Solaris: /dev/rdsk/cxtxdxsx
/dev/rdsk/cxtxdx

where x is an integer.
When using multiple data streams, you can include multiple entries in the Backup
Selections list.
For example:
CACHE=/dev/rdsk/c1t4d0s0

Chapter 5, FlashBackup Configuration 107

Configuring FlashBackup in the Earlier Manner (UNIX only)

/dev/rdsk/c1t4d0s7

CACHE=/dev/rdsk/c1t4d0s1

/dev/rdsk/c1t4d0s3

/dev/rdsk/c1t4d0s4

Requirements for the cache partition

Must reside on the same host as the raw partitions containing the source data that
you are backing up.
Cannot be the raw partition being backed up.
Cannot be a mounted file system. Any data configured on this device may be
overwritten by the copy-on-write process.
On Solaris, the same cache partition may be used simultaneously by multiple
backups (two policies can use the same cache partition at the same time). On
HP-UX, multiple backups cannot use the same cache partition simultaneously. If
multiple policies list the same cache partition on HP-UX systems, backups
naming those policies must run at different times to prevent a failure.
The cache partition must have enough space to hold all the writes to the source
data that may occur during the backup. Backups during nonworking hours
normally require a smaller cache than during peak activity. See Sizing the Cache
Partition on page 135.

Using Multiple Data Streams with Pre-5.0 Clients

Certain directives must be added to the policys Backup Selections tab.
The number of backups started depends on the directives in the Backup Selections
tab.
The number of backups that can run concurrently depends on the number of available
drives in the storage units and the maximum jobs parameters (for example, Limit
jobs per policy).

Note Only one data stream is created for each physical device on the client. You cannot
include the same partition more than once in the Backup Selections list.

The directives that you can use in the Backup Selections list for a FlashBackup policy are:
NEW_STREAM

CACHE=value (this directive is required; see Cache Partition in Earlier Manner on

page 106 earlier in this chapter)
UNSET

108 NetBackup Advanced Client System Administrators Guide

 Configuring FlashBackup in the Earlier Manner (UNIX only)

UNSET_ALL

Each backup begins as a single stream of data. The start of the Backup Selections list up to
the first NEW_STREAM directive (if any) is the first stream. Each NEW_STREAM entry causes
NetBackup to create an additional stream or backup.
Note that all file paths listed between NEW_STREAM directives are in the same stream.
The Backup Selections list in the following example generates four backups:

On Solaris systems: On HP systems:

1 CACHE=/dev/rdsk/c1t3d0s3
1 CACHE=/dev/cache_group/rvol1c

/dev/rdsk/c1t0d0s6
/dev/vol_grp/rvol1

2 NEW_STREAM
2 NEW_STREAM

/dev/rdsk/c1t1d0s1
UNSET CACHE

3 NEW_STREAM
CACHE=/dev/cache_group/rvol2c

UNSET CACHE
/dev/vol_grp/rvol2

CACHE=/dev/rdsk/c1t3d0s4
3 NEW_STREAM

/dev/rdsk/c1t2d0s5
UNSET CACHE

/dev/rdsk/c1t5d0s0
CACHE=/dev/cache_group/rvol3c

4 NEW_STREAM
/dev/vol_grp/rvol3

UNSET CACHE
/dev/vol_grp/rvol3a

CACHE=/dev/rdsk/c0t2d0s3
4 NEW_STREAM

/dev/rdsk/c1t6d0s1
UNSET CACHE

CACHE=/dev/cache_group/rvol4c

/dev/vol_grp/rvol4

1. The first stream is generated automatically and a backup is started for

/dev/rdsk/c1t0d0s6 (Solaris) or /dev/vol_grp/rvol1 (HP). The CACHE= entry
sets the cache partition to /dev/rdsk/c1t3d0s3 (Solaris) or
/dev/cache_group/rvol1c (HP).

2. The first NEW_STREAM directive starts a second stream to back up

/dev/rdsk/c1t1d0s1 (Solaris) or /dev/vol_grp/rvol2 (HP). On Solaris
systems, this backup uses the same cache partition. On HP systems, a different cache
partition must be defined for each stream (CACHE=/dev/cache_group/rvol2c).

3. The second NEW_STREAM starts a backup for /dev/rdsk/c1t2d0s5 and

/dev/rdsk/c1t5d0s0 (Solaris) or /dev/vol_grp/rvol3 and
/dev/vol_grp/rvol3a (HP). These two partitions are backed up serially within the
stream. In addition, the UNSET CACHE directive unsets the previous cache setting and
the CACHE= directive sets a new cache partition for this backup.

Chapter 5, FlashBackup Configuration 109

Configuring FlashBackup in the Earlier Manner (UNIX only)

4. The last NEW_STREAM directive starts a backup for /dev/rdsk/c1t6d0s1 (Solaris)

or /dev/vol_grp/rvol4 (HP). Like the third stream, this one also unsets the cache
directive and defines a new cache partition.
As shown in the example, policy-specific directives such as CACHE are passed to the client
with the current stream and all subsequent streams, until the directive is redefined or
unset.
If the directive is encountered again, its value is redefined.
An UNSET or UNSET_ALL directive unsets a directive that was previously defined in the
Backup Selections list.
UNSET unsets a policy-specific directive so it is not passed with additional streams.
The directive that was unset can be defined again later in the Backup Selections list so
it is included in the current or a later stream.
UNSET_ALL has the same effect as UNSET but affects all policy-specific directives that
have been defined up to this point in the Backup Selections list. If used, UNSET_ALL
must appear immediately after the second or later NEW_STREAM directive.

110 NetBackup Advanced Client System Administrators Guide

NAS Snapshot Configuration 6
This chapter explains how to configure NetBackup policies to create snapshots of client
data on NAS (NDMP) hosts.

Note NetBackup for NDMP software is required, in addition to Advanced Client.

The following topics are covered in this chapter:

NAS Snapshot Overview

Notes and Prerequisites

Setting up a Policy for Snapshots

111
NAS Snapshot Overview

NAS Snapshot Overview

By means of the snapshot feature of Advanced Client and the NDMP V4 snapshot
extension, NetBackup can make snapshots of client data on a NAS (NDMP) host. The
client data must reside on the NAS host and be mounted on the client by means of NFS on
UNIX or CIFS on Windows.
A snapshot is a point-in-time disk image. Snapshots can be retained on disk as long as
desired. The data can be efficiently restored from disk by means of the Advanced Client
Instant Recovery feature.

NDMP snapshot environment

NetBackup for NDMP master server,

with Advanced Client software

LAN / WAN

NetBackup client,
with Advanced NAS host
Client
CIFS or NFS mount

Data is mounted on client but resides on NAS

(NDMP) host.
In NetBackup policy, enter:
Snapshot of client
for Windows client: data is made on
NAS-attached
\\ndmp_hostname\volume disk
for UNIX client:
/NFS_mountpoint

NOTE: Windows pathnames must use the

Universal Naming Convention (UNC).

NetBackup creates snapshots on the NAS-attached disk only, not on storage devices
attached to the NetBackup server or the client.

112 NetBackup for NDMP System Administrators Guide

 Notes and Prerequisites

Notes and Prerequisites

Snapshots of NAS host data are supported for NetBackup clients running Solaris 7, 8,
and 9, and Windows 2000/2003.
Both NetBackup for NDMP and Advanced Client software must be installed and
licensed on the NetBackup server; Advanced Client must be installed on NetBackup
clients. The NetBackup for Oracle database agent software is required for Oracle
clients.
The NAS host must support NDMP protocol version V4 and the NDMP V4 snapshot
extension. The NetBackup for NDMP Supported OS and NAS Appliance Information
online document contains a list of NAS vendors that NetBackup supports for NAS
snapshots, and includes requirements specific to your NAS vendor. Refer to the
preface for help accessing that document.
NetBackup must have access to each NAS host on which a snapshot will be created.
To create this authorization, you can use the NetBackup Device Configuration wizard,
or the set_ndmp_attr -auth command as explained in the NetBackup for NDMP
System Administrators Guide.
The client data must reside on a NAS host and be mounted on the client by means of
NFS on UNIX or CIFS on Windows. For NFS mounts, the data must not be
auto-mounted, but must be hard (or manually) mounted.
Only the NetBackup binary catalog is supported. For sites running the NetBackup
ASCII catalog, NetBackup will automatically create the NAS_Snapshot in the binary
catalog format.

Setting up a Policy for Snapshots

This section explains how to set up a policy for making snapshots of NAS data, using the
Advanced Client features of the NetBackup policy dialog.

1. Start the NetBackup Administration Console on the NetBackup for NDMP server as
follows:
On Windows NT/2000/2003: from the Windows Start menu, select Programs,
VERITAS NetBackup, NetBackup Administration Console.
On UNIX servers, enter the following:
/usr/openv/netbackup/bin/jnbSA &

2. Click on Policies in the left pane.

Chapter 6, NAS Snapshot Configuration 113

Setting up a Policy for Snapshots

3. For Policy type, select Standard for UNIX clients, MS-Windows-NT for Windows
clients, or Oracle for UNIX clients configured in an Oracle database.

Note The NDMP policy type is not supported for snapshots in this 5.1 release.

4. For Storage unit, select Any_available.

Note Although the policy cannot execute without a specified storage unit, NetBackup
does not use the storage unit. The snapshot is created on disk regardless of which
storage unit you select.

For Oracle policies, the policy uses the storage unit you specify, but only for backing
up archive logs and control files.

5. Select Perform snapshot backups and Retain snapshots for instant recovery.

6. Select Perform offhost backup and Use data mover.

7. From the pull-down under Use data mover, pick Network Attached Storage.
When the policy executes, NetBackup will automatically select the NAS_Snapshot
method for creating the snapshot.
As an alternative, you can manually select the NAS_Snapshot method using the
Advanced Snapshot Options dialog from the policy display. For the Maximum
Snapshots (Instant Recovery only) parameter, refer to the Policy Configuration
chapter of this guide, Selecting the Snapshot Method.

8. On the Schedule Attributes tab, select Instant recovery backups to disk only.

9. For the Backup Selections list, specify the directories or files from the client perspective,
not from the NDMP host perspective. For example:
On a UNIX client, if the data resides in /vol/vol1 on the NDMP host nas1, and
is NFS mounted to /mnt2/home on the UNIX client, specify /mnt2/home in the
policy Backup Selections list.
On a Windows client, if the data resides in /vol/vol1 on the NDMP host nas1,
and is CIFS mounted (mapped to) \\nas1\vol\vol1 on the Windows client,
specify \\nas1\vol\vol1 in the policy Backup Selections list.

Note the following:

Windows pathnames must use the Universal Naming Convention (UNC), in the form
\\server_name\share_name.

114 NetBackup for NDMP System Administrators Guide

 Setting up a Policy for Snapshots

The client data must reside on a NAS host and be mounted on the client by means of
NFS on UNIX or CIFS on Windows. For NFS mounts, the data must be manually
mounted by means of the mount command, not auto-mounted.
All paths for a given client in the policy must be valid, or the backup will fail.
The ALL_LOCAL_DRIVES entry is not allowed in the Backup Selections list.

Chapter 6, NAS Snapshot Configuration 115

Setting up a Policy for Snapshots

116 NetBackup for NDMP System Administrators Guide

Instant Recovery Configuration 7

This chapter explains how to prepare for and configure a policy that uses the Instant
Recovery feature.
The following topics are covered in this chapter:
Instant Recovery Capabilities
Requirements
Restrictions
Instant Recovery Overview
Configuring a Policy for Instant Recovery
Configuring VxVM
Instant Recovery for Databases

117
Instant Recovery Capabilities

Instant Recovery Capabilities

The Instant Recovery feature of Advanced Client enables high-speed data retrieval from
disk by means of the standard NetBackup user interface. Note the following features:
Supports NetBackup clients running on Solaris 7, 8, 9, HP-UX 11i, or Windows
2000/2003. The master server can be running on any VERITAS-supported operating
system.
Uses snapshot technologies to create disk images.
Creates a snapshot and a backup to tape, from one policy.
Enables random-access (non-sequential) restores of dispersed files from full backups.
Enables block-level restore and file promotion from images created using VERITAS
VxFS Storage Checkpoints with the VxFS_Checkpoint method.
Enables rollback from backups created using VERITAS VxFS Storage Checkpoints,
VxVM, or NAS_Snapshot method.
Can restore to a different path or host.
Provides resource management by means of a rotation schedule.
Supports Oracle, Microsoft Exchange, DB2, and SQL-Server databases.

Requirements
For snapshots using Storage Checkpoints (using NetBackups VxFS_Checkpoint
method), all clients must have VxFS 3.4 or later with the Storage Checkpoint feature.
For VxVM snapshot volumes on UNIX, clients must have VxVM 3.2 or later with the
FastResync feature. Windows clients must have Storage Foundations for Windows
version 3.1.
For Instant Recovery with DB2, Oracle, Exchange, or SQL-Server databases, refer to
the appropriate NetBackup database agent guide.
For replication hosts (using NetBackups VVR method), clients must have VxVM 3.2
or later with the VERITAS Volume Replicator feature.

Restrictions
For snapshots using Storage Checkpoints, Instant Recovery supports file systems with
the Version 4 disk layout or later. Older disk layouts must be upgraded to Version 4 or
later.

118 NetBackup Advanced Client System Administrators Guide

 Instant Recovery Overview

No-data Storage Checkpoints (those containing file system metadata only) are not
supported.
Instant Recovery snapshots must not be manually removed or renamed, otherwise the
data cannot be restored.
Block-level restore is available only when restoring files to the original location on the
original client and when the snapshot method used in the backup was
VxFS_Checkpoint. This feature requires the VxFS File System.
Alternate client backup is supported in the split-mirror configuration only, using a
mirror-type snapshot method (vxvm or VVR).
For Instant Recovery backups of data configured on VxVM volumes on Windows, the
VxVM volume names must be 12 characters or less. Otherwise, the backup will fail.

Instant Recovery Overview

Standard NetBackup can use disks for backup and restore. The Instant Recovery feature of
Advanced Client extends that capability by exploiting the speed of snapshots. A snapshot
is created with minimal impact on clients access to data or on the speed of their
transactions. If snapshots are made frequently, a user who accidentally deletes or modifies
a file can restore the file from the latest snapshot in a matter of seconds.

Instant Recovery from Snapshot on Disk

NetBackup master
server with Instant Snapshots on same

Recovery policy or different disk Time of creation

snapshot A Dec 9 07:30:16

2003

NetBackup Clients working Dec 9 08:30:11

snapshot B
client data 2003

snapshot C Dec 9 09:30:15

restore 2003

The on-disk snapshots become point-in-time versions of file systems or volumes, to be

kept or discarded as needed. The data can be restored from local disk; no need to mount a
tape or other remote storage.

Chapter 7, Instant Recovery Configuration 119

Instant Recovery Overview

In the figure above, NetBackup Instant Recovery creates snapshot A of the client data on
disk. One hour later, as scheduled, NetBackup creates snapshot B, also on disk, followed
one hour later when it creates snapshot C. When needed, a user can restore data directly
from disk, from the appropriate snapshot.

Note NetBackup Instant Recovery keeps the snapshot. The snapshot can be used for
restore even if the client has been rebooted.

The following sections provide background information on Storage Checkpoints and

snapshot volumes, and describe how NetBackup creates and maintains each type of
snapshot.

Snapshot and Backup

An Instant Recovery backup creates a snapshot on disk and optionally backs up the
clients data to a storage device. The snapshot is made in one of three places:
On the same disk file system that contains the clients original data (for VxFS file
systems, using the VxFS_Checkpoint snapshot method)
On a local mirror disk (for VxVM volumes, using the snapshot vxvm method)
On a mirror disk on a replication host (for VERITAS Volume Replication, using the
VVR snapshot method)

Maintaining the NetBackup Catalogs

For Instant Recovery backups, NetBackup automatically updates the catalog to keep it
correlated with the snapshots on the client. If not kept up-to-date, the catalog might
eventually include references to snapshots that no longer exist on the client, due to user
activity (snapshot replacement or deletion).
NetBackup includes a maintenance program (bppficorr) that can be run manually to
update the catalog, if an Instant Recovery snapshot is accidentally deleted or renamed.
For man page information on the bppficorr command, refer to the NetBackup Commands
guide.

VxFS Storage Checkpoints

Storage Checkpoints are a feature of the VERITAS File System (VxFS). They are produced
by a copy-on-write mechanism that creates a snapshot by identifying only the file system
blocks that have changed since the last checkpoint was taken. When restoring from a
storage checkpoint, NetBackup retrieves only the data that changed since the last backup,
not the entire file system, if block-level restore is applicable.

120 NetBackup Advanced Client System Administrators Guide

 Instant Recovery Overview

For more detail on Storage Checkpoints, refer to the VERITAS File System Administrators
Guide. For an introduction to the copy-on-write process, refer to How Copy-on-Write
Works on page 249.
A Storage Checkpoint has the following features:
Persists after a system reboot or failure.
Identifies the blocks that have changed since the last Storage Checkpoint.
Shares the same pool of free space as the file system. The number of checkpoints is
limited only by the available disk space.
Supports mounting a VxFS 4.0 file system over a VxVM 4.0 volume set (multi-device
system).

VxVM Split Mirrors

The second type of snapshot that Instant Recovery can make is a split mirror. This is a
complete disk copy of the clients primary volume. It is stored on separate disk(s),
physically independent of the primary volume. Since mirroring requires a complete copy
of the primary disk on a separate disk (equal in size to the original disk), it consumes more
disk space than a copy-on-write type of snapshot such as a Storage Checkpoint
(VxFS_Checkpoint).
For Instant Recovery, the vxvm snapshot method supports split-mirror snapshots.

Chapter 7, Instant Recovery Configuration 121

Configuring a Policy for Instant Recovery

Configuring a Policy for Instant Recovery

This section explains how to configure a policy for backups enabled for Instant Recovery.

1. Start the NetBackup Administration Console as follows:

On UNIX, enter: /usr/openv/netbackup/bin/jnbSA &
On Windows, click Start > Programs > VERITAS NetBackup > NetBackup
Administration Console.

2. Click Policies. In the All Policies pane, open a policy or create a new one.

Select policy type.

For backup to storage,

select appropriate
storage unit. The
storage unit is ignored
if making snapshot
only (see step 9).

Click here.

Click here to enable

Instant Recovery.

3. For the policy type, select Standard, MS-Windows-NT, or the database agent type
appropriate for the client(s).

122 NetBackup Advanced Client System Administrators Guide

 Configuring a Policy for Instant Recovery

4. Select a storage unit (disk or tape).

If you select Instant recovery backups to disk only on the Schedules tab (see step 9),
the storage unit is not used; NetBackup creates a snapshot only.

5. Select Perform snapshot backups.

6. Select Retain snapshots for instant recovery.

NetBackup retains the snapshot so that Instant Recovery can be performed from the
snapshot. A normal backup to storage is also performed, if you do not select Instant
recovery backups to disk only on the Schedules tab (step 9).

7. To save the settings, click Apply.

8. Click the Advanced Snapshot Options button to select the snapshot method.
If this is a new policy, you can skip this step to let NetBackup select the method (auto
is the default).

Select snapshot method.

Choose auto if you want
NetBackup to select the
method.

Change values if needed

If auto is the method, only
one snapshot can be
retained at a time.

a. Select a snapshot method from the pull-down list. For creating an Instant
Recovery snapshot, the available methods are:
auto (UNIX or Windows) - NetBackup selects the snapshot method.
NAS_Snapshot (UNIX or Windows) - uses the NDMP V4 snapshot extension
to create the snapshot on the NAS-attached disk. Refer to the NAS Snapshot
Configuration chapter for help in setting up a policy for NAS_Snapshot.
VxFS_Checkpoint (UNIX only) - uses VxFS Storage Checkpoint to create the
snapshot.

Chapter 7, Instant Recovery Configuration 123

Configuring a Policy for Instant Recovery

A new Storage Checkpoint is created whenever a backup using the

VxFS_Checkpoint method is executed.
vxvm (UNIX or Windows) - uses VxVM FastResync to create the snapshot.
VxVM mirrors must be configured as explained under Configuring VxVM
on page 126. Can also use instant snapshots; see VxVM Instant Snapshots
on page 142.
VVR (UNIX) - creates a snapshot of a VxVM volume on a VERITAS Volume
Replication host.

b. Change parameter values for the method, if needed. The parameters are
described under step 6 on page 92.

c. When finished, click OK.

9. Use the Schedule tab to configure a schedule.

Click here for a

snapshot only.
Leave unchecked if
you want a normal
backup to tape in
addition to a
snapshot for Instant
Recovery

a. For a snapshot only, select Instant recovery backups to disk only.

If Instant recovery backups to disk only is selected, the snapshot is not
backed up to tape or other storage. NetBackup creates a snapshot on disk
only. This option is required for the NAS_Snapshot method.

Note If the snapshot uses the VxFS_Checkpoint method or is a vxvm space-optimized

snapshot, the snapshot is created on the same device as the one containing the
original data. In that case, you may want to create another policy to back up the
data to a separate device.

124 NetBackup Advanced Client System Administrators Guide

 Configuring a Policy for Instant Recovery

If Instant recovery backups to disk only is not selected, NetBackup creates

(and keeps) a snapshot and backs up the clients data to the storage unit
specified for the policy.

b. Make other schedule selections as desired, and click OK.

10. Use the Backup Selections tab to enter the files and folders to be backed up.
If backing up Oracle database clients, refer to the NetBackup for Oracle System
Administrators Guide for instructions.
Advanced Client policies do not support the ALL_LOCAL_DRIVES entry in the
policys Backup Selections list.
If you use the Backup Policy Configuration wizard, see Backup Policy
Configuration Wizard on page 100.

11. Use the Clients tab to specify clients to be backed up by this policy.

Configuring Third Mirror Split Volume Rotation

NetBackup policies can be configured to use one or more pre-allocated VxVM mirrors as a
data protection medium, thus making one or more levels of undo available, for instant
data recovery. See Maximum Snapshots (Instant Recovery only) on page 93. Changes
made to the files on the primary file system are reflected in a mirror, until the backup
policy is again executed. At that point, further changes are reflected in the second
pre-allocated mirror, not the first. The snapshot on the first mirror is retained, in case a
restore is needed from that data. When a third backup is executed, the mirrors are
rotated again and further changes are reflected in the third mirror (if a third mirror was
allocated). If a third mirror was not allocated, the changes are reflected in the first mirror,
and the data that had been retained on that mirror is overwritten.
The next figure shows rotation of three mirror volumes for Instant Recovery. Rotation
occurs each time the Instant Recovery policy is executed.

Chapter 7, Instant Recovery Configuration 125

Configuring VxVM

VxVM Volume Mirror Rotation, For Multiple Undo Levels

Primary file system

Oldest snapshot: Most recent snapshot:

12:00 noon 4:00 PM

Older snapshot: 2:00 PM

In the above example, the next Instant Recovery backup will overwrite the mirror
snapshot that was made at 12:00 noon.
Before a mirror can be used for creating a backup for Instant Recovery, it must be
initialized (see Configuring VxVM on page 126).

Configuring VxVM
Note For Instant Recovery backups of data configured on VxVM volumes on Windows,
the VxVM volume names must be 12 characters or less. Otherwise, the backup will
fail.

Before using an Instant Recovery policy for backing up VxVM volumes, one or more
mirrors must be created. The primary volumes must be enabled for FastResync. Note that
on Windows, FastResync is enabled by default.

Creating a Snapshot Mirror

You can create a snapshot mirror using VxVM commands.

Windows
There are two ways to create the snapshot mirror:

For a volume that is associated with a drive letter:

vxassist snapstart X:
where X is the drive letter. This creates a snapshot mirror of the designated drive.

For a volume that is not associated with a drive letter:

126 NetBackup Advanced Client System Administrators Guide

 Configuring VxVM

vxdg -g disk_group dginfo

This shows information for the specified disk group, including the names of the
volumes configured for that group. Create the snapshot by entering the following:
vxassist snapstart \Device\HarddiskDmVolumes\disk_group\Volume_name

This creates a snapshot mirror of the designated Windows volume.

UNIX

1. Add dco (data change object) logs to the primary volume:

/usr/sbin/vxassist -g disk_group addlog volume_name logtype=dco

2. Enable FastResync on the volume:

/usr/sbin/vxvol -g disk_group set fmr=on volume_name

3. Prepare a new mirror for the Instant Recovery snapshot:

a. Create the mirror:

/usr/sbin/vxassist -g disk_group snapstart primary_volume

Wait until the mirror is synchronized (status SNAPDONE, or State field reads Ready
in the volumes properties display).

b. To verify that the mirror is synchronized, enter:

/usr/sbin/vxprint -g disk_group -q -t -e 'assoc="primary_volume"'

Creating an Instant Snapshot

The following procedure is required for Instant Recovery backups when using the
full-sized and space-optimized instant snapshot options in VxVM 4.0. For Instant
Recovery backups, VxVM 4.0 instant snapshots are supported by Advanced Clients
vxvm and VVR methods.

Creating space-optimized snapshots

A cache object called NBU_CACHE must be created in the disk group containing the
volume to be backed up. NetBackup will recognize the cache object and use it to create a
space-optimized snapshot.

Chapter 7, Instant Recovery Configuration 127

Configuring VxVM

1. Create the cache object:

/usr/sbin/vxassist -g disk_group make cache_volume size

layout=layout init=active

2. Label the cache object:

/usr/sbin/vxmake -g disk_group cache NBU_CACHE

cachevolname=cache_volume

Creating full-sized snapshots

Unlike the space-optimized snapshot, VxVM full-sized instant snapshots cannot be
created by NetBackup: you must create them prior to running the backup, as explained in
the following procedure. You must create one full-sized instant snapshot for each backup
you are going to run.

1. Create the parent volume:

/usr/sbin/vxassist -g disk_group make volume length layout=layout

logtype=dco dcoversion=20 [drl=no|sequential|yes] [ndcomirror=number]

[fastresync=on]

Where:
Brackets [ ] indicate optional items.
make volume specifies the name of the volume snapshot.

2. Create a volume for a full-sized instant snapshot:

a. Determine the required size for the snapshot volume:

# LEN='vxprint -g disk_group -F%len volume'

b. Find the name of the DCO volume:

# DCOVOL='vxprint -g disk_group -F%dconame volume'

c. Discover the DCO volume's region size (in blocks):

# RSZ='vxprint -g disk_group -F%regionsz $DCOVOL'

d. Create a volume called snapvol of the required size and redundance:

vxassist -g disk_group make snapvol $LEN layout=mirror

nmirror=number logtype=dco drl=no dcoversion=20

ndcomirror=number regionsz=$RSZ init=none

[storage attributes ...]

128 NetBackup Advanced Client System Administrators Guide

 Configuring VxVM

The number for nmirror should equal the number for ndcomirror.

e. Create the mirror:

vxsnap -g disk_group make source=volume/snapvol=snapvol/syncing=on

3. Set the Maximum Snapshots (Instant Recovery only) value on the Advanced
Snapshot Options dialog.

Using the VxVM Graphical User Interface

These are the steps for using the VxVM graphical user interface to configure VxVM
mirrors for Instant Recovery backups.

Using VERITAS Enterprise Administrator (VxVM 3.5)

To prepare a new mirror

1. Make sure that FastResync is enabled on the primary VxVM volume.

a. In the VEA console, right-click on the volume and click Properties from the
pop-up menu.
The FastResync field states whether or not FastResync is enabled.

b. Click Cancel.

c. If FastResync is disabled, right-click on the volume again and select

Fast Resync > Add from the pop-up menu. This enables FastResync.

2. Prepare a new mirror for Instant Recovery backup.

Chapter 7, Instant Recovery Configuration 129

Configuring VxVM

a. Create a new mirror by right-clicking on the volume and selecting Snap > Snap
Start.

b. Make sure FastResync is enabled on the mirror. Click OK to create the mirror and
start full synchronization.

3. On the Mirrors tab, ensure that synchronization has completed as indicated by Snap
Ready in the Status field.

130 NetBackup Advanced Client System Administrators Guide

 Instant Recovery for Databases

Instant Recovery for Databases

To configure an Instant Recovery policy for database clients, refer to the appropriate
NetBackup database agent guide.

Chapter 7, Instant Recovery Configuration 131

Instant Recovery for Databases

132 NetBackup Advanced Client System Administrators Guide

Snapshot Configuration Notes 8
This chapter provides notes and configuration assistance for the Advanced Client
snapshot methods.
For help using the NetBackup Administration Console to select a snapshot method for a
policy, refer to Configuring an Advanced Client Policy on page 85.
The following topics are covered in this chapter:
nbu_snap
VxFS_Checkpoint
VxFS_Snapshot
vxvm
FlashSnap
VVR
NAS_Snapshot
Array-Specific Snapshot Methods
(TimeFinder, ShadowImage, BusinessCopy)

133
nbu_snap

nbu_snap
The nbu_snap snapshot method is for Solaris clients only. It is for making
copy-on-write snapshots for UFS or VERITAS VxFS file systems.
The information in this section applies to either Standard or FlashBackup policy
types.

Note nbu_snap is not supported in clustered file systems, either as the selected snapshot
method, or as the default snapctl driver when configuring FlashBackup in the
earlier manner.

An alternative copy-on-write snapshot method, for clustered file systems, is

VxFS_Snapshot with a FlashBackup policy.

Cache device
The cache device is a raw disk partition: either a logical volume or physical disk. This
is used for storing the portions of the clients data that are changed by incoming write
requests while the copy-on-write is in progress.
Specify the actual character special device file (such as /dev/rdsk/c2t0d0s4).
nbu_snap will not work for block special device files.
Enter the full path name of the raw partition. Do not specify wildcards (such as
/dev/rdsk/c2*) as paths.
For the cache device, do not specify an active partition containing valuable data. Any
data in that partition will be lost when the nbu_snap process is complete.
The cache partition must be unmounted.
The cache partition must reside on the same host as the snapshot source (the clients
data to back up).
The partition must have enough space to hold all the writes to the partition that may
occur during the backup. Note that backups during non-working hours normally
require a smaller cache than a backup during peak activity. (See Sizing the Cache
Partition on page 135 for more suggestions on cache size.)
For the Media Server or Third-Party Copy method, the host containing the snapshot
source and cache must be visible to the media server or third-party copy device (refer
to the chapter titled SAN Configuration for Advanced Client).
For the Media Server or Third-Party Copy method, the disk containing the cache must
meet the requirements spelled out under Disk Requirements for Media Server/
Third-Party Copy on page 178.

134 NetBackup Advanced Client System Administrators Guide

 nbu_snap

Sizing the Cache Partition

The size needed for the cache partition depends on user write activity during the backup,
not on the size of the clients file system. If the backup occurs when user activity is heavy,
a larger cache is required.
Use the following procedure to determine a size for the cache partition:

1. Consider the period in which the backup will occur: the more user activity expected,
the larger the cache required.
You should execute the following procedure at an appropriate period, when your
nbu_snap backups typically run. If user activity at your site is known to vary with the
time of day, a different time could bring very different results.

2. Make sure a raw partition is available on a separate disk (see Cache device on
page 134 for cache partition requirements).

3. During the appropriate backup period, create an nbu_snap snapshot by entering the
following as root:
/usr/openv/netbackup/bin/driver/snapon snapshot_source cache

where snapshot_source is the partition on which the clients file system is mounted, and
cache is the raw partition to be used as copy-on-write cache. For example:
/usr/openv/netbackup/bin/driver/snapon /omo_cat3

/dev/vx/rdsk/zeb/cache

Example output:
matched /omo_cat3 to mnttab entry /omo_cat3

mount device: /dev/vx/dsk/omo/vol03 fstype: vxfs

snapshot 29 enabled on /omo_cat3 at 06/05/03 15:16:02

4. In /usr/openv/netbackup/bin/driver, enter the snaplist and

snapcachelist commands.
snaplist - shows the id of each snapshot, the size of the partition containing
the client file system, and the amount of file system write activity in 512-byte
blocks that occurred during the nbu_snap snapshot (under the cached column).
The more blocks cached as a result of user activity, the larger the cache partition
required.
snapcachelist - shows each cache device in use and what percentage has
been used (busy). For each cache device listed, busy shows the total space used
in the cache. This value indicates the size of the raw partition that may be
required for nbu_snap cache.

Chapter 8, Snapshot Configuration Notes 135

nbu_snap

For more details on the snap commands, refer to nbu_snap Commands on

page 255.

Note The snap commands can be used in a script.

5. If the cache partition is not large enough, the backup will fail with status code 13, file
read failed. The /var/adm/messages log may contains errors such as the
following:
Mar 24 01:35:58 bison unix: WARNING: sn_alloccache: cache

/dev/rdsk/c0t2d0s3 full - all snaps using this cache are now

unusable

6. Using the information provided by snaplist and snapcachelist, you have

several options:
Specify a larger (or smaller) partition as cache, depending on the results shown by
snaplist and snapcachelist.
Reschedule backups to a period when less user activity is expected.
If multiple backups are using the same cache, reduce the number of concurrent
backups by rescheduling some of them.

7. When finished with the snapshot, you can remove it by entering the following:
/usr/openv/netbackup/bin/driver/snapoff snapid
where snapid is the numeric id of the snapshot created at step 3.

Note A snapshot created manually with the snapon command is not controlled by a
NetBackup policy. When run manually, snapon creates a copy-on-write snapshot
only. The snapshot will remain on the client until it is removed by entering
snapoff or the client is rebooted.

How to Enter the Cache

For the nbu_snap and VxFS_Snapshot methods, you must identify a raw partition to be
used by the copy-on-write process. Depending on how the policy is configured, this can
be done in any of several ways:
If manually selecting the snapshot method on the Advanced Snapshot Options dialog,
you have two options for specifying the raw cache partition:

136 NetBackup Advanced Client System Administrators Guide

 nbu_snap

a. Use the Host Properties > Clients > Client Properties dialog > UNIX Client >
Client Settings to specify the raw partition in the Default cache device path for
snapshots field. This setting applies to the client in all policies.

b. Use the Advanced Snapshot Options dialog, Cache device path Value field. This
cache setting applies to all clients in the current policy, and overrides the cache
setting in the Client Properties dialog.
If you want NetBackup to be able to select the nbu_snap or VxFS_Snapshot methods
by means of the auto method, specify the cache on the Host Properties > Clients >
Client Properties dialog > UNIX Client > Client Settings as described above.
In a FlashBackup policy: if Perform snapshot backups is NOT selected, you must use
a CACHE= directive in the Backup Selections tab. This cache setting applies to all
clients in the current policy, and overrides the cache setting in the Host Properties
dialog. (This means of configuring cache will be discontinued in a future release.)

Chapter 8, Snapshot Configuration Notes 137

VxFS_Checkpoint

VxFS_Checkpoint
The VxFS_Checkpoint snapshot method is for making copy-on-write snapshots (Solaris or
HP only). This is one of several snapshot methods that support Instant Recovery backups.
Note that for VxFS_Checkpoint, the Instant Recovery snapshot is made on the same disk
file system that contains the clients original data.
For VxFS_Checkpoint, VxFS 3.4 or later with the Storage Checkpoints feature must be
installed on the NetBackup clients (HP requires VxFS 3.5).
The VxFS_Checkpoint method is not supported for backing up raw partitions
(whether FlashBackup or Standard policies).
Make sure there is enough disk space available for the checkpoint. The file system
containing the snapshot source should have at least 10% free space in order to
successfully implement the checkpoint.

VxFS Multi-Device System

VxFS_Checkpoint is the only snapshot method in Advanced Client that supports the
multi-device system (MDS) feature of VxFS 4.0.
The multi-device file system feature requires a VxVM 4.0 volume set. With volume sets,
you can group related volumes into one volume set, and mount a VxFS file system on it.
This means that a VxFS file system can be mounted on more than one volume. This
feature allows file systems to make best use of the different performance and availability
characteristics of the underlying volumes. For example, file system metadata could be
stored on volumes with higher redundancy, and user data on volumes with better
performance.
For background and configuration assistance with multi-device file system, refer to the
VERITAS File System 4.0 Administrators Guide and the VERITAS Volume Manager 4.0 Administrators
Guide.

Note Offhost backup is not supported for a VxFS 4.0 multi-device system.

Storage Checkpoint Disk Usage

The ls command does not list Storage Checkpoint disk usage. This means that the
primary volume may appear to have available space even if it is full. You must use the
fsckptadm list command to show Storage Checkpoint disk usage. Refer to the
VERITAS File System Administrators Guide for more information on fsckptadm.

138 NetBackup Advanced Client System Administrators Guide

 VxFS_Checkpoint

Note A new Storage Checkpoint is created whenever a VxFS_Checkpoint policy is

executed.

Checkpoint Retention Schedules

For Instant Recovery storage checkpoints, no data is moved between the client and the
media server. Instead, a VxFS Storage Checkpoint is created on the client, and, for Oracle
clients, the file and directory names are sent to the server for the catalog. In the case of file
systems only (non-database clients), only the directory name is sent, not the file data.
Changes made to the files on the clients primary fileset are reflected in the Storage
Checkpoint until the backup policy is again executed, creating another Storage
Checkpoint. Storage Checkpoints are created and retained until the maximum
checkpoints threshold is exceeded, when the oldest checkpoint is removed.
The next figure shows the Instant Recovery checkpoint expiration schedule. If the
maximum checkpoint value is set to 2 and a third checkpoint is created, the oldest is
removed.

Instant Recovery Retention Schedule for Storage Checkpoints.

Clients 3rd 1st (oldest)

primary checkpoint 2nd checkpoint is
fileset (most recent) checkpoint removed

A new Storage Checkpoint is created

each time a backup is started.

Block-Level Restore
If only a small portion of a file system or database changes on a daily basis, full restores
are unnecessary. The VxFS Storage Checkpoint mechanism keeps track of data blocks
modified since the last checkpoint was taken. Block-level restores take advantage of this
by restoring only changed blocks, not the entire file or database. This leads to faster
restores when recovering large files.
Refer to Instant Recovery: Block-Level Restore (UNIX Clients Only) on page 185 for
instructions on setting up this feature.

Chapter 8, Snapshot Configuration Notes 139

VxFS_Snapshot

VxFS_Snapshot
The VxFS_Snapshot method is for making copy-on-write snapshots of local Solaris or
HP-UX clients. Offhost backup is not supported with this snapshot method.

Note VxFS_Snapshot supports the FlashBackup policy type only.

Please note the following:

The VxFS_Snapshot method can only be used to back up a single file system. If
multiple file systems are specified in the policys Backup Selections list when using
this method, the backup will fail.

Note In a FlashBackup policy, if the Backup Selections list is configured with CACHE=
entries (see Configuring FlashBackup in the Earlier Manner (UNIX only) on
page 106), FlashBackup does support the backup of multiple file systems from a
single policy. For each file system, a separate cache must be designated with the
CACHE= entry.

You must designate a raw partition to be used for copy-on-write cache.

Raw partition example:
Solaris:
/dev/rdsk/c1t0d0s3
HP:
/dev/rdsk/c1t0d0
See Cache device on page 134 for general requirements for cache partitions. See also
How to Enter the Cache on page 136.
VxFS_Snapshot is the default snapshot method for FlashBackup clients running
HP-UX, when Perform snapshot backup is not selected for the backup policy.

140 NetBackup Advanced Client System Administrators Guide

 vxvm

vxvm
The vxvm snapshot method is for making mirror snapshots with VERITAS Volume
Manager 3.1 or later snapshot mirrors. (On Windows, make sure that VxVM has the
latest VxVM service packs and updates.)
The vxvm snapshot method works for any file system mounted on a VxVM volume.
However, before the backup is performed, the data must be configured with a VxVM
3.1 or later snapshot mirror or a VxVM 4.0 or later cache object (otherwise, the backup
will fail).
For help configuring a snapshot mirror, refer to Creating a Snapshot Mirror of
the Source, below, or refer to your VERITAS Volume Manager documentation.
For help configuring a cache object, refer to your VERITAS Volume Manager
documentation, and to VxVM Instant Snapshots on page 142.
For Instant Recovery backups of data configured on VxVM volumes on Windows,
the VxVM volume names must be 12 characters or less. Otherwise, the backup
will fail.

Note Since VxVM does not support fast mirror resynchronization on RAID 5 volumes,
the vxvm snapshot method must not be used with VxVM volumes configured as
RAID 5. If the vxvm snapshot method is selected for a RAID 5 volume, the backup
will fail.

Creating a Snapshot Mirror of the Source

To use the vxvm snapshot method with VxVM volumes in the third-mirror (split-mirror)
configuration, you must create a snapshot mirror of the snapshot source before making a
backup. Before specifying a snapshot source as described in this chapter, create a snapshot
mirror on the client by using one of the following methods:
In the Volume Manager Storage Administrator interface:
UNIX: select the volume that will be the snapshot source, right click on it, and select
Snapshot from the pop-up menu. In the Volume Snapshot dialog, select Enable FMR
(if available, see note below) and click the Snapstart button.
Windows: select the volume that will be the snapshot source, right click on it, select
Snap, and choose Snapstart.
For details, refer to your VERITAS Volume Manager documentation.
Or (for UNIX):
Enter the following commands:
/usr/sbin/vxassist -g disk_group snapstart volume_name

Chapter 8, Snapshot Configuration Notes 141

vxvm

/usr/sbin/vxvol -g disk_group set fmr=on volume_name

where:
disk_group is the Volume Manager disk group to which the volume (snapshot
source) belongs.
volume_name is the name of the volume designated at the end of the snapshot
source path (for example, vol1 in /dev/vx/rdsk/dg/vol1).
fmr=on sets the Fast Mirror Resynchronization attribute, which resynchronizes
the mirror with its primary volume but only copies the blocks that have changed,
rather than performing a full resynchronization. Fast mirror resynchronization
can dramatically reduce the time required to complete the backup.
Fast Mirror Resynchronization (FMR) is a separately available product for
VERITAS Volume Manager.

Note If the Media Server or Third-Party Copy method is used, the disks that make up the
disk group must meet the requirements spelled out under Disk Requirements for
Media Server/ Third-Party Copy on page 178.

VxVM Instant Snapshots

Advanced Client supports two additional kinds of snapshot volumes included in Volume
Manager 4.0: full-sized instant snapshots and space-optimized instant snapshots. These
offer advantages over traditional third-mirror snapshots, such as immediate availability
and easier configuration and administration.
Full-sized instant snapshot
This is a variation on the third-mirror volume snapshot model. It makes a snapshot
volume available for access as soon as the snapshot plexes have been created. Like
traditional third-mirror volumes, this volume, after resynchronization, can be moved
into a separate disk group or turned into an independent volume.
Space-optimized instant snapshot
This type of snapshot volume contains only the blocks that changed during the
snapshot, and uses a storage cache (cache object) to store them. The size of this cache
may be configured when the snapshot is created. This volume cannot be moved into a
separate disk group or turned into an independent volume.
Refer to VERITAS Volume Manager 4.0 Administrators Guide for complete descriptions of instant
snapshot volumes and for configuration assistance.

142 NetBackup Advanced Client System Administrators Guide

 vxvm

NetBackup Snapshot Methods

The instant snapshot feature of VxVM 4.0 is supported by NetBackups vxvm, FlashSnap,
and VVR methods. Please note:
FlashSnap supports full-sized instant snapshots only.
Space-optimized snapshots are not supported for offhost backup.
Apart from configuring the VxVM volumes and selecting vxvm, FlashSnap, or VVR as the
NetBackup snapshot method, there are no special parameters in NetBackup to be
specified.

Space-Optimized Instant Snapshots

When using the space-optimized snapshot feature of VxVM, you must create a cache
object for the snapshot. For the procedure, refer to Creating space-optimized snapshots
on page 127.

Chapter 8, Snapshot Configuration Notes 143

FlashSnap

FlashSnap
FlashSnap uses the Persistent FastResync and Disk Group Split and Join features of
VERITAS Volume Manager (VxVM).
The FlashSnap snapshot method can be used for alternate client backups only, in the split
mirror configuration, which is described under Split mirror on page 11.

Note FlashSnap supports VxVM full-sized instant snapshots but not space-optimized
snapshots. For more information, refer to VxVM Instant Snapshots on page 142.

Testing Volumes for FlashSnap

Before running an alternate client backup with the FlashSnap snapshot method, you
should test your FlashSnap volume configuration as described below. The goal of this
procedure is to ensure that the disk(s) containing the volume can be moved (deported and
imported) to the alternate client without errors. For instance, if the disks containing the
backup volumes contain some part of a volume that has another part on a disk not in the
backup disks, then the backup disks cannot be split off into a new disk group and
deported.1

On UNIX
The following steps are described in more detail in the VERITAS FlashSnap Point-In-Time Copy
Solutions Administrators Guide.

1. On the primary host:

a. If not already done, create a snapshot mirror:

vxassist -g diskgroup -b snapstart volume

b. Add a DCO log to the volume:

vxassist -g diskgroup addlog volume logtype=dco

c. Enable FastResync on the volume:

vxvol -g diskgroup set fastresync=on volume

d. Create a snapshot volume from the primary volume:

vxassist -g diskgroup snapshot volume snap_volume

1. Deporting a disk group means disabling access to that disk group. See the Volume Manager
Administrators Guide for more information on deporting disk groups.

144 NetBackup Advanced Client System Administrators Guide

 FlashSnap

e. Move the disks containing the snapshot volume to a separate (split) disk group:
vxdg split diskgroup split_diskgroup snap_volume
If the volume has not been properly configured, you may see an error similar to
the following:
host-name# vxdg split lhdvvr lhdvvr_split SNAP-emc_concat

vxvm:vxdg: ERROR: vxdg split lhdvvr lhdvvr_split failed

vxvm:vxdg: ERROR: emc_dis05 : Disk moving, but not all

subdisks on it

Look again at the layout of the disks and the volumes assigned to them, and
reassign the unwanted volumes to other disks as needed. Consult the VERITAS
FlashSnap Point-In-Time Copy Solutions Administrators Guide for examples of disk groups
that can and cannot be split.

f. Deport the split disk group:

vxdg deport split_diskgroup
2. On the secondary host:

a. Import the disk group that was deported from the primary:
vxdg import split_diskgroup

b. Enable the imported volume:

vxdg recover -g split_diskgroup -m snap_volume

c. Start the volume:

vxvol -g split_diskgroup start snap_volume
If the above commands execute successfully, the volume setup is correct.

Note After doing this test, you must re-establish the original configuration to what it was
prior to testing the volumes. 1, deport the disk group on the alternate client, then 2,
import the disk group on the primary client, and 3, recover and join the original
volume group. For directions, refer to For FlashSnap (Solaris and HP): on
page 209.

On Windows
1. On the primary host:

a. If not already done, create a snapshot mirror:

vxassist snapstart \Device\HarddiskDmVolumes\diskgroup\volume

Chapter 8, Snapshot Configuration Notes 145

FlashSnap

b. Create a snapshot volume from the primary volume:

vxassist snapshot \Device\HarddiskDmVolumes\diskgroup\volume

DrivePath=C:\Temp\Mount SNAP-Volume

c. Move the disks containing the snapshot volume to a separate (split) disk group.
Disk group is also deported after this command completes:
vxdg -g DskGrp -n SPLIT-DskGrp split

\Device\HarddiskDmVolumes\diskgroup\snap_volume

2. On the secondary host:

a. Rescan to make the deported disk group visible on the secondary host:
vxassist rescan

b. Import the disk group that was deported from the primary:
vxdg -g split_diskgroup import

c. Assign the snapshot volume to an empty NTFS directory.

This example uses C:\Temp\Mount.
vxassist assign

\Device\HarddiskDmVolumes\split_diskgroup \snap_volume

DrivePath=C:\Temp\Mount

If the above commands execute successfully, the volume setup is correct.

146 NetBackup Advanced Client System Administrators Guide

 VVR

VVR
The VVR snapshot method relies on the VERITAS Volume Replicator, which is a licensed
component of VxVM. The Volume Replicator maintains a consistent copy of data at a
remote site. Volume Replicator is described in the VERITAS Volume Replicator Administrators
Guide.

The VVR snapshot method can be used for alternate client backups only, in the data
replication configuration, described under Data Replication (UNIX Only) on page 13.
VVR makes use of the VxVM remote replication feature. The backup processing is done
by the alternate client at the replication site, not by the primary host or client.

Note VVR supports VxVM instant snapshots. For more information, refer to VxVM
Instant Snapshots on page 142.

Set Up Volume Replication

Before doing a replication backup with VVR, make sure to configure the Volume
Replicator as explained in the Volume Replicator Administrators Guide.

Name Registration
Inband Control (IBC) messages are used to exchange control information between
primary and secondary hosts. A name has to be registered at both primary and secondary
host for each replicated volume group before IBC messaging can be used. The VVR
snapshot method assumes that the application name is APP_NBU_VVR. To avoid an
initial backup failure, you should register that name as described in step 1 on page 148.

Note If APP_NBU_VVR is not registered, NetBackup will register the name when the
first backup is attempted, but the backup will fail. Subsequent backups, however,
will succeed.

Primary/Secondary Disk Group and Volume Names

For the VVR snapshot method, the disk group and volume must have the same name on
both the primary and secondary host. If the names are different, the VVR backup will fail.

Chapter 8, Snapshot Configuration Notes 147

VVR

Test the Replication Setup

Before running an alternate client backup with VVR, you should test your replication
setup as follows.

1. On both primary and secondary host, register the APP_NBU_VVR name:

vxibc -g diskgroup register APP_NBU_VVR replicated_group
The above command must be executed twice, once on each host.

2. On the primary host, send an IBC message to the secondary host:

vxibc -g diskgroup send APP_NBU_VVR replicated_group

replication_link

3. On the secondary host, receive the IBC message from the primary host:
vxibc -g diskgroup -R10 receive APP_NBU_VVR replicated_group

4. On the secondary host, restart replication:

vxibc -g diskgroup unfreeze APP_NBU_VVR replicated_group
If the above commands execute successfully, the replication setup is correct.

148 NetBackup Advanced Client System Administrators Guide

 NAS_Snapshot

NAS_Snapshot
NetBackup can make point-in-time snapshots of data on NAS (NDMP) hosts using the
NDMP V4 snapshot extension. The snapshot is stored on the same device that contains the
NAS client data. From the snapshot, you can restore individual files or roll back a file
system or volume by means of the Instant Recovery feature.

Note NetBackup for NDMP software is required on the server, and the NAS vendor must
support the NDMP V4 snapshot extension.

For help in setting up a policy for NAS snapshot, see the NAS Snapshot Configuration
chapter.

Chapter 8, Snapshot Configuration Notes 149

Array-Specific Snapshot Methods

Array-Specific Snapshot Methods

This section describes the snapshot methods that are designed for particular hardware
arrays.

Configuration Checklist
This checklist includes major caveats and important information. READ THIS TABLE
before setting up your disk arrays for the array-specific snapshot methods. The right
column refers to sources of further information.

CHECK THE FOLLOWING ! Refer to these topics for help

If you want your client data configured over Volume Manager Refer to the NetBackup Release Notes, or
volumes, make sure your arrays and operating system are see Advanced Client Information on
supported by Volume Manager (VxVM). the Web on page xix.

Make sure the client data is correctly mirrored to secondary disks See Configuring Primary and
in the array. Secondary Disks on page 156.

When configuring a backup policy, be sure to select a snapshot See The Snapshot Methods on
method that supports your arrays. page 151.

For NetBackup Media Server or Third-Party Copy Device offhost See Disk Configuration
methods: ask your array support technician to configure your Requirements on page 154.
array as follows:
The NetBackup clients must have access to primary and
secondary disks in the array.
The media server must have access to the secondary disks.

Solaris: If client data is configured over Volume Manager volumes, See Disk Configuration
label all secondary disks using the format command (label Requirements on page 154.
option).

Solaris: The EMC Symmetrix array must be configured in Common See Multiple Connectivity to EMC
Serial Number Mode to support multiple client SCSI and/or fibre Array: Common Serial Number mode
channel connections. on page 156

Do not include secondary disks in a Volume Manager disk group. See Disk Types on page 168.
Be sure to follow this and other restrictions when using Volume
Manager.

Read the Best Practices section. See Best Practices on page 172.

150 NetBackup Advanced Client System Administrators Guide

 Overview

Overview
This section describes the array-specific snapshot methods provided in Advanced Client,
explains the need for data mirroring, and introduces terms used in this chapter.

The Snapshot Methods

For mirror-type snapshot backups on supported disk arrays, select the TimeFinder,
ShadowImage, or BusinessCopy snapshot method according to the vendor-type of disk
array that contains the client data. If the snapshot method does not match the vendor-type
of the array, the backup will fail.
TimeFinder
The TimeFinder snapshot method is for making mirror snapshots on EMC Symmetrix
disk arrays with TimeFinder Symapi (with or without VERITAS Volume Manager 3.1
or later). Supports UFS and VxFS file systems, VxVM logical volumes, and raw
partitions.
ShadowImage
The ShadowImage snapshot method is for making mirror snapshots on Hitachi Data
Systems (HDS) disk arrays with ShadowImage (HOMRCF). Supports UFS and VxFS
file systems, VxVM logical volumes, and raw partitions.
BusinessCopy
The BusinessCopy snapshot method is for making mirror snapshots on HP XP series
disk arrays with BusinessCopy Services. Supports UFS and VxFS file systems, VxVM
logical volumes, and raw partitions.

Chapter 8, Snapshot Configuration Notes 151

Overview

As shown in the following table, each of the array methods must be used for its own
array-type.

Match Snapshot Method to Type of Array

To back up the following: Use this snapshot method:

EMC Symmetrix disk arrays TimeFinder

Hitachi disk arrays ShadowImage

HP XP disk arrays BusinessCopy

These snapshot methods cannot be switched: for example,

selecting TimeFinder to back up an Hitachi array will cause
the backup to fail.
For the latest information on supported disk arrays, see
Advanced Client Information on the Web on page xix.

Configuration of client data over VxVM volumes is supported only on certain

combinations of disk arrays and platforms. You can go to the VERITAS support web site
to locate an up-do-date list of VxVM supported arrays/platforms (see Advanced Client
Information on the Web on page xix).
If client data is not configured over VxVM, all above arrays are supported (no
restrictions).

Note As an alternative, the vxvm snapshot method can be used in backing up any of the
above disk arrays on either Solaris or HP, if the client data is configured over
Volume Manager volumes.

152 NetBackup Advanced Client System Administrators Guide

 Overview

Client Data Must Be Mirrored

When NetBackup makes a mirror-type snapshot, the client data on the primary disk must
be mirrored on a secondary disk prior to the backup.

For mirror snapshot, Done by the system

client data must be on administrator, or by the
mirror disk that has synchronize mirror before
been: backup option. For
Primary disk Mirror is information on this option,
containing 1, synched (brought
synched with see step 6 on page 92.
client data up-to-date with primary
primary disk
disk)

and

2, split (separated from Done automatically

Primary disk Mirror is split primary, no further by NetBackup
containing from primary writes allowed to
client data disk mirror).

If the clients data is distributed across two or more primary disks by means of a VxVM
volume, an equal number of mirror disks must also contain the same data.

NetBackup client data in Clients data must

vol_1 on vol_1 on be distributed on
VxVM volumes:
primary disk 1 mirror disk 1 mirror disks just as
where vol_1 is distributed
on primary disks
across primary disk_1
and disk_2
vol_1 on vol_1 on
primary disk 2 mirror disk 2

Disk Terms
The terms used in this manual for array disk mirroring are primary and mirror (or primary
and secondary). Some array vendors refer to these as follows:
EMC: The primary is called the standard, and the mirror is called a BCV.

Chapter 8, Snapshot Configuration Notes 153

Disk Configuration Requirements

Hitachi and HP: Primary and secondary are called primary volume and secondary
volume.

Disk Configuration Requirements

Contact the arrays support technicians for help in configuring arrays to your
specifications.

Access to Disk Arrays

For NetBackup Media Server or Third-Party Copy Device methods only: NetBackup
clients must have access to both primary and secondary disks (via SCSI or fibre
channel, or both). If the clients do not have access to both primary and secondary
disks, the backup will fail.
For NetBackup Media Server or Third-Party Copy Device methods only: the media
server requires access to the secondary disks only (via SCSI or fibre channel, or both).
If the media server does not have access to the secondary disks, the backup will fail.

Note Although the above configuration is required by NetBackup, a support technician

for your disk array vendor must configure this access.

Connection to Disk Array: SCSI and Fibre Channel

NetBackup supports three configurations, each requiring setup assistance from your array
vendor. Note that Fibre Channel and SCSI are both supported.

Configuration 1: Local SCSI (no Fibre Channel)

LAN

NetBackup NetBackup
client media server/
alternate client
Arrays

SCSI SCSI
primary mirror
disk disk

154 NetBackup Advanced Client System Administrators Guide

 Disk Configuration Requirements

Configuration 2: Array on Fibre Channel

LAN

NetBackup NetBackup
client media server/
alternate client

Fibre Channel

Arrays

primary mirror
disk disk

Configuration 3: Array Behind Bridge; Bridge on Fibre Channel

LAN

NetBackup NetBackup
client media server

Fibre Channel

Bridge

SCSI

Arrays

primary mirror
disk disk

Chapter 8, Snapshot Configuration Notes 155

Configuring Primary and Secondary Disks

Multiple Connectivity to EMC Array: Common Serial Number

mode
EMC Symmetrix disk arrays with multiple connections from NetBackup clients and
media servers (multiple SCSI or both fibre channel and SCSI) must be configured in
Common Serial Number Mode.
If there are multiple SCSI connections (or fibre channel and SCSI connections) to the same
disks in the array, NetBackup and Volume Manager will be presented with two different
serial numbers for the same disk: one for the SCSI path and one for the fibre channel path.
As a result:
Volume Manager will be confused and will not be able to run in DMP (Dynamic
Multipath) mode.
NetBackup data movement services will be confused, since the disk serial number is
used to identify the proper disk to back up.
To prevent these problems, the array must be configured in Common Serial Number
Mode.

Caution If Common Serial Number Mode is not configured for an EMC disk array that
has multiple client and media server connections, the backup may fail.

Configuring Primary and Secondary Disks

The following is a brief listing of the commands required for making the
primary-to-mirror disk association, for EMC, Hitachi, and HP arrays. The
primary-to-mirror association can be set up before or after installation of NetBackup
Advanced Client, but must be done prior to running the backup.The primary-to-mirror
association is done on the NetBackup client only.

Note If a mirror disk is not correctly associated and synchronized with the primary disk,
a snapshot of the clients data cannot be made. (A snapshot has to be made on the
mirror, not on the primary.) In that case, if the backup policy is configured with a
mirror-type snapshot, the backup will fail.

156 NetBackup Advanced Client System Administrators Guide

 Configuring Primary and Secondary Disks

EMC Symmetrix
For an EMC Symmetrix disk array on the NetBackup client, you need to create device
groups, add primary and mirror (secondary) devices to the groups, and associate or pair
the primaries with the secondaries. Once associated, the secondary disks must be
synchronized with the primary disks. During synchronization, the primary disks are
copied to the secondaries.
Use the following commands.

Note Please refer to your EMC TimeFinder SYMCLI documentation for more details on
these commands.

symdg
Creates a disk group.

symld
Adds primary disk to the disk group.

symbcv
Associates a secondary disk with a primary disk.

Create the EMC Symmetrix disk groups

1. Create a disk group that will contain any number of primary and secondary disks.
symdg create nbfim_test
Creates disk group named nbfim_test.

2. Add primary disks to the disk group.

symld -g nbfim_test add dev 02A
Adds primary disk 02A to disk group nbfim_test.

3. Add secondary (BCV) disks to the disk group.

symbcv -g nbfim_test associate dev 08C BCV001
Adds the secondary disk 08C to the same disk group.

4. Synchronize the secondary disks with the primaries in the disk group.
symmir -g nbfim_test establish

Chapter 8, Snapshot Configuration Notes 157

Configuring Primary and Secondary Disks

Pairs, or associates, the primary with the mirror, and synchronizes the mirror with the
primary. If there are multiple primaries and mirrors, they are paired according to the
order in which they were added to the group.

5. Show the output.

symmir -g nbfim_test query
When the above commands are successfully entered, NetBackup can execute snapshot
requests involving primary device 02A and its associated mirror 08C.

Hitachi and HP Arrays

The procedure for setting up Hitachi arrays is identical to that for HP arrays. For more
detail on the commands and files described below, refer to your Hitachi Data Systems or
HP documentation.
The basic steps are these (details follow), entered on the NetBackup client only:

1. Create array configuration files.

2. Add array service names to /etc/services file.

3. Start the RAID Manager daemons.

4. Set the instance number and enable the listing of output.

5. View the state of the arrays.

6. Configure the arrays, depending on your requirements.

Create array configuration files

You need a configuration file for each set of primary disks, and another file for each set of
mirror (secondary) disks. The entries in the file must be space-delimited.

1. Create a configuration file for your primary disks. Use this path and file name:
/etc/horcmX.conf

where X is an integer. For example: /etc/horcm0.conf. This integer is called the

instance number.

2. Create a configuration file for your mirror disks, using the same path and file name as
above, but with a different instance number.
For example: /etc/horcm1.conf

158 NetBackup Advanced Client System Administrators Guide

 Configuring Primary and Secondary Disks

Following are two example files. Note that entries must be separated by spaces.
Except for comment lines (#), the file must contain the HORCM_MON, HORCM_CMD,
HORCM_DEV, and HORCM_INST parameters, followed by appropriate entries
(explained below).

Example 1: configuration file /etc/horcm0.conf Entries must be space

for three primary disks delimited

HORCM_MON
Host where the
configuration file #host service poll(10ms) timeout(10ms)
resides.
turnip horcmgr0 1000 3000
Port name for
this instance.
HORCM_CMD

#cmd_dev_file cmd_dev_file cmd_dev_file

/dev/rdsk/c2t8d14s2

HORCM_DEV

#dev_group dev_name port# TargetID LU# MU#

One line per

primary disk. wiltest dev1 CL1-A 8 0
wiltest dev2 CL1-A 8 1
wiltest dev3 CL1-A 8 2

HORCM_INST
#dev_group partner host partner service
Also enter this in wiltest turnip horcmgr1
/etc/services file

HORCM_MON

Enter values for the following:

host: the NetBackup client where this configuration file resides. The NetBackup
client accesses the disks specified in this file under HORCM_DEV, when backing up or
restoring data using the ShadowImage or BusinessCopy snapshot method.

Chapter 8, Snapshot Configuration Notes 159

Configuring Primary and Secondary Disks

service: the port name of the RAID Manager instance (for this configuration file) to be
registered in the /etc/services file.
poll: the interval at which the disks are monitored, expressed as tens of milliseconds.
timeout: time-out period for attempting to communicate with the partner service,
expressed as tens of milliseconds.
HORCM_CMD

Enter values for the following:

cmd_dev_file: the command device file(s) for the array. For example,
/dev/rdsk/c2t8d14s2.
You can use the NetBackup bptpcinfo command to determine the command device
file, as follows:
bptpcinfo -d /dev/rdsk -o- | grep CM

Below is sample output showing a command device file for an Hitachi device and for
an HP device.

Command
device files p=/dev/rdsk/c2t8d14s2 s=HITACHI:OPEN-9-CM:60159001C00
(note CM): p=/dev/rdsk/c2t5d35s2 s=HP:OPEN-3-CM:30436002500

The format of the output is:

p=/dev/rdsk/c#t#d#s2s=VID:PID:SN
where:
VID (vendor ID) must be HP or HITACHI.
PID (product ID) must include -CM.
The first five characters of the serial number (SN) must match the serial number
of the disks.
In this example, the command device file for the Hitachi array is
/dev/rdsk/c2t8d14s2 and for the HP array it is /dev/rdsk/c2t5d35s2.
HORCM_DEV

Enter values for the following:

dev_group: a user-defined name of a logical grouping of primary and secondary
disks.
dev_name: a user-defined name assigned to a primary-secondary pair of disks within
the logical group.

160 NetBackup Advanced Client System Administrators Guide

 Configuring Primary and Secondary Disks

The dev_group and dev_name parameters are used on the pair configuration
commands described later in this section.
port #: the port number specified for the disk, configured by means of the arrays
dedicated console (not from a NetBackup host).
Target ID: the SCSI or fibre channel target ID number of the disk, configured by
means of the arrays dedicated console (not from a NetBackup host).
LUN: the SCSI or fibre channel logical unit number of the disk, configured by means of
the arrays dedicated console (not from a NetBackup host).
MU: a numeric mirror descriptor for cascading disks (default 0). If you are not using
cascading disks, this value may be left blank. A cascading disk has more than one
mirror (secondary) associated with a given primary.
HORCM_INST

Enter values for the following:

dev_group: same as under HORCM_DEV.
partner host: the host where the corresponding secondary (or primary)
configuration file resides (may be the same as the host specified under HORCM_MON).
For this example, the host and partner host are both turnip. (See under partner
service for a discussion of partner.)
partner service: the port name of the RAID Manager instance for the
corresponding secondary (or primary) configuration file, to be registered in the
/etc/services file.
For the example /etc/horcm0.conf file, the partner service for horcmgr0 (entered
under HORCM_MON, service) is horcmgr1. For the secondary-disk configuration
example /etc/horcm1.conf file (below), the partner service is the opposite:
horcmgr0.
Partner is a relative term. From the viewpoint of the configuration file for the primary
disks (/etc/horcm0.conf file), the partner file would be /etc/horcm1.conf (for
the secondary disks). It is the same with partner service and partner host: each refers
to the secondary from the viewpoint of the primary, or to the primary from the
viewpoint of the secondary.

Note The partner service value must be entered in the /etc/services file.

Chapter 8, Snapshot Configuration Notes 161

Configuring Primary and Secondary Disks

Example 2: configuration file /etc/horcm1.conf, for Entries must

three secondary disks be space
delimited
Contains same HORCM_MON

parameters
(HORCM_MON, etc.) #host service poll(10ms) timeout(10ms)

as in config file for

primary disks.
turnip horcmgr1 1000 3000

Disk-related entries
refer to the
secondary disks. HORCM_CMD
#cmd_dev_file cmd_dev_file cmd_dev_file
/dev/rdsk/c2t8d14s2
Port name for this
instance
HORCM_DEV

#dev_group dev_name port# TargetID LU# MU#

One line per wiltest dev1 CL2-A 16 32

secondary
disk wiltest dev2 CL2-A 16 33

wiltest dev3 CL2-A 16 34

HORCM_INST

#dev_grouppartner host partner service

Also enter this in wiltest turnip horcmgr0

/etc/services file

See under first example configuration file (/etc/horcm0.conf) for a description of

these entries.

Add array service names to /etc/services file

The values listed under service in the configuration files (horcmgr1 and horcmgr0 in
the above examples) must be entered in /etc/services file.

Restart the inetd daemon

For example:
kill -SIGHUP pid_of_inetd

162 NetBackup Advanced Client System Administrators Guide

 Configuring Primary and Secondary Disks

Start the RAID Manager daemons

Enter the following command to start the RAID Manager daemons:
/bin/horcmstart.sh x x

where x is the instance number of each configuration file. For the above example files, the
command would be:
/bin/horcmstart.sh 0 1

The daemons must be running in order to configure your primary and secondary disks.

Set the instance number and enable the listing of output

If you are using the Bourne shell, and the instance number for your primary disks is 0,
enter the following:
HORCMINST=0

HORCC_MRCF=1

export HORCMINST HORCC_MRCF

If using the C shell, enter the following:

setenv HORCMINST 0

setenv HORCC_MRCF 1

The HORCMINST parameter determines three things:

The array to which commands will be sent.
Which disk is the primary and which is the secondary, when using the paircreate
command (described below).
Which disk (primary or secondary) is listed first in each pair when using the
pairdisplay command to view the state of the arrays (described below). In this
example (HORCMINST=0), the primaries are listed first. That is because the
configuration file that defines the primary disks is named /etc/horcm0.conf, with
0 as the instance number.

View the state of the arrays

1. To display status information on all the disks, enter the following:

pairdisplay -g groupname -CLI -fc
where groupname is the name specified in the configuration files under dev_group.
CLI and fc are options:
-CLI formats headers and columns in the resulting display.
-fc includes the percentage of synchronization progress in the display.
For example:

pairdisplay -g wiltest -CLI -fc

Chapter 8, Snapshot Configuration Notes 163

Configuring Primary and Secondary Disks

Resulting output:
Group PairVol L/R Port# TID LU-M Seq# LDEV# P/S Status % P-LDEV# M

wiltest dev1 L CL1-A 8 0 0 60159 0 P-VOL PAIR 100 43

wiltest dev1 R CL2-A 16 32 0 60159 43 S-VOL PAIR 100 0

wiltest dev2 L CL1-A 8 1 0 60159 1 P-VOL PSUS 99 44 W

wiltest dev2 R CL2-A 16 33 0 60159 44 S-VOL SSUS 99 1 -

wiltest dev3 L CL1-A 8 2 0 60159 2 SMPL - - -

wiltest dev3 R CL2-A 16 34 0 60159 45 SMPL - - - -

2. For status information on a particular pair of disks, enter the following:

pairdisplay -g groupname -d dev_name [-CLI] [-fc]
where dev_name is the name specified in the configuration files under dev_name.

Note If no primary-secondary associations (pairings) exist, all disks are listed as SMPL in
the P/S column. To create a primary-secondary pairing, see If disks are not
paired: on page 166.

The following describes important headers in the pairdisplay listing.

Group

This is the dev_group name defined in the configuration file.

PairVol

Lists the devices by device name. In the above output, dev1 is listed twice: the first line is
the primary disk, the second is the mirror (secondary). This is shown under the P/S
column: P-VOL indicates the primary, S-VOL the secondary.
L/R
Indicates local or remote host, with respect to the current instance number.
Port#

The port number for the disk, configured by means of the arrays dedicated console (not
from a NetBackup host).
TID
The SCSI or fibre channel target ID number of the disk, configured by means of the arrays
dedicated console (not from a NetBackup host).
LU-M

164 NetBackup Advanced Client System Administrators Guide

 Configuring Primary and Secondary Disks

LU indicates the SCSI or fibre channel logical unit number of the disk, configured by
means of the arrays dedicated console (not from a NetBackup host). M is the numeric
mirror descriptor for cascading disks. A cascading disk has more than one mirror
(secondary) associated with a given primary.
Seq#
This is the unit serial number of the array.
LDEV#
Logical device number of the disk.
P/S
Indicates whether or not the disk is configured in a primary-secondary pair:

P-VOL: the disk is the primary.

S-VOL: the disk is the secondary.

SMPL: the disk is not paired (associated) with any other disk.

Status

Shows the current state of each disk in the array:

PAIR: the secondary disk in the pair is synchronized with the primary.
PSUS: the pair is split (primary disk).
SSUS: the pair is split (secondary disk).
COPY: a synch or split is in progress. If synchronizing, the status changes to PAIR at
completion of the COPY; if splitting, the result is PSUS for primary disk, or SSUS for
secondary disk.

Note If a backup is attempted while a disk is split (PSUS, SSUS), the backup fails with a
status code 11. If a backup is attempted while a disk is in the COPY state, there are
two possible results: if the disks synchronize (shown as PAIR), the backup proceeds;
if the disks split (PSUS, SSUS), the backup fails with a status code 11.

%
Shows the percentage of the status that has completed.
P-LDEV#
The LDEV number of the partner disk in the pair.
M
Indicates whether the secondary is writable, as a result of being split from the primary.

Chapter 8, Snapshot Configuration Notes 165

Configuring Primary and Secondary Disks

Configure the arrays, depending on your requirements

The next steps depend on the results of the pairdisplay listings and the requirements
of your site.
If all required disks are correctly paired (status of PAIR), the primary-secondary
configuration is finished.
If required disks are paired but currently split (PSUS, SSUS), or if they are not paired
at all (SMPL), you must resynchronize or configure them, respectively.

Note If a mirror-type snapshot backup attempts to access a disk that is split or not paired,
the backup fails with a status code 11.

If disks are paired but need to be unpaired or otherwise reconfigured, you must split
them and create a new association.

If disks are split:

1. Enter the following to resynchronize the split disks:

pairresync -g groupname -d dev_name

where groupname is the name listed under dev_group, and dev_name is the device
name, as defined in the configuration files. To resynchronize the disks listed as split
(PSUS, SSUS) in the above example (see Resulting output: on page 164), enter:
pairresync -g wiltest -d dev2

2. Enter the following to view the result:

pairdisplay -g wiltest -d dev2 -CLI -fc
When the resynchronization starts, the Status column reads COPY. When it is nearly
completed, the Status column reads PAIR (see the % column for percentage
completion).

If disks are not paired:

1. Enter the following to create a pair of primary and secondary:

paircreate -g groupname -d dev_name -vl
where groupname is the name listed under dev_group, dev_name is the device
name, as defined in the configuration files, and -vl specifies that the current instance
number is the primary.
To associate the dev3 disks as a pair (the ones listed as SMPL in the above example;
see Resulting output: on page 164), enter the following:

166 NetBackup Advanced Client System Administrators Guide

 Configuring Primary and Secondary Disks

paircreate -g wiltest -d dev3 -vl

2. Enter the following to view the result:

pairdisplay -g wiltest -d dev3 -CLI -fc
When the synchronization starts, the Status column reads COPY. When it is nearly
completed, the Status column reads PAIR (see the % column for percentage
completion).

If disks are paired but need to be split or reconfigured:

1. To split the secondary disk from the primary but maintain the pair association, enter
the following:
pairsplit -g groupname -d dev_name
where groupname is the name listed under dev_group, and dev_name is the device
name, as defined in the configuration files. The pairdisplay command will show a
status of PSUS and SSUS.
For example:
pairsplit -g wiltest -d dev1

This splits the secondary from the primary in the dev1 pair.

2. To split the secondary disk from the primary and remove the pair association between
them, enter the following:
pairsplit -g groupname -d dev_name -S
where -S means break the pair association. The pairdisplay command will show
SMPL in the P/S column for the affected disks, meaning the disks are no longer
paired.
For more information on array configuration, refer to the documentation provided by
the arrays vendor.

Chapter 8, Snapshot Configuration Notes 167

Volume Manager Configuration

Volume Manager Configuration

Disk Label
On Solaris only: If client data is configured in Volume Manager volumes, be sure to label
all secondary devices using the format command (label option). Labeling the
secondary disks prevents Volume Manager from marking the disks as disabled (if they are
split from their primary disks) during a system reboot.
While a secondary disk is synchronized with its primary, the secondary is invisible to
Volume Manager. When the secondary is split off from its primary disk, the secondary
becomes visible again. If the secondaries are labeled (using the format label
command), Volume Manager will not disable the disks when they are split.

Disk Types
There are important restrictions involving the use of Volume Manager with Advanced
Client.

Note If these restrictions are not observed, the backup will fail.

Do not include secondary (mirror) disks in a Volume Manager disk group.

The Volume Manager disk group must contain disks of one vendor type only. Do not
configure disks of different vendors in the same Volume Manager disk group.
The vendor type of the snapshot method must match the vendor-type of the disks in
the Volume Manager disk group.
Concerning these restrictions, refer to the next two diagrams.

168 NetBackup Advanced Client System Administrators Guide

 Volume Manager Configuration

Example VxVM Disk Groups: the Good and the Bad

Disk arrays on SAN

Fibre Channel

EMC Symmetrix Hitachi HP

P1 M1 P3 M3 P5 M5

P2 M2 P4 M4 P6 M6

P = primary disk in array

M = mirror (secondary)
disk in array

Consider the following VxVM disk

groups:

If disk group contains P1, P2 Good: group contains

only primary devices, of
same vendor.

If disk group contains P3, M3 Bad: group contains a

secondary (mirror) disk.

If disk group contains P1, P3, P5 Bad: group contains disks

of different vendors.

As shown above, no secondary (mirror) disks should be included in VxVM disk groups,
and groups must contain disks of the same vendor.

Note These restrictions apply when using any of the array-specific snapshot methods;
they do NOT apply if you are using the vxvm snapshot method.

Chapter 8, Snapshot Configuration Notes 169

Volume Manager Configuration

When Using Volume Manager and Array-Specific Methods

For each of these Volume Manager volumes:

/dev/vx/rdsk/dg/emc-vol /dev/vx/rdsk/dg/hds-vol /dev/vx/rdsk/dg/hp-vol

EMC Hitachi HP
array array array

select this snapshot method:

TimeFinder ShadowImage BusinessCopy

Disk Group Clones

When using the array-specific snapshot methods with client data configured over a
Volume Manager volume, NetBackup creates a temporary disk group (clone) of the disks
containing the mirror volume. To avoid a naming conflict in the Volume Manager,
NetBackup names the temporary disk group as follows:
diskgroup_name_clone

For alternate client backup, the temporary disk group is named as follows:
client_name_diskgroup_name_clone

While the backup is in progress, this clone appears in the output of the Volume Manager
vxdg command. This is normal. When the backup completes, NetBackup automatically
removes the disk group clone.

170 NetBackup Advanced Client System Administrators Guide

 Volume Manager Configuration

Disk Group Cloning Example:

Client data is in:

- file system /fs_1
- configured over VxVM volume /dev/vx/rdsk/dg_1/vol01

vol01 on vol01 on
primary_disk1 mirror_disk1

VxVM disk group NetBackup creates a temporary VxVM disk

dg_1 on group (clone) client_name_dg_1_clone on
primary_disk1. mirror_disk1.

In the above example, NetBackup removes the VxVM disk group

client_name_dg_1_clone after the backup has completed. If a major system
interruption occurs (such as a crash or unexpected reboot), NetBackup may not be able to
remove the clone. In that case, you must use the bpdgclone command with the -c
option to remove the clone, and then resynchronize the mirror disk with the primary disk.
Refer to Removing a VxVM Volume Clone on page 211 for assistance.

When Secondary Disks are Split and Synched

This situation is described for your information only. A backup occurring in this
circumstance should complete normally, in spite of the Volume Manager error described
as follows.
When the secondary (mirror) device is split from its primary, Volume Manager will see the
secondary disk as a separate device. But when the secondary disk is then re-synched to its
primary disk (provided Volume Manager had seen it before), the synched secondary disk
is no longer visible and VxVM issues an I/O error. In addition, if DMP is enabled, the
secondary disks are marked as disabled. The next time the secondary is split, it will
reappear in Volume Manager, only to disappear when the disk is again synched to its
primary.

Chapter 8, Snapshot Configuration Notes 171

Best Practices

Best Practices
The recommendations in this section apply primarily to the use of the array-specific
snapshot methods and Volume Manager, except where noted.

NetBackup Access to Arrays

In connection with the information listed under Access to Disk Arrays on page 154, note
the following recommendation:
The NetBackup media server only needs read access to the secondary disks in the
array; it does not need access to the primary disks.

Resynchronizing Disks At End of Backup

Resynchronizing very large mirror disks can take time. If disk-resynchronization
significantly delays completion of the backup, set the Resynchronize mirror in
background option to Yes. This allows the backup to complete without waiting for the
mirror disks to be resynchronized. The disks are resynchronized after the backup
completes. Refer to step 6 on page 92 for more information on this option.

Hardware-Level Disk Restore

Caution Hardware-level disk restore (such as by means of the symmir command with
the -restore option) can result in data loss if the primary disk is shared by
more than one file system or more than one VxVM volume. The hardware-level
restore overwrites the entire primary disk with the contents of the mirror disk.

This can be a problem if you are attempting to restore a snapshot of one of the
file systems or one of the VxVM volumes that share the same disk: the other file
systems or volumes sharing the disk may have older data that you do not want
to write back to the primary. When the hardware-level disk restore takes place,
the older data will replace the newer data on the primary disk.

Volume Manager Disk Groups

When creating a VxVM disk group, it is best to create a group that corresponds to the
primary disks that were grouped as described under Configuring Primary and
Secondary Disks on page 156. If you create an array disk group with two primary disks,
a VxVM disk group should be created with the same primaries. The VxVM disk group
configuration should follow the array disk group configuration for the primaries.

172 NetBackup Advanced Client System Administrators Guide

 Best Practices

Volume Manager with Dynamic Multipathing (DMP)

If you are using Volume Manager with DMP enabled, and there are multiple paths to the
same disk array (for instance, one fibre channel connection and one SCSI), DMP will
rename the arrays disks with DMP encapsulated names.

Backups Concurrently Accessing Same Disk (no VxVM)

A conflict occurs if two or more backups using array-specific snapshot methods attempt
to access a disk at the same time.
When the snapshot process is started, NetBackup reserves or locks the secondary
(mirror) disk for that backup job, denying any other backup jobs access to that disk. If a
second backup job requests a snapshot involving data on the same disk before the first job
is complete, access is denied and the second job fails.
This conflict can arise when there are two backup policies, each using an array-specific
method, and each requires access to the same disk at the same time (see diagram).

Backup Policies in Conflict: Two Backups Accessing Same Disk

secondary disk (mirror)

Backup policy A
partition 0,
Backup Selections list:
/file_sys1 Backup policy B
/file_sys1
Backup Selections list:
partition 1,
/file_sys2 /file_sys2

Backup policy A starts job to back up /file_sys1;

mirror is locked to make snapshot. Backup policy B
attempts to back up /file_sys2 and requests
snapshot: mirror is already lockedaccess denied.

Note Snapshot disk locks are applied to the entire disk: when a backup job requires a
snapshot, the entire disk is locked.

To avoid this conflict, see Avoiding Concurrent Access Conflicts later in this chapter.

Chapter 8, Snapshot Configuration Notes 173

Best Practices

Backups Concurrently Accessing VxVM Volumes

A conflict occurs if two or more concurrent backups using an array-specific method
attempt to access data configured in the same Volume Manager volume or in volumes
configured on the same disk(s).

Concurrent Access to Same VxVM Volume

In this case, a conflict occurs if two or more backups using an array-specific method
attempt to access the same Volume Manager volume at the same time.

Backup Policies in Conflict: Two Backups Accessing Same Volume

secondary disk (mirror)

Backup policy A Backup policy B
Backup Selections list: Backup Selections list:
/dev/vx/rdsk/dg/vol_1 /dev/rdsk/cxtxdxs2 /dev/vx/rdsk/dg/vol_1

Backup policy A starts job to back up /vol_1; mirror is locked to make

snapshot. Backup policy B attempts to back up /vol_1 and requests
snapshot: mirror is already lockedaccess denied.

The above diagram shows /dev/vx/rdsk/dg/vol_1 on a single disk. The same conflict
will occur if /vol_1 is distributed across two or more disks.
To avoid this conflict, see Avoiding Concurrent Access Conflicts later in this chapter.

Concurrent Access to Volume Manager Volumes on Same Disks

A conflict can occur if two or more concurrent backups using an array-specific method
attempt to access Volume Manager volumes that are distributed across the same disks.
VERITAS Volume Manager (VxVM) supports three means of distributing volumes across
disks: striping, concatenating, and RAID 5 (described in the Volume Manager
documentation). Use of these distribution methods can lead to access problems for
NetBackup with an array-specific method. The following diagram shows two VxVM
volumes, /dev/vx/rdsk/dg/vol_1 and /dev/vx/rdsk/dg/vol_2. Each is
distributed across two disks in an array (using any of the three distribution methods). If
two backups request snapshots of these volumes at the same time, a conflict occurs, even

174 NetBackup Advanced Client System Administrators Guide

 Best Practices

though the two backups are attempting to access different volumes. This happens because
the array-specific snapshot methods split the mirror disk from the primary disk at the disk
device layer, not at the volume layer.

Backup Policies in Conflict: Two Backups Accessing Volumes Distributed on Same Disks

secondary disks (mirrors)

A
Backup policy A
Backup Selections list:
/dev/rdsk/cxt0dxs2 /dev/vx/rdsk/dg/vol_1

Backup policy B
/dev/rdsk/cxt1dxs2
Backup Selections list:
/dev/vx/rdsk/dg/vol_2

Backup policy A starts to back up /vol_1; both disks A and B are locked to
make a snapshot of /vol_1. Backup policy B attempts to back up /vol_2 and
requests snapshot: disks A and B are already lockedaccess denied.

Avoiding Concurrent Access Conflicts

These are recommendations for backups that encounter any of the concurrent-access
problems when using an array-specific method.
Schedule the policies so that none can start a backup at the same time as another.
If possible, combine the separate policies into one policy. Snapshots will be created
before backups begin, on a global basis across all streams at once.
If you want the backups to run concurrently, combine the separate policies into one
and configure that policy for multiple data streaming. Multiple data streaming
prevents concurrent backups from encountering snapshot conflicts. See the NetBackup
System Administrators Guide for help with multiple data streams.

If the data to back up is configured in Volume Manager (VxVM) volumes, use the
vxvm snapshot method. The vxvm method allows snapshot backups to run
concurrently without conflicts, provided that the backup data consists of file systems
mounted on VxVM volumes. See Creating a Snapshot Mirror of the Source on
page 141 for help with the vxvm method.
Use the Volume Manager administration interface to determine which disks the
volumes are configured on, and configure the volumes on different disks.

Chapter 8, Snapshot Configuration Notes 175

Best Practices

176 NetBackup Advanced Client System Administrators Guide

Notes on Media Server and Third-Party
Copy 9
This chapter provides notes and restrictions regarding the NetBackup Media Server and
Third-Party Copy Device methods.
The following topics are covered in this chapter:
Disk Requirements for Media Server/ Third-Party Copy
ALL_LOCAL_DRIVES
Storage Units
Multiplexing
Raw Partition Backups

177
Disk Requirements for Media Server/ Third-Party Copy

Disk Requirements for Media Server/ Third-Party Copy

For the NetBackup Media Server or Third-Party Copy Device backup method, the clients
data must be on one or more disks that meet the following criteria:
The disk must be either a SCSI or Fibre Channel device.
The disk must be visible to both the NetBackup client and to the NetBackup media
server. The disk must be connected through a fibre channel SAN or through a disk
array that has dual port SCSI connections.
The disk must be able to return its SCSI serial number in response to a serial-number
inquiry (serialization), or the disk must support SCSI Inquiry Page Code 83.

ALL_LOCAL_DRIVES
The policys Backup Selections list must not contain the ALL_LOCAL_DRIVES entry.

Storage Units
Any_available is not supported for NetBackup Media Server and Third-Party Copy
Device backup methods.
Disk storage units are not supported for the Third-Party Copy Device method.

Multiplexing
The Third-Party Copy Device backup method is incompatible with multiplexing (the
writing of two or more concurrent backup jobs to the same storage device). To prevent
multiplexing on a third-party copy backup, you must set Maximum multiplexing per
drive to 1 (on the Add New Storage Unit or Change Storage Unit dialog).

Raw Partition Backups

When entering a raw partition in the Backup Selections list (for a policy that uses the
NetBackup Media Server or Third-Party Copy method), do not specify a block device as the
raw partition. For these two backup methods, NetBackup does not support block devices.
Instead, specify the raw partition as a character device.
Examples:
Solaris:

178 NetBackup Advanced Client System Administrators Guide

 Raw Partition Backups

/dev/rdsk/c1t3d0s3
HP:
/dev/rdsk/c1t0d0

Chapter 9, Notes on Media Server and Third-Party Copy 179

Raw Partition Backups

180 NetBackup Advanced Client System Administrators Guide

Backup and Restore Procedures 10

The following topics are covered in this chapter:

Performing a Backup

Performing a Restore

Configurations for Restore

Restoring from a Disk Snapshot

181
Performing a Backup

Performing a Backup
Before proceeding, please note the following for the array integration snapshot methods.
For the EMC TimeFinder, Hitachi ShadowImage, or HP BusinessCopy snapshot method,
the client data to be backed up must reside on a mirror disk made by the corresponding
vendor (EMC, Hitachi, or HP). Assistance from the disk array vendors technical support
may be required. For NetBackup-related items, refer to the chapter titled Snapshot
Configuration Notes.

Automatic Backup
The most convenient way to back up client data is to configure a policy and then set up
schedules for automatic, unattended backups. To use NetBackup Advanced Client, you
must enable snapshot backup as described in the appropriate configuration chapter of this
guide. To add new schedules or change existing schedules for automatic backups, you can
follow the guidelines in the NetBackup System Administrators Guide.

Manual Backup
The administrator can use the NetBackup Administration interface on the master server to
execute a backup for a policy. To use NetBackup Advanced Client, you must enable
snapshot backup as described in the appropriate configuration chapter of this guide.
See the NetBackup System Administrators Guide for instructions on doing manual backups.

User-Directed Backup and Archive

From a NetBackup client, the user can execute an Advanced Client backup. The
NetBackup administrator must configure an appropriate snapshot policy with schedule.
See the NetBackup Backup, Archive, and Restore Getting Started Guide for instructions on
doing user-directed backups and archives.

182 NetBackup Advanced Client System Administrators Guide

 Performing a Restore

Performing a Restore
You can use the Backup, Archive, and Restore interface to restore individual files or
directories, or a volume or raw partition. See the NetBackup Backup, Archive, and Restore
Getting Started Guide for instructions on performing the restore. The following sections
include restore notes and procedures unique to certain components of Advanced Client.

Restores from a FlashBackup Backup

Using the Backup, Archive, and Restore interface, you can restore individual directories or
files (or an entire raw partition) from a FlashBackup backup. The procedure is the same as
that for restoring from a regular backup as described in the NetBackup Backup, Archive, and
Restore Getting Started Guide.

Note In the Backup, Archive, and Restore interface, set the policy type to FlashBackup
for UNIX clients and FlashBackup-Windows for Windows clients.

Note the following before starting the restore.

A backup made from a FlashBackup (or FlashBackup-Windows) policy supports both
individual file restore and raw partition restore; that is, you can do either type of
restore from the same backup. To restore individual files, select Normal Backups on
the Restore Files tab; to restore an entire raw partition, select Raw Partition Backups.
To restore a raw partition, you must have administrator capabilities on the NetBackup
server.
An entire raw partition can be restored from a full backup only. FlashBackup
incremental backups only support individual file restores.
Ensure that the device file for the raw partition exists prior to the restore.
The overwrite option is ignored during raw partition restores. The device file must
exist and the disk partition is overwritten whether this option is set or not.
If you are restoring a very large number of files and individual file restore would take
too long, you can do a raw-partition restore by redirecting it to another raw partition
of the same size and then copying individual files to the original file system.
For UNIX clients:
To restore an entire raw partition, ensure that the partition is not mounted and not
in use. (For this reason, you cannot perform a raw partition restore to the root
partition, or to the partition on which NetBackup is installed.) If the partition is
being used by a database, shut down the database. The partition must be the same
size as when it was backed up; otherwise, the results of the restore are
unpredictable.

Chapter 10, Backup and Restore Procedures 183

Performing a Restore

After a raw partition restore of a VxFS file system, a file system consistency check
(fsck) is usually required before the file system can be mounted.
For Windows clients
To restore an entire raw partition, ensure that the partition is mounted
(designated as a drive letter) but not in use. (For this reason, you cannot perform a
raw partition restore to the root partition, or to the partition on which NetBackup
is installed.) If the partition is being used by a database, shut down the database.
The partition must be the same size as when it was backed up; otherwise, the
results of the restore are unpredictable.

Restores in Clustered File System (VxFS on UNIX Only)

In a clustered file system, if you are restoring a large number of files (such as 100,000 or
more), the restore will finish sooner if the file system is mounted on a local host. The file
system can then be mounted as a shared mount after the restore.
Do the following:

1. Stop all applications (on any nodes) that are using the file system.

2. Unmount the file system.

3. Mount the file system locally.

4. Perform the restore.

5. Share the mounted file system again, and restart applications (if any).

Instant Recovery Restore

You can restore files from an Instant Recovery backup in the same way as from a normal
backup. Restore procedures are described in the NetBackup Backup, Archive, and Restore
Getting Started Guide.
There are three restore features unique to Instant Recovery.
Block-level restore (for VxFS_Checkpoint snapshots)
File promotion (for VxFS_Checkpoint or NAS_Snapshot snapshots)
Rollback (for VxFS_Checkpoint, vxvm, or NAS_Snapshot snapshots)
These require special instructions, as follows.

184 NetBackup Advanced Client System Administrators Guide

 Performing a Restore

Instant Recovery: Block-Level Restore (UNIX Clients Only)

If the Instant Recovery snapshot was made with the VxFS_Checkpoint method, large files
can be recovered faster by means of block-level restore. Only the blocks that have changed
are moved from the snapshot to the clients primary fileset.

Important Notes
Block-level restore requires the VxFS File System.
Block-level restore is available only when restoring files to the original location on the
client, AND when the snapshot method used for the Instant Recovery backup was
VxFS_Checkpoint.
If the snapshot method for the backup was VxFS_Checkpoint and the files to be
restored are in an Oracle database, block-level restore is automatically enabled.

To activate block-level restore

Create the following (empty) file on the client:
/usr/openv/netbackup/PFI_BLI_RESTORE

After this file is created, all subsequent restores of the clients data will use block-level
restore.

To de-activate block-level restore

Delete (or rename) the PFI_BLI_RESTORE file.

Note When block-level restore is activated, it is used for all files in the restore. This may
not be appropriate for all of the files. It may take longer to restore a large number of
small files, because they must first be mapped.

Instant Recovery: File Promotion (UNIX Clients Only)

If the Instant Recovery snapshot was made with the VxFS_Checkpoint or NAS_Snapshot

method for a UNIX client, large files that have had many changes since the backup can be

recovered more quickly by means of file promotion. File promotion optimizes single-file

restore by using a minimum of I/O to recover the files.

Notes on File Promotion

Only regular files can be promoted, not file links or directories.

Notes on VxFS_Checkpoint:

Chapter 10, Backup and Restore Procedures 185

Performing a Restore

File promotion requires the VxFS File System version 4.0 or later.
File promotion can be done only from the last Instant Recovery snapshot that was
made with the VxFS_Checkpoint method.
File promotion is available only when restoring files to the original location on the
original client.
Notes on NAS_Snapshot:
File promotion is available when restoring to the original volume on the original
client.
File promotion can be done from older snapshots, but any newer NAS snapshots
are deleted after the file promotion takes place.
The file system requirements depend on the NAS vendor. For further
requirements specific to your NAS vendor, see the NetBackup for NDMP Supported
OS and NAS Appliance Information online document (refer to the preface for help
accessing that document).

To Use File Promotion

The procedure for restoring individual files with file promotion is the same as the
standard restore procedure for NetBackup (refer to the NetBackup Backup, Archive, and
Restore Getting Started Guide). No special settings or choices are required when using the
Backup, Archive, and Restore interface.

How File Promotion Is Selected By NetBackup

If the above requirements are met (see Notes on File Promotion), NetBackup
automatically attempts file promotion first. If the file cannot be promoted, thenfor
VxFS_Checkpoint onlyblock-level restore is used if it is activated as described under
Instant Recovery: Block-Level Restore (UNIX Clients Only) on page 185. Otherwise, the
restore takes place in the standard manner, without file promotion: all file data is copied
from the snapshot to the primary file system. The NetBackup progress log indicates how
many files were promoted, and how many files that could not be promoted were restored
by means of tar.

Instant Recovery: Snapshot Rollback

You can also restore a snapshot of an entire file system or volume with minimal I/O. This
type of restore is called point in time rollback. All the data in the snapshot is restored; single
file restore is not available in a rollback.

186 NetBackup Advanced Client System Administrators Guide

 Performing a Restore

Notes on Rollback
Rollback can be done only from backups that were enabled for Instant Recovery and
made with the VxFS_Checkpoint, vxvm, or NAS_Snapshot methods.
For backups made with the VxFS_Checkpoint method, rollback requires the VxFS File
System 4.0 or later and Disk Layout 6. For NAS_Snapshot, the file system
requirements depend on the NAS vendor.
Rollback deletes any snapshots (and their catalog information) that were created after
the creation-date of the snapshot that you are restoring.
Rollback deletes all files that were created after the creation-date of the snapshot that
you are restoring. Rollback returns a volume to a given point in time. Any data
changes or snapshots that were made after that time are lost.
Rollback is available only when restoring the file system or volume to the original
location on the client.
When a rollback of a file system is initiated, NetBackup verifies, by default, that the
primary file system does not contain files created after the snapshot was made;
otherwise, the rollback aborts.

To perform snapshot rollback (UNIX)

This procedure requires root access.

1. Start the Backup, Archive, and Restore interface.

/usr/openv/netbackup/bin/jbpSA &

2. Click the Restore Files tab.

3. Click Actions > Specify NetBackup Machines to specify the server, source client,
policy type, and destination client.

4. For the Restore Type, select Point in Time Rollback.

The Browse directory field is grayed out, with root (/) as default.

Chapter 10, Backup and Restore Procedures 187

Performing a Restore

Instant Recovery backups are displayed in the Backup History window, for all dates
(you cannot set a range).

5. Select an image from the list and click OK.

The image contents are displayed in the Directory Structure pane of the Restore Files
tab.

You can select root

level or mount points
(file systems or
volumes), but not
folders or files at a
lower level

188 NetBackup Advanced Client System Administrators Guide

 Performing a Restore

6. In the Directory Structure list, click the check box next to the root node or a mount
point beneath root.
You can select a file system or volume, but not lower-level components.

7. Click the Restore button.

Select this
option only if you
are sure you
want to replace
all files in the
original location.

The only available destination option is Restore everything to its original location.

8. For file systems, you can choose to skip file verification by placing a check in the
Overwrite existing files option.

Caution Click on Overwrite existing files only if you are sure you want to replace all the
files in the original location with the snapshot. Rollback deletes all files that
were created after the creation-date of the snapshot that you are restoring.

If Overwrite existing files is not selected, NetBackup performs several checks on the
file system as described under Notes on Rollback on page 187. If the checks do not
pass, the rollback aborts and a message is written to the Task Progress tab stating that
rollback could not be performed because file verification failed.
The rest of the procedure is identical to a normal restore as explained in the NetBackup
Backup, Archive, and Restore Getting Started Guide.

Chapter 10, Backup and Restore Procedures 189

Performing a Restore

To perform snapshot rollback (Windows)

This procedure requires Administrator privilege.

1. Start the Backup, Archive, and Restore interface.

Click Start > Programs > VERITAS NetBackup > Backup, Archive, and Restore.

2. From the Select for Restore drop-down list, select Restore from Point in Time
Rollback.

3. Click File > Specify NetBackup Machines and Policy Type to specify the server,
source client, policy type, and destination client.

4. In the NetBackup History pane, click on the backup image to restore.

Only Instant Recovery backups are displayed in the NetBackup History pane, for all
dates (you cannot set a date range).

You can select root

level or mount points
(file systems or
volumes), but not
folders or files at a
lower level

5. In the All Folders pane, click the check box next to the root node or a mount point
beneath root.
You can select a file system or volume, but not lower-level components.

6. Click Actions > Start Restore of Marked Files.

190 NetBackup Advanced Client System Administrators Guide

 Performing a Restore

Select this
option only if you
are sure you
want to replace
all files in the
original location.

Note The only destination option is Restore everything to its original location.

7. For file systems, you can choose to skip file verification by placing a check in the
Overwrite the existing file option.

Caution Click on Overwrite the existing file only if you are sure you want to replace all
the files in the original location with the snapshot. Rollback deletes all files that
were created after the creation-date of the snapshot that you are restoring.

If Overwrite the existing file is not selected, NetBackup performs several checks on
the file system as described under Notes on Rollback on page 187. If the checks do
not pass, the rollback aborts and a message is written to the progress log stating that
rollback could not be performed because file verification failed.
The remainder of the procedure is identical to a normal restore as explained in the
NetBackup Backup, Archive, and Restore Getting Started Guide.

Chapter 10, Backup and Restore Procedures 191

Configurations for Restore

Configurations for Restore

Advanced Client backups can be restored in several ways, depending on your
configuration:
Restore over the LAN: data can be restored from the storage device to the media
server, and from the media server via the LAN to the client. This kind of restore is also
used for ordinary (non-Advanced Client) backups.

LAN / WAN

2 media
client
SCSI server
client
disks SCSI
3
1
storage
1. Media server reads data from local storage.
2. Media server sends the data to the client over the LAN.
3. Client restores the data to disk (disk can be locally
attached or on SAN).

Restore over the SAN to a host acting as both client and media server. This requires
the FORCE_RESTORE_MEDIA_SERVER option in the servers bp.conf file (see the
NetBackup System Administrators Guide for details on this option).

LAN / WAN

client/media media
server server

SAN
2

client disks 1 storage

on SAN

1. Client/media server reads data from tape over the SAN.

2. Client restores the data to disk (disk can be locally attached or on SAN)
(Requires use of FORCE_RESTORE_MEDIA_SERVER option in bp.conf file.)

192 NetBackup Advanced Client System Administrators Guide

 Restoring from a Disk Snapshot

Restore directly from a snapshot (this is not the Instant Recovery feature): if the Keep
snapshot after backup option was turned on for the backup, the data can be restored
from a mirror disk by restoring individual files from the snapshot, or by restoring the
entire snapshot. Note that this type of restore must be done from the command
prompt (using, for instance, a copy command such as UNIX cp), not from the
NetBackup Administration Console.
For details, refer to Restoring from a Disk Snapshot on page 193.

Restoring from a Disk Snapshot

If the Keep snapshot after backup parameter is set to Yes (on the Snapshot Options
dialog), the snapshot is retained on the mirror disk after the backup completes. From the
command line, you can restore individual files or the entire snapshot directly from the
disk, rather than restoring from tape.

Note Unless the backup was made with the Instant Recovery feature, you cannot restore
from a snapshot by means of the Backup, Archive, and Restore interface; you must
perform the restore manually at the command line.

On UNIX

To restore individual files

To restore individual files, locate and mount the snapshot file system and copy files from
that file system using system commands, such as cp and ftp, as follows.

1. To list the identifiers of current snapshots, use the bpfis command with the query
option:
/usr/openv/netbackup/bin/bpfis query

This returns the ID (FIS IDs) of all current snapshots. For example:
INF - BACKUP START 3629

INF - FIS IDs: 1036458302

INF - EXIT STATUS 0: the requested operation was successfully

completed

2. For each snapshot identifier, enter bpfis query again, specifying the snapshot ID:
/usr/openv/netbackup/bin/bpfis query -id 1036458302
This returns the path of the original file system (snapshot source) and the path of the
snapshot file system. For example:

Chapter 10, Backup and Restore Procedures 193

Restoring from a Disk Snapshot

INF - BACKUP START 3629

INF - Snapshot image host : ricopico

INF - Snapshot image owner: GENERIC

INF - Time created : Mon Oct 7 19:35:38 2002

INF - REMAP FILE BACKUP /mnt/ufscon USING /tmp/_vrts_frzn_img_26808/mnt/ufscon

OPTIONS:ALT_PATH_PREFIX=/tmp/_vrts_frzn_img_26808,FITYPE=MIRROR,

MNTPOINT=/mnt/ufscon,FSTYPE=ufs

INF - EXIT STATUS 0: the requested operation was successfully completed

In this example, the primary file system is /mnt/ufscon and the snapshot file
system is /tmp/_vrts_frzn_img_26808/mnt/ufscon.
For further examples using bpfis, see the Managing Snapshots from the Command
Line appendix in this manual.

3. Copy the files from the mounted snapshot file system to the original file system.

To restore the entire snapshot

There are additional ways to recover data from the disk snapshot, depending on your
hardware configuration and the snapshot method used by the policy.

If the snapshot method was FlashSnap, you can restore the snapshot volume as follows:

1. Unmount the snapshot source (original file system) and the snapshot file system on
the alternate client:
umount original_file_system

umount snapshot_image_file_system

To locate the file systems, refer to step 1 and step 2 on page 193.

2. Deport the snapshot on the alternate-client:

a. Find the VxVM disk group:

vxdg list
The format of the disk group name is as follows:
SPLIT-primaryhost_diskgroup

If vxdg list does not show the disk group, the group might have been
deported. You can discover all the disk groups, including deported ones, by
entering:
vxdisk -o alldgs list

The disk groups listed in parentheses are not imported on the local system.

b. Deport the VxVM disk group:

194 NetBackup Advanced Client System Administrators Guide

 Restoring from a Disk Snapshot

vxdg deport SPLIT-primaryhost_diskgroup

3. Import and join the VxVM disk group on the primary (original) client:
vxdg import SPLIT-primaryhost_diskgroup

vxrecover -g SPLIT-primaryhost_diskgroup -m

vxdg join SPLIT-primaryhost_diskgroup diskgroup

4. Start the volume and snap back the snapshot volume, using the
-o resyncfromreplica option:
vxvol -g SPLIT-primaryhost_diskgroup start SNAP_diskgroup_volume

vxassist -g SPLIT-primaryhost_diskgroup -o resyncfromreplica snapback SNAP_diskgroup_volume

If the snapshot was made on an EMC, Hitachi, or HP disk array:

WITH CAUTION, you can use hardware-level restore to restore the entire mirror or
secondary disk to the primary disk. If the disk is shared by more than one file system or
VxVM volume, there may be unintended results. Please read the following:

Caution Hardware-level disk restore (such as by means of the symmir command with
the -restore option) can result in data loss if the primary disk is shared by
more than one file system or more than one VxVM volume. The hardware-level
restore overwrites the entire primary disk with the contents of the mirror disk.

This can be a problem if you are attempting to restore a snapshot of one of the
file systems or one of the VxVM volumes that share the same disk: the other file
systems or volumes sharing the disk may have older data that you do not want
to write back to the primary. When the hardware-level disk restore takes place,
the older data will replace the newer data on the primary disk.

On Windows

To restore individual files

1. To list the identifiers of current snapshots, use the bpfis command with the query
option:
/usr/openv/netbackup/bin/bpfis query

This returns the ID (FIS IDs) of all current snapshots. For example:
INF - BACKUP START 3629

INF - FIS IDs: 1036458302

INF - EXIT STATUS 0: the requested operation was successfully

completed

Chapter 10, Backup and Restore Procedures 195

Restoring from a Disk Snapshot

2. For each snapshot identifier, enter bpfis query again, specifying the snapshot ID:
/usr/openv/netbackup/bin/bpfis query -id 1036458302
This returns the path or the original file system (snapshot source) and the GUID
(Global Universal Identifier) representing the snapshot volume. For example:
INF - BACKUP START 2708

INF - Snapshot method: FlashSnap

INF - Snapshot image host : tire

INF - Snapshot image owner : NBU

INF - Time created : Sat Oct 25 15:26:04 2003

INF - REMAP FILE BACKUP H:\ USING

\\?\Volume{54aa666f-0547-11d8-b023-00065bde58d1}\

OPTIONS:ALT_PATH_PREFIX=C:\Program Files\VERITAS\NetBackup\

Temp\_vrts_frzn_img_2408,FITYPE=MIRROR,MNTPOINT=H:\,FSTYPE=NTFS

INF - EXIT STATUS 0: the requested operation was successfully completed

In this example the snapshot file system is H:\ and the GUID is
\\?\Volume{54aa666f-0547-11d8-b023-00065bde58d1}\.

3. To restore individual files from the snapshot volume:

a. Mount the GUID to an empty NTFS directory:

mountvol C:\Temp\Mount

\\?\Volume{54aa666f-0547-11d8-b023-00065bde58d1}\

b. Copy the file to be restored from the temporary snapshot mountpoint (in this
example, C:\Temp\Mount)to the primary volume.

To restore the entire snapshot

There are additional ways to recover data from the disk snapshot, depending on your
hardware configuration and the snapshot method used by the policy.

If the snapshot method was FlashSnap, you can restore the snapshot volume as follows:

1. Deport the snapshot on the alternate-client:

a. Find the VxVM disk group:

vxdg list
The format of the disk group name is as follows:
SPLIT-primaryhost_diskgroup

b. Deport the VxVM disk group:

196 NetBackup Advanced Client System Administrators Guide

 Restoring from a Disk Snapshot

vxdg -g split_diskgroup deport

2. Import and join the VxVM disk group on the primary (original) client:
vxassist rescan

vxdg -g split_diskgroup import

vxdg -g split_diskgroup -n diskgroup join

3. Snap back the snapshot volume, using the -o resyncfromreplica option:

vxassist -o resyncfromreplica snapback \Device\HarddiskDmVolumes\diskgroup\snap_volume

Chapter 10, Backup and Restore Procedures 197

Restoring from a Disk Snapshot

198 NetBackup Advanced Client System Administrators Guide

Troubleshooting 11

This chapter covers the following topics.

Gathering Information and Checking Logs

Important Notes

Particular Issues

Removing a Snapshot

Removing a VxVM Volume Clone

Note For explanations of NetBackup status codes, refer to the NetBackup Status Codes
and Messages chapter in the NetBackup Troubleshooting Guide.

199
Gathering Information and Checking Logs

Gathering Information and Checking Logs

You can resolve many problems on your own by creating the logging directories listed
below, reproducing the problem, and then checking the logs. For an in-depth description
of NetBackup logs, refer to the NetBackup Troubleshooting Guide.

Logging Directories for UNIX Platforms

Advanced Client messages are written to the following directories if they exist.

Note To create detailed log information, place a VERBOSE entry in the bp.conf file on
the NetBackup master and client, or set the Global logging level to a high value in
the Logging dialog, under both Master Server Properties and Client Properties.

Note These directories can eventually require a lot of disk space. Delete them when you
are finished troubleshooting and remove the VERBOSE option from the bp.conf
file or reset the Global logging level to a lower value.

Logging Directories for Backup

During a backup, Advanced Client messages are logged to the following directories.
Create these directories using an access mode of 755 so NetBackup can write to the logs.

Path of log directory Where to create the directory

/usr/openv/netbackup/logs/bpbrm NetBackup server

/usr/openv/netbackup/logs/bptm NetBackup media server

/usr/openv/netbackup/logs/online_util NetBackup client or alternate

client

/usr/openv/netbackup/logs/bpbkar NetBackup client or alternate

client

/usr/openv/netbackup/logs/bpfis NetBackup client or alternate

client

/usr/openv/netbackup/logs/bppfi NetBackup client or alternate

client

200 NetBackup Advanced Client System Administrators Guide

 Gathering Information and Checking Logs

Logging Directories for Restore

During a restore, Advanced Client messages are logged to the following directories on the
master server. Create these directories using an access mode of 755.

Path of log directory Where to create the directory

/usr/openv/netbackup/logs/bprestore NetBackup master server

/usr/openv/netbackup/logs/bprd NetBackup master server

/usr/openv/netbackup/logs/bpbrm NetBackup master server

/usr/openv/netbackup/logs/bptm NetBackup master server

snapctl Driver Messages

Messages from the snapctl driver are logged in the clients /var/adm/messages file
along with other kernel messages.

Logging Folders for Windows platforms

During a backup, Advanced Client messages are written to the folders listed below if they
exist. You can use the following command to create these folders:
install_path\NetBackup\logs\mklogdir.bat
The default path for the logs is C:\Program Files\VERITAS\NetBackup\logs.
Since a different path can be set during installation, the path listed below is shown as
install_path\NetBackup\logs.

Note To create detailed log information, set the Global logging level to a high value, in
the Logging dialog, under both Master Server Properties and Client Properties.

Note The log folders can eventually require a lot of disk space. Delete them when you are
finished troubleshooting and set the logging level on master and client to a lower
value.

Chapter 11, Troubleshooting 201

Gathering Information and Checking Logs

Logging Folders for Backup

During a backup, Advanced Client messages are logged to the following folders.

Path of log directory Where folder is created

install_path\NetBackup\logs\bpbrm NetBackup server

install_path\NetBackup\logs\bptm NetBackup media server

install_path\NetBackup\logs\bppfi NetBackup client or alternate

client

install_path\NetBackup\logs\online_util NetBackup client or alternate

client

install_path\NetBackup\logs\bpbkar NetBackup client or alternate

client

install_path\NetBackup\logs\bpfis NetBackup client or alternate

client

Logging Folders for Restore

During a restore, Advanced Client messages are logged to the following folders on the
master server.

Path of log directory Where folder is created

install_path\NetBackup\logs\bprestore NetBackup master server

install_path\NetBackup\logs\bprd NetBackup master server

install_path\NetBackup\logs\bpbrm NetBackup master server

install_path\NetBackup\logs\bptm NetBackup master server

install_path\NetBackup\logs\tar NetBackup master server

202 NetBackup Advanced Client System Administrators Guide

 Important Notes

Contacting VERITAS Customer Support

Before calling customer support, please gather as much log information as possible. Be
sure to have the following information ready:
NetBackup version
Operating system version of the NetBackup master and media server and NetBackup
Advanced Client client
Note whether or not the action that failed had ever worked, and whether the problem
is repeatable
Log information

Important Notes
The disk containing the clients data (the files to back up) must be a SCSI or Fibre
Channel device if you are using NetBackup Media Server or Third-Party Copy
Device.
The disk containing the clients data must be visible to both the client and the media
server if you are using the NetBackup Media Server or Third-Party Copy Device
method. The disk can be connected through SCSI or fibre channel.
For the NetBackup Media Server or Third-Party Copy Device method, a disk device
must be able to return its SCSI serial number in response to a serial-number inquiry
(serialization), or the disk must support SCSI Inquiry Page Code 83.
When configuring the Third-Party Copy Device or NetBackup Media Server method,
a particular storage unit or group of storage units must be specified for the policydo
not choose Any_available. For configuration instructions, refer to Configuring an
Advanced Client Policy on page 85.
The storage_unit_name portion of a mover.conf.storage_unit_name file name
must exactly match the actual storage unit name (such as nut-4mm-robot-tl4-0)
that you have defined for the policy. See Naming the Mover File on page 73 for help
creating a mover.conf.storage_unit_name file.
Similarly, the policy_name portion of a mover.conf.policy_name file name must
match the actual name of the policy that the third-party copy device is to be associated
with.
For the TimeFinder, ShadowImage, or BusinessCopy snapshot methods, the client
data must reside in a device group, with the data on the primary disk and
synchronized with a mirror disk. Assistance from the disk array vendor may also be
required. Refer to Array-Specific Snapshot Methods on page 150.

Chapter 11, Troubleshooting 203

Important Notes

If the Keep snapshot after backup option for the snapshot method is changed from
yes to no, the last snapshot created for that policy must be deleted manually before
the backup is run again. Use the bpfis command to delete the snapshot. Refer to the
man page for bpfis, or to Managing Snapshots from the Command Line in this
guide.
During a third-party copy device backup, if tape performance is slow, increase the
buffer size by creating one of the following files on the media server:
/usr/openv/netbackup/db/config/SIZE_DATA_BUFFERS_TPC.policy_name

/usr/openv/netbackup/db/config/SIZE_DATA_BUFFERS_TPC.storage_unit_name

/usr/openv/netbackup/db/config/SIZE_DATA_BUFFERS_TPC

By default, the size of the data buffer for third-party copy backup is 65536 bytes (64K).
To increase it, put a larger integer in the SIZE_DATA_BUFFERS_TPC file. For a buffer
size of 96K, put 98304 in the file. If the value is not an exact multiple of 1024, the value
read from the file will be rounded up to a multiple of 1024.
The file name with no extension (SIZE_DATA_BUFFERS_TPC) applies as a default to
all third-party copy backups, if neither of the other file-name types exists. A
SIZE_DATA_BUFFERS_TPC file with the .policy_name extension applies to backups
executed by the named policy, and the .storage_unit_name extension applies to
backups using the named storage unit. If more than one of these files applies to a
given backup, the buffers value is selected in this order:
SIZE_DATA_BUFFERS_TPC.policy_name
SIZE_DATA_BUFFERS_TPC.storage_unit_name
SIZE_DATA_BUFFERS_TPC
As soon as one of the above files is located, its value is used. A .policy_name file that
matches the name of the executed policy will override the value in both the
.storage_unit_name file and the file with no extension. The .storage_unit_name file will
override the value in the file with no extension.
You can set the maximum buffer size that a particular third-party copy device can
support.

Note A third-party copy device will not be used if it cannot handle the buffer size set for
the backup.

204 NetBackup Advanced Client System Administrators Guide

 Particular Issues

Particular Issues

Installation Problem
If you receive the following message during installation:
/usr/openv/netbackup/bin/version not found.

Add-On Product Installation Aborted.

you have tried to install the Advanced Client software before installing the base
NetBackup software.

FlashBackup and Status code 13

Status 13 can result from any of the following:
The FlashBackup cache partition may have run out of space. In that case, the cache
partition may not be large enough for the backup requirements.
If the cache partition is full, you will see messages such as the following in the system
log:
WARNING: sn_alloccache: cache /dev/vx/rdsk/flash1dg/f full - all

snaps using this cache are now unusable

WARNING: sn_failsnap: snapshot id 5 failed error 28

Specify a larger cache partition, or designate additional cache partitions in the Backup
Selections list. See the FlashBackup Configuration chapter in this guide for cache
partition requirements.
On Solaris: If your cache partition runs out of space, there may be stale snapshots
taking up space on the cache partition. Stale snapshots are those that were not
automatically deleted by FlashBackup.

a. Determine if there are stale snapshots on your Solaris client by executing the
following:
/usr/openv/netbackup/bin/driver/snaplist

b. For each snapshot listed, execute the following to make sure there is a bpbkar
process associated with it:
ps -eaf |grep ident
where ident is the snapshot process id displayed by the snaplist command.

c. Remove snapshots that do not have an associated bpbkar process by entering the
following:

Chapter 11, Troubleshooting 205

Removing a Snapshot

/usr/openv/netbackup/bin/driver/snapoff snapn
where snapn is the snapshot id displayed by the snaplist command.

Removing a Snapshot
NetBackup ordinarily removes snapshots after the Advanced Client backup completes
(unless the Keep snapshot after backup parameter was set to Yes). However, as a result
of certain kinds of system failures, such as a system crash or abnormal termination of the
backup, the snapshot may not be removed.

To identify and remove a left-over snapshot

1. Use the bpfis command with the query option to list the current snapshots:
Do this on the client or alternate client, depending on the type of backup:
/usr/openv/netbackup/bin/bpfis query

This returns the IDs (FIS IDs) of all current snapshots. For example:
INF - BACKUP START 3629

INF - FIS IDs: 1036458302

INF - EXIT STATUS 0: the requested operation was successfully

completed

In this example, the snapshot ID is 1036458302.

2. If the bpfis output shows the ID of the snapshot, delete it as follows:

bpfis delete -id snapshot_id

If bpfis removed the snapshot, you can skip the rest of this procedure.

3. Solaris or HP only: if bpfis could not remove the snapshot, enter the following (on the
client or alternate client) when no backups are running:
df -k

This displays all mounted file systems, including any snapshots of a mounted file
system.

Note It is important to enter df -k when no backups are running. If a snapshot backup is

currently running, the snapshot should not be deleted. NetBackup will delete it
when the backup completes.

Here are two snapshots from a df -k listing:

/dev/dsk/c1t3d2s4 1048800 73076 914742 8% /tmp/_vrts_frzn_img__wil_vxfs_1299000

206 NetBackup Advanced Client System Administrators Guide

 Removing a Snapshot

/dev/vx/dsk/clone_qes_clone/ufs 38383 21678 12867 63% /tmp/_vrts_frzn_img__mix_ufs_1299000

The snapshot appears in the following form:

/tmp/_vrts_frzn_img__filesystemname_pid

4. Solaris or HP only: unmount the unneeded snapshot file systems (on the client or
alternate client, depending on the type of backup).

5. The next step depends on the type of snapshot.

For nbu_snap (Solaris only):

a. Enter the following to display leftover snaps:

/usr/openv/netbackup/bin/driver/snaplist

b. To remove a leftover snap, enter

/usr/openv/netbackup/bin/driver/snapoff snap1 ... snapn
For more information on snaplist and snapoff, refer to the Managing
nbu_snap (Solaris) appendix of this guide.

For vxvm and VVR: (Solaris and HP)

Do the following on the client for vxvm, and on the alternate client for VVR:

a. Enter the following to display unsynchronized mirror disks:

vxprint -g diskgroup

b. Enter the following to resynchronize the mirror disks:

vxassist -g diskgroup -v volume snapback

For vxvm: (Windows)

a. Enter the following to display unsynchronized mirror disks:

vxdg -g diskgroup dginfo

b. Enter the following to resynchronize the mirror disks:

vxassist snapback

\Device\HarddiskDmVolumes\diskgroup\snap_volume

Chapter 11, Troubleshooting 207

Removing a Snapshot

For VxFS_Checkpoint (Solaris or HP only):

a. Enter the following VxFS command to display the name of the checkpoint:
/usr/lib/fs/vxfs/fsckptadm list /file_system
where file_system is the mount point of the primary or original file system that
was backed up, NOT the snapshot file system that was unmounted in step 4.
For example, if the snapshot file system that was unmounted is the following:
/tmp/_vrts_frzn_img__vm2_1765

The original file system, which should be specified on the fsckptadm list
command, is this:
/vm2

Example entry:
/usr/lib/fs/vxfs/fsckptadm list /vm2

Output:
/vm2

NBU+2004.04.02.10h53m22s:

ctime = Fri Apr 02 10:53:23 2004

mtime = Fri Apr 02 10:53:23 2004

flags = removable

In this example, the name of the checkpoint is NBU+2004.04.02.10h53m22s.

b. Remove the checkpoint by entering the following:

/usr/lib/fs/vxfs/fsckptadm remove name_of_checkpoint /file_system
For example:
/usr/lib/fs/vxfs/fsckptadm remove NBU+2004.04.02.10h53m22s /vm2

c. If the checkpoint cannot be removed, unmount it (umount) and retry step b.

For more detail on removing VxFS clones, refer to the recommended actions for
NetBackup status code 11 in the NetBackup Troubleshooting Guide.

For TimeFinder, ShadowImage, BusinessCopy (Solaris or HP only):

Do the following on the client or alternate client, depending on the type of backup:

a. To discover and remove any VxVM clones, follow the steps under Removing a
VxVM Volume Clone on page 211.

b. Enter the following to resynchronize the mirror disks:

208 NetBackup Advanced Client System Administrators Guide

 Removing a Snapshot

For EMC arrays (TimeFinder):

symmir -g device_group establish LdevName
where LdevName is the logical device name of the standard device.
For Hitachi and HP arrays (ShadowImage, BusinessCopy):
pairresync -g groupname -d dev_name
For more information about EMC, Hitachi, and HP arrays and resynchronizing
disks, refer to the section Array-Specific Snapshot Methods on page 150.

For VxFS_Snapshot (Solaris or HP only):

Using the mounted file system found at step 3, unmount the snapshot as follows:
umount -F vxfs /tmp/_vrts_frzn_img__filesystemname_pid

For FlashSnap (Solaris and HP):

Do the following on the client or alternate client, depending on the type of backup:

a. Find the VxVM disk group:

vxdg list
The format of the disk group name is as follows:
SPLIT-primaryhost_diskgroup

If vxdg list does not show the disk group, the group might have been deported.
You can discover all the disk groups, including deported ones, by entering:
vxdisk -o alldgs list

The disk groups listed in parentheses are not imported on the local system.

b. Deport the VxVM disk group:

vxdg deport SPLIT-primaryhost_diskgroup
Enter the following on the primary (original) client:

c. Import and join the VxVM disk group:

vxdg import SPLIT-primaryhost_diskgroup

vxrecover -g SPLIT-primaryhost_diskgroup -m

vxdg join SPLIT-primaryhost_diskgroup diskgroup

d. Start the volume and snap back the snapshot volume:

vxvol -g SPLIT-primaryhost_diskgroup start SNAP-diskgroup_volume

Chapter 11, Troubleshooting 209

Removing a Snapshot

vxassist snapback SNAP-diskgroup_volume

Example
In this example, chime is the primary client and rico is the alternate client. 1hddg is
the name of the original disk group on chime, and chime_lhddg is the split group
that was imported on rico and must be rejoined to the original group on the primary
chime.
On alternate client rico, enter:
vxdg deport chime_lhddg

On primary client chime, enter:

vxdg import chime_lhddg

vxrecover -g chime_lhddg -m

vxdg join chime_lhddg lhddg

vxvol start SNAP-lhddg-vol01

vxassist snapback SNAP-lhddg-vol01

For FlashSnap (Windows):

a. Find the VxVM disk group:

vxdg list
The format of the disk group name is as follows:
SPLIT-primaryhost_diskgroup

b. Deport the VxVM disk group:

vxdg -g split_diskgroup deport
Enter the following on the primary (original) client:

c. Import and join the VxVM disk group:

vxassist rescan

vxdg -g split_diskgroup import

vxdg -g split_diskgroup -n diskgroup join

d. Snap back the snapshot volume:

vxassist snapback \Device\HarddiskDmVolumes\diskgroup\snap_volume

210 NetBackup Advanced Client System Administrators Guide

 Removing a VxVM Volume Clone

Removing a VxVM Volume Clone

A form of snapshot that might need manual deletion is a VxVM volume clone. See Disk
Group Clones on page 170 for a description of disk clones.
Major system interruptions, such as a system crash or unexpected reboot, may prevent
NetBackup from removing the clone. If the clone is not removed, subsequent backups of
the clients data will fail. Examine the /usr/openv/netbackup/logs/online_util
log for text such as the following:
19:13:07.686 [14981] <2> onlfi_vfms_logf: INF - do_cmd: Command failed with status=20:

/usr/openv/netbackup/bin/bpdgclone -g wil_test -n vol01 -f /var/tmp/HDSTFCAAs7aOqD

</dev/null >/var/tmp/VfMSAAAq7aOqD 2>/var/tmp/VfMSBAAr7aOqD

19:13:07.687 [14981] <2> onlfi_vfms_logf: INF - --- Dumping file /var/tmp/VfMSAAAq7aOqD

(stdout):

19:13:07.687 [14981] <2> onlfi_vfms_logf: INF - --- End of file /var/tmp/VfMSAAAq7aOqD

19:13:07.687 [14981] <2> onlfi_vfms_logf: INF - --- Dumping file /var/tmp/VfMSBAAr7aOqD

(stderr):

19:13:07.687 [14981] <2> onlfi_vfms_logf: INF - clone group and volume already exists

19:13:07.688 [14981] <2> onlfi_vfms_logf: INF - --- End of file /var/tmp/VfMSBAAr7aOqD

In this case, you must use the bpdgclone command with the -c option to remove the
clone, and then resynchronize the mirror disk with the primary disk.

How to remove the clone

The following commands should be run on the client or alternate client, depending on the
type of backup.

1. When no backups are running, use the following VxVM command to list any clones.
vxdg list

Note If a backup configured with an array-specific snapshot method is currently running,

a clone for that backup will appear in the vxdg output. Do not delete the clone;
NetBackup will delete it when the backup completes.

Example vxdg output:

NAME STATE ID

rootdg enabled 983299491.1025.turnip

VolMgr enabled 995995264.8366.turnip

wil_test_clone enabled 1010532924.21462.turnip

wil_test enabled 983815798.1417.turnip

In this example, the name suffix indicates wil_test_clone was created for a
snapshot backup that was configured with an array-specific snapshot method. If a
backup failed with log entries similar to those included above, the clone must be
manually deleted.

Chapter 11, Troubleshooting 211

Removing a VxVM Volume Clone

2. To remove the clone, enter the following:

/usr/openv/netbackup/bin/bpdgclone -g disk_group -n volume -c clone

For the above example, you would enter:

/usr/openv/netbackup/bin/bpdgclone -g wil_test -n vol01 -c wil_test_clone

where wil_test is the name of the disk group, vol01 is the name of the VxVM
volume, and wil_test_clone is the name of the clone. Use the Volume Manager
vxprint command to display volume names and other volume information.
For more information, refer to the bpdgclone man page. For assistance with
vxprint and other Volume Manager commands, refer to the VERITAS Volume
Manager Administrators Guide.

3. To verify that the clone has been removed, re-enter vxdg list.
Sample output:
NAME STATE ID
rootdg enabled 983299491.1025.turnip
VolMgr enabled 995995264.8366.turnip
wil_test enabled 983815798.1417.turnip
The clone no longer appears in the list.

212 NetBackup Advanced Client System Administrators Guide

Managing Snapshots from the Command

Line A

This appendix presents basic procedures for creating and managing snapshots using the
NetBackup command-line interface. As compared to the NetBackup Administration
Console, use of NetBackup commands (such as in a backup script) provides additional
flexibility that may be desirable in certain circumstances.

Note The command for creating a snapshot is bpfis, which is described in man page
format in the NetBackup Commands manual.

213
Reasons for Using the Command Line

Reasons for Using the Command Line

Configuration of snapshot backups is normally done from the Policy dialogs in the
NetBackup Administration Console, as explained in Policy Configuration chapter.
When these snapshot backups are run automatically by schedule, or started manually
from the Administration Console, NetBackup takes care of all snapshot processing and
backs up the snapshot as the policy dictates. In most cases, these GUI-driven backups
offer all the control that is needed; no command-line intervention is required on your part.
Certain managerial tasks, however, cannot be performed from the Administration
Console:
Creating a snapshot without an immediate backup. From the command line, you can
create the snapshot as a separate action, and either delay the backup or omit it
entirely.
Using a snapshot to gather diagnostic or statistical information about client and
system resources. Snapshots can provide a unique point-in-time view.
Coordinating with database quiesce and job scheduling.

Notes on Using the Command Line

The NetBackup command for creating a snapshot is bpfis. This command is used by all
snapshot backups, including those configured in the Administration Console. The bpfis
command can create, delete, and catalog snapshots. Snapshots can be created on primary
or alternate clients. Snapshot sources1 can be raw disk partitions, volumes, or file systems.
Snapshots created by the bpfis command can be backed up at any time using the
bpbackup command. As an alternative, the snapshot may be examined or processed in
some other way, for example, to gather statistics on the state of application data at a point
in time.
After the snapshot is created, it remains mounted and read-only by default. The snapshot
may be remounted in a writable state if desired.
When creating a snapshot, bpfis returns an identifier for the snapshot. A default
identifier is returned if none is specified by the user. The identifier can be used to query
the status of the snapshot, or to delete it. The bpfis command also returns a set of
Backup Selections list directives which can be used to back up the snapshot.

1. A snapshot source is the entity containing the data to be captured, where entity means raw
disk partition, volume, or file system.

214 NetBackup Advanced Client System Administrators Guide

 Examples

Examples
The following examples show how to use the command line interface to create and back
up a snapshot.

Note The man pages for bpfis and other NetBackup commands are in the NetBackup
Commands manual. For a list of the snapshot options available with bpfis, refer to the
<opt_params> area of each snapshot method (FIM) listed in the
/usr/openv/netbackup/vfm.conf file.

Example 1 (using nbu_snap method)

This example creates and backs up an nbu_snap snapshot. It assumes the following:
The file system to back up is mounted on /mnt. This is the snapshot source.
A policy has been created (mypol) with an empty Backup Selections list and
appropriate storage units, schedule, and client.
If you are using the policy dialogs of the Administration Console, creating a policy
with no Backup Selections entries for the client will generate an error message. When
prompted, click Yes to ignore the error.
The commands in this example were entered on a Solaris system in the Korn shell.

Note All commands in this example are executed on the (primary) client.
You must have root privileges to execute the bpfis command. bpfis is located in
/usr/openv/netbackup/bin.

1. Create the snapshot, using nbu_snap as the method:

bpfis create -id my_id -fim "nbu_snap:cache=/dev/rdsk/c1t1d1s2" /mnt

where my_id is a user-defined identifier for the snapshot, cache= designates a raw
partition for the copy-on-write mechanism used by the nbu_snap method, and /mnt
is the source for the snapshot (the file system to back up).
In bpfis output, the original mount point of the snapshot source (/mnt) and the
active mount point of the snapshot are embedded in the directive labeled REMAP
FILE, where USING /path_name indicates the mount point of the snapshot. The
OPTIONS directives are internal to NetBackup and should not be changed. These
directives are also written to file /tmp/filelist.my_id. (This file is written on the
alternate backup client if alternate client backup is employed.)

2. Add Backup Selections list directives to the policy:

bpplinclude mypol -add "FIM=NONE"

cat /tmp/filelist.my_id | while read dir; do

Appendix A, Managing Snapshots from the Command Line 215

Examples

> bpplinclude mypol -add "$dir"

> done

The above command takes the directives from the /tmp/filelist.my_id file
(created by bpfis) and adds them to the policy.

3. Make sure the policy is enabled for snapshots and that the backup copy type is local
(not offhost).
bpplinfo mypol -modify -bc 0 -fi 1

where -bc specifies that the backup copy method is local and -fi specifies enable
snapshots for this policy.

4. Make sure the NetBackup catalog is updated by listing policy attributes and the
Backup Selections list.
bpplinfo mypol -L

bpplinclude mypol -l

The output should include snapshot: yes, and Backup Copy: 0, where 0 indicates
local.

5. Back up the snapshot.

bpbackup -i -w 0 -p mypol

where -i means make the backup immediately, -w 0 means wait indefinitely for
completion status from the server before returning to the system prompt, and -p
specifies the policy.

6. If desired, delete the snapshot.

bpfis delete -id my_id

where my_id is the identifier you entered at step 1.

Example 2 (for alternate client backup, using FlashSnap method)

This example creates and backs up a snapshot on the alternate client alt, and uses the
FlashSnap snapshot method. All other assumptions are the same as those in example 1.

Note You must have root privileges to execute the bpfis command.
bpfis is located in /usr/openv/netbackup/bin.

1. Create the snapshot, using FlashSnap as the snapshot method:

bpfis create -id my_id -fim FlashSnap -rhost alt /mnt

216 NetBackup Advanced Client System Administrators Guide

 Examples

where my_id is a user-defined identifier for the snapshot, -rhost designates the
alternate client named alt, and /mnt is the source for the snapshot on the primary
client (the file system to back up).
For a brief description of bpfis output, refer to step 1 on page 215.

2. Add Backup Selections list directives to the policy:

bpplinclude mypol -add "FIM=NONE:alternate_backup_client=alt"

rsh alt cat /tmp/filelist.my_id | while read dir; do

> bpplinclude mypol -add "$dir"

> done

3. Make sure the policy is enabled for snapshots and that the backup copy type is local
(not offhost).
bpplinfo mypol -modify -bc 0 -fi 1

where -bc specifies that the backup copy method is local and -fi specifies enable
snapshots for this policy.

4. Make sure the NetBackup catalog is updated by listing policy attributes and the
Backup Selections list.
bpplinfo mypol -L

bpplinclude mypol -l

The output should include snapshot: yes, and Backup Copy: 0, where 0 indicates
local.

5. Back up the snapshot.

bpbackup -i -w 0 -p mypol

where -i means make the backup immediately, -w 0 means wait indefinitely for
completion status from the server before returning to the system prompt, and -p
specifies the policy.

6. If desired, delete the snapshot.

bpfis delete -id my_id

where my_id is the identifier you entered at step 1.

Appendix A, Managing Snapshots from the Command Line 217

Examples

218 NetBackup Advanced Client System Administrators Guide

Processing Background B

This appendix presents background information on snapshot processing and

backup/restore processing.
Processing Before and After the Snapshot
FlashBackup: Local and Offhost Diagrams
Process Diagrams: Backup and Restore
How Copy-on-Write Works

219
Processing Before and After the Snapshot

Processing Before and After the Snapshot

NetBackup performs several vital functions prior to creating a snapshot, as outlined
below and in the following text. Without this pre-processing, the integrity of the snapshot
cannot be guaranteed and the backup data may be of no value.

NetBackup Processing Before and After Creating the Snapshot

1. Backup process requests database quiesce.

2. Database application quiesces (must wait for transactions

to complete).
Steps 1, 2, and 6 apply
3. Lock and flush the file system.
only to databases, such
as those requiring
NetBackup for Oracle 4. Create the snapshot.
Advanced Client.
5. Unlock the file system.

6. Release (unquiesce) the application.

7. Back up the snapshot.

8. (Optional:) Remove the snapshot.

Quiescing the System

Before a useful snapshot can be created, the data to back up must be transactionally
consistent or complete. A transaction is a single data action, such as updating a patients
record in a medical database, or creating a record for a new patient. Such a transaction is
composed of multiple I/O requests (search, copy, send, write, and so forth). Until the
transactions I/O requests are complete, the data is inconsistent and may be unsuitable for
backup.
Transactions affect all levels of the storage management stack (file system, volume
manager, and so forth), generating further transactions as a request is handed off to the
next level of the stack. From the viewpoint of the file system, for instance, an I/O request
issued by a database application constitutes a transaction and may be split into many disk
references, all of which must be complete for the original request to be fulfilled. As a
result, the creation of the snapshot must be coordinated with any application or other
process that can affect the transactional consistency of the data.

220 NetBackup Advanced Client System Administrators Guide

 Processing Before and After the Snapshot

The means of coordination is called quiesce (literally, to make quiet or place in repose). This
involves pausing the database application or process until the data is transactionally
consistent. Applications and the storage management stack must all be quiesced before a
useful snapshot can be made.

Quiescing the Database Application

Most database applications are transactionally consistent only at particular points in time.
Sometimes, they are consistent only after they have been shut down. Since there is a
growing need for database applications to remain up and available constantly, many
applications are now designed to reach a point of transactional consistency at regular
intervals or in response to an external event. This process is called application quiesce,
described below.
In database application quiesce, an external signal or message is sent to a receptive
database. In response, the database finishes the current transaction or group of
transactions, signaling the snapshot consumer when this is complete. The database then
waits for a second signal indicating that normal operations can resume. After the database
signals that it has reached a state of transactional consistency, the final steps of creating the
snapshot can proceed.
Once the snapshot has been created, another signal is sent to the waiting database telling
it to resume processing. This is called unquiescing the application.

Appendix B, Processing Background 221

Processing Before and After the Snapshot

Dialog for Quiesce/Unquiesce

Snapshot Consumer
Database Application (NetBackup)
1 Quiesce.

2
Finish transactions.
3
Quiesce acknowledge.

... In quiesce mode .... 4

Quiesce stack,
trigger snapshot.
5
Unquiesce.
6
Out of quiesce mode.
Continue processing.

Quiescing the Stack

The storage management stack is a layered arrangement of software elements. An I/O
request originated by a database application passes from element to element until a
hardware request to move data reaches the storage network. Each stack element performs
a variety of functions, some of which treat I/O requests like transactions to assure their
completion. Before a snapshot is created, the stack must be quiesced (made transactionally
consistent).
Since the file system is the front-line interface to applications for managing files and
performing I/O, file system quiesce is a critical part of quiescing the stack.

File System
Two of the principal tasks of quiescing the file system are the following:
Prohibit new I/O requests from initiating, which is called locking the file system.
Flush file system cache (write cached data back to disk). The system must complete
any outstanding application I/O and note completion of outstanding metadata
updates.

222 NetBackup Advanced Client System Administrators Guide

 FlashBackup: Local and Offhost Diagrams

Volume Manager
As in a file system, the volume managers data caching may have to be flushed and
disabled until the snapshot is created. As long as volume manager caching is enabled,
data required for a consistent image may be lingering in volume manager cache rather
than being available on disk when the snapshot is created.

FlashBackup: Local and Offhost Diagrams

This section is an overview of FlashBackup processing with respect to local and offhost
backup. The FlashBackup process is slightly different from other Advanced Client backup
types described under Offhost Backup Methods on page 10.

Local Client
A FlashBackup image can be backed up from local disks to a storage device on the media
server, using the local client. If Perform snapshot backups is not selected on the policy
Attributes display, nbu_snap (on Solaris) or VxFS_Snapshot (on HP) creates the snapshot.
If Perform snapshot backups is selected, a different snapshot method can be configured.
In either case, a fibre channel network or SAN is not required for a local backup.
The following diagram shows a network configuration sufficient for a local client backup.
The network configuration is identical to that for normal NetBackup.

NetBackup master
server
1
LAN / WAN

3 4

Client Media
SCSI
server
2 SCSI
5
storage
Client disks

1. Client backup is initiated by master server, which tells the NetBackup client to create the
snapshot data on disk.
2. Client creates file system map of the snapshot.
3. Client sends the file system map to the media server.
4. Client sends the backup data to the media server.
5. Media server writes data to local storage.

Appendix B, Processing Background 223

FlashBackup: Local and Offhost Diagrams

NetBackup Media Server

In this offhost backup method, the NetBackup media server performs the backup I/O
processing (data movement). Choose this option if processing time on the NetBackup
client is critical and off-loading backup processing to the media server would save time.
The figure below shows the basic components used in the NetBackup Media Server
method with FlashBackup.

Note If you have a multi-ported SCSI disk array, a fibre channel SAN is not required. See
Offhost Backup Without a SAN (UNIX Only) on page 16.

FlashBackup with NetBackup Media Server Offhost Method

LAN / WAN

1 4 3
NetBackup NetBackup NetBackup
master client media server
server

2 3 6
5

Fibre Channel/SAN
Robot on
SAN
Disks of client
data on SAN

1. Backup is initiated by master server, which tells the client to create the snapshot data on the disk.
2. Client creates file system map of the snapshot and sends the file system map over the LAN to the
media server.
3. Media server reads the file system map over the LAN and writes it to tape storage.
4. Client creates the disk block map over the snapshot and sends the disk block map over the LAN
to the media server.
5. Media server reads the data over the SAN from the disk block map provided by the client.
6. Media server writes data across the SAN to storage.

224 NetBackup Advanced Client System Administrators Guide

 FlashBackup: Local and Offhost Diagrams

Third-Party Copy Device

In this offhost backup method, a third-party copy device (not a component of NetBackup)
performs the backup data movement. Choose this option if processing time on the
NetBackup client is critical and off-loading backup processing to a third-party copy
device would save time. The figure below shows the basic components used in the
Third-Party Copy Device method with FlashBackup.

FlashBackup with Third-Party Copy Device Offhost Method

LAN / WAN

1 4 3
NetBackup NetBackup NetBackup
master client media server
server

2 5
3

Fibre Channel/SAN

Robot on
7 SAN
6 Third-party
copy device
Disks of client SCSI
data on SAN

1. Client backup is initiated by master server, which tells the NetBackup client to create the snapshot data
on the disk.
2. Client creates file system map of snapshot and sends file system map over the LAN to the media server.
3. Media server reads the file system map over the LAN and writes it to tape storage.
4. Client creates the disk block map over the snapshot and sends the disk block map over the LAN to the
media server.
5. Media server creates third-party copy commands based on disk block map and sends the commands to
the third-party copy device over the SAN.
6. The third-party copy commands instruct the third-party copy device to read the client data from either
SAN-attached or SCSI-attached disk.
7. The third-party copy commands instruct third-party copy device to write the client data to SAN-attached
or SCSI-attached storage.

Appendix B, Processing Background 225

Process Diagrams: Backup and Restore

Process Diagrams: Backup and Restore

The following diagrams show the major components (NetBackup daemons) involved in
offhost and snapshot backup and restore.

Alternate Client Backup: Split-Mirror/VVR

This diagram shows the major components for making a backup using alternate client
backup with a mirror snapshot. Numbers refer to descriptions on the next page.

Overview of Alternate Client Backup

Master
server Admin Backup
bprd
Console Request
1

bpsched bpdbm

Media server 2
bpcd Tape or optical
device

bptm
bpbrm 6
5

Primary client Alternate client

4
bpbkar
Backup Request 3
jbpSA or
bpfis bpfis
bp

Active Shared disk or snapshot

client data replication to back up

226 NetBackup Advanced Client System Administrators Guide

 Process Diagrams: Backup and Restore

1. The NetBackup master server or primary client initiates the backup, causing the
NetBackup request daemon bprd to start the scheduler, bpsched. bpsched
processes the policy configurations depending on the initiator of the backup
(scheduled, immediate manual, or user directed). Refer to Appendix A of the
NetBackup Troubleshooting Guide for more information on this stage of the backup
operation.

2. bpsched uses bpcd (client daemon) to start the backup/restore manager (bpbrm) on
the media server.

3. bpbrm starts bpfis on the primary client and creates a snapshot on the alternate
client.

4. bpbrm starts bpbkar on the alternate client to back up the snapshot.

5. bpbrm starts bptm on the media server to send the backup to the storage device.

6. A standard backup occurs on the alternate client. bpbrm optionally removes the
snapshot.

Appendix B, Processing Background 227

Process Diagrams: Backup and Restore

NetBackup Media Server: Copy-On-Write Process

The following diagram shows the major components for making a backup using the
NetBackup Media Server method with a copy-on-write snapshot. Numbers refer to
descriptions on the next page.

Overview of NetBackup Media Server Process: Copy-On-Write

1 Solaris client
Master Backup
server Request
jnbSA or bprd
bpadm jbpSA or
Backup bp
Request
bpdbm
bpsched

Media server
Fil
e
inf

bpcd 2
or
ma
tio
n

bpcd
3
bpbrm File
infor 4
mati
bptm on
(parent) bpbkar
Shared
bptm Make
Memory list
Extent
Tape

(child) snapshot
request

7
Ba

extent
ck

9
up

10 list info 5
snapshot
i
ma

mapping
6 services
ge

Ba

8 services
ltid
cku

File/volume
mo
p im

un mapping
t info snapshot
age

created

Tape or optical snapshot

Active
device of client data
client data
(cached)

228 NetBackup Advanced Client System Administrators Guide

 Process Diagrams: Backup and Restore

1. The NetBackup master server or client initiates the backup, causing the NetBackup
request daemon bprd to start the scheduler, bpsched. bpsched processes the policy
configurations depending on the initiator of the backup (scheduled, immediate
manual, or user directed). Refer to Appendix A of the NetBackup Troubleshooting Guide for
more information on this stage of the backup operation.

2. bpsched uses bpcd (client daemon) to start the backup/restore manager (bpbrm) on
the media server.

3. bpbrm starts the Media Manager process bptm (parent) and also starts the actual
backup by using bpcd on the client to start the clients backup and archive program
bpbkar.

4. bpbkar sends information about files within the image to the backup/restore
manager bpbrm, which directs the file information to bpdbm for the NetBackup file
database on the master server.

5. bpbkar requests creation of a snapshot of the clients active data. bpbkar uses the
snapshot method that was configured for the snapshot source.

6. bpbkar requests file/volume mapping information about the client data. bpbkar
uses one or more mapping services to decompose the clients data into physical disk
addresses (also referred to as disk extents). The file/volume mapping information (list
of extents) comes from one of two places: the clients active (primary) data, and from
the snapshot of the client data (cached).

7. On the media server, bptm creates a child process, which reads the mapping
information (extent list) from bpbkar.

8. Based on the extent list received from bpbkar, bptm reads the client data (backup
image) from two places: from the clients active data (for those blocks that have not
changed since the backup was initiated), and from the snapshot cache (to obtain the
original contents of the blocks that have changed since the backup was initiated).

9. The bptm child stores the client data block-by-block in shared memory.

10. The parent bptm process then takes the backup image from shared memory and
sends it to the storage device. For information on how the tape request is issued, refer
to Appendix A of the NetBackup Troubleshooting Guide.

Appendix B, Processing Background 229

 Process Diagrams: Backup and Restore

NetBackup Media Server: Mirror Process

The following diagram shows the major components for making a backup using the
NetBackup Media Server method with a mirror snapshot. Numbers refer to descriptions
on the next page.

Overview of NetBackup Media Server Process: Mirror

Master server Solaris or HP client
1
Backup
jnbSA or Request jbpSA or
bprd
bpadm Backup bp
Request

bpdbm
bpsched

Media server
Fil
ei
nfo

bpcd 2
r
ma
tio
n

bpcd
3
bpbrm File
in 4
form
bptm ation
(parent) 7 bpbkar
Shared Make
bptm ts
Memory Disk exten snapshot
Tape

(child)
request

g )
Ba

(mappin Disk
ck

9 5
up

10 extents
snapshot
im

mapping
services
ag

6
Ba

8 services
e

ltid
cku

mo
p

un
ima

t snapshot
File/volume
ge

mapping created
info

Tape or optical snapshot

of client data Active
device
(cached) client data

Mirror (secondary) Primary

230 NetBackup Advanced Client System Administrators Guide

 Process Diagrams: Backup and Restore

1. The NetBackup master server or client initiates the backup, causing the NetBackup
request daemon bprd to start the scheduler, bpsched. bpsched processes the policy
configurations depending on the initiator of the backup (scheduled, immediate
manual, or user directed). Refer to Appendix A of the NetBackup Troubleshooting Guide for
more information on this stage of the backup operation.

2. bpsched uses bpcd (client daemon) to start the backup/restore manager (bpbrm) on
the media server.

3. bpbrm starts the Media Manager process bptm (parent) and also starts the actual
backup by using bpcd on the client to start the clients backup and archive program
bpbkar.

4. bpbkar sends information about files within the image to the backup/restore
manager bpbrm, which directs the file information to bpdbm for the NetBackup file
database on the master server.

5. bpbkar requests creation of a snapshot of the clients active data. bpbkar uses the
snapshot method that was configured for the snapshot source.

6. bpbkar requests file/volume mapping information about the client data. bpbkar
uses one or more mapping services to decompose the clients data into physical disk
addresses (also referred to as disk extents). The file/volume mapping information (list
of disk extents) comes from the snapshot of the client data.

7. On the media server, bptm creates a child process, which reads the mapping
information (disk extent list) from bpbkar.

8. Based on the extent list received from bpbkar, bptm reads the client data (backup
image) from the snapshot on the mirror (secondary) disk.

9. The bptm child stores the client data block-by-block in shared memory.

10. The parent bptm process then takes the backup image from shared memory and
sends it to the storage device. For information on how the tape request is issued, refer
to Appendix A of the NetBackup Troubleshooting Guide.

Appendix B, Processing Background 231

Process Diagrams: Backup and Restore

Third-Party Copy Device: Copy-On-Write Process

The following diagram shows the major components for making a backup using the
NetBackup Third-Party Copy Device method with a copy-on-write snapshot. Numbers
refer to descriptions on the next page.

Overview of Third-Party Copy Device Process: Copy-On-Write

Master server
Backup 1 Solaris client
Request
jnbSA or bprd
bpadm jbpSA or
Backup bp
Request
bpdbm
bpsched

Media server
Fil
ei
nfo

bpcd 2
rm
ati
on

bpcd
3
bpbrm F ile 4
infor
mati
on
bptm
bpbkar
Extent list Make
Tap 7
e extent snapshot
req
ue st
8 list info
ltid mapping
Third-party snapshot 5
services services
copy command mount
with extent list 6

10 File/volume
mapping info
ge
up ima
k Tape or optical
Bac snapshot
device Active
created
Third-party copy client data
device
Backup image
9
snapshot of
client data
(cached)

232 NetBackup Advanced Client System Administrators Guide

 Process Diagrams: Backup and Restore

1. The NetBackup server or client initiates the backup, causing the NetBackup request
daemon bprd to start the scheduler, bpsched. bpsched processes the policy
configurations depending on the initiator of the backup (scheduled, immediate
manual, or user directed). Refer to Appendix A of the NetBackup Troubleshooting Guide for
more information on this stage of the backup operation.

2. bpsched uses bpcd (client daemon) to start the backup/restore manager (bpbrm) on
the media server.

3. bpbrm starts the Media Manager process bptm and also starts the actual backup by
using bpcd on the client to start the clients backup and archive program bpbkar.

4. bpbkar sends information about files within the image to the backup/restore
manager bpbrm, which directs the file information to the NetBackup file database on
the master server.

5. bpbkar requests creation of a snapshot of the clients active data, by means of a

snapshot method. bpbkar uses the snapshot method that was configured for the
snapshot source.

6. bpbkar requests file/volume mapping information about the client data. bpbkar
uses one or more mapping services to decompose the clients data into physical disk
addresses (also referred to as disk extents). This file/volume mapping information (list
of extents) comes from one of two sources: the clients active (primary) data, or from
the snapshot of the client data (cached).

7. bptm reads the mapping information (extent list) from bpbkar.

8. bptm sends the third-party copy command with the extent list to the third-party copy
device. For information on how the tape request is issued, refer to Appendix A of the
NetBackup Troubleshooting Guide.

9. The third-party copy device reads the backup image (client data) from two places:
from the clients active data (for those blocks that have not changed since the backup
was initiated), and from the snapshot cache (for the original contents of the blocks that
have changed since the backup was initiated).

10. The third-party copy device sends the backup image to the storage device.

Appendix B, Processing Background 233

Process Diagrams: Backup and Restore

Third-Party Copy Device: Mirror Process

The following diagram shows the major components for making a backup using the
NetBackup Third-Party Copy Device method with a mirror snapshot. Numbers refer to
descriptions on the next page.

Overview of NetBackup Third-Party Copy Process: Mirror

Master server
Solaris or HP client
Backup 1
Request
jnbSA or bprd
bpadm jbpSA or
Backup bp
Request
bpdbm
bpsched

Media server
Fil
ei

bpcd 2
nfo
rm
ati
o

bpcd
n

3
bpbrm F ile 4
infor
mati
on
bptm
bpbkar
Extent list 7 Make
Tap
e req extent snapshot
ues
t list info
8 ltid
mapping snapshot
Third-party 5
services services
copy command mount
with extent list
6 snapshot
File/volume created
10
mapping info
e
mag
ckup i Tape or optical
Ba
device
Third-party copy snapshot Active
device of client data client data
Backup image
9
Mirror (secondary) Primary

234 NetBackup Advanced Client System Administrators Guide

 Process Diagrams: Backup and Restore

1. The NetBackup server or client initiates the backup, causing the NetBackup request
daemon bprd to start the scheduler, bpsched. bpsched processes the policy
configurations depending on the initiator of the backup (scheduled, immediate
manual, or user directed). Refer to Appendix A of the NetBackup Troubleshooting Guide for
more information on this stage of the backup operation.

2. bpsched uses bpcd (client daemon) to start the backup/restore manager (bpbrm) on
the media server.

3. bpbrm starts the Media Manager process bptm and also starts the actual backup by
using bpcd on the client to start the clients backup and archive program bpbkar.

4. bpbkar sends information about files within the image to the backup/restore
manager bpbrm, which directs the file information to the NetBackup file database on
the master server.

5. bpbkar requests creation of a snapshot of the clients active data, by means of a

snapshot method. bpbkar uses the snapshot method that was configured for the
snapshot source.

6. bpbkar requests file/volume mapping information about the client data. bpbkar
uses one or more mapping services to decompose the clients data into physical disk
addresses (also referred to as disk extents). This file/volume mapping information (list
of extents) comes from the snapshot of the client data on the mirror (secondary) disk.

7. bptm reads the mapping information (extent list) from bpbkar.

8. bptm sends the third-party copy command with the extent list to the third-party copy
device. For information on how the tape request is issued, refer to Appendix A of the
NetBackup Troubleshooting Guide.

9. The third-party copy device reads the backup image (client data) from the snapshot
on the mirror (secondary) disk.

10. The third-party copy device sends the backup image to the storage device.

Appendix B, Processing Background 235

Process Diagrams: Backup and Restore

Instant Recovery: Backup Process Overview

The following diagram shows the major components for making an Instant Recovery
image. Numbers refer to descriptions on the following page.

Instant Recovery Image: Backup Process Overview

Master server Client

Backup 1
Backup
Request
Admin bprd Request user
Console interface

bpsched
bpdbm

2
Fi
le bpcd
inf bpcd
or
m
at 3
ion
bpbrm File
6 in form
a tion 4

bpfis

Make
snapshot

snapshot 5
services

snapshot
captured

snapshot of
client data

236 NetBackup Advanced Client System Administrators Guide

 Process Diagrams: Backup and Restore

1. The NetBackup master server or client initiates the backup, causing the NetBackup
request daemon bprd to start the scheduler, bpsched. bpsched processes the policy
configurations depending on the initiator of the backup (scheduled, immediate
manual, or user directed). Refer to Appendix A of the NetBackup Troubleshooting Guide for
more information on this stage of the backup operation.

2. bpsched uses bpcd (client daemon) to start the backup/restore manager (bpbrm) on
the master server.

3. bpbrm starts the backup by using bpcd on the client to start the clients backup and
archive program bpbkar.

4. bpfis requests creation of a snapshot of the clients active data. bpbkar uses the
snapshot method that was configured for the snapshot source.

5. bpfis captures the snapshot, mounts it, sends catalog information, then unmounts
the snapshot.

6. bpfis sends information about files within the snapshot to the backup/restore
manager bpbrm, which directs the file information to bpdbm for the NetBackup file
database on the master server.

Appendix B, Processing Background 237

Process Diagrams: Backup and Restore

Instant Recovery: Backup to disk only

The following diagram shows the major components in a backup to disk only, with Instant
Recovery enabled.

Master server Client

1
bpsched
2
bpbrm bpfis

bpdbm
Snapshot of
client data

Catalog

1. bpsched receives and validates the policy attribute Retain snapshots for instant
recovery and schedule attribute Instant recovery backups to disk only to
determine what type of backup should be started.

2. In either case, bpsched calls bpbrm which in turn invokes bpfis to create the necessary
snapshots.

3. For Instant Recovery backups to disk only, bpfis creates the catalog entries.

238 NetBackup Advanced Client System Administrators Guide

 Process Diagrams: Backup and Restore

Instant Recovery: Backup to tape

The following diagram shows the major components in a tape backup with Instant
Recovery enabled.

Master server Client

1
bpsched bpbrm bpfis

4 3
Snapshot of
client data

bptm bpbkar
5

1. bpsched starts bpfis on the client through bpbrm.

2. bpfis creates snapshot and returns the modified backup selections list to bpbrm and
bpsched.

3. bpbrm starts bpbkar with the modified backup selections list.

4. bpbrm starts bptm.

5. bpbkar and bptm back up the snapshot data to tape.

Appendix B, Processing Background 239

Process Diagrams: Backup and Restore

Instant Recovery: Normal Restore

The following diagram shows the major components in a normal backup restore from an
Instant Recovery image.

Instant Recovery Image: Restore Process

Master server Client

2 Restore 1
Request
bprd Backup, Archive,
Restore

bpcd 3
5

bpbrm bppfi
4

6
Persistent
bpbkar snapshot of
client data

8 7

tar Client file

system
Restore to client

1. User submits the restore request.

2. A handler for the restore request is started by the request service (bprd) on the master
server. The restore request handler uses catalog information to produce a detailed file
selection list, identifying which files are to be restored from which image. For each
image, a restore request is sent to the server.

3. The client request service (bpcd) on the server starts a backup and restore manager
(bpbrm) to process the restore.

4. bpbrm starts two agents on the client: tar, and bppfi (the Instant Recovery agent).

240 NetBackup Advanced Client System Administrators Guide

 Process Diagrams: Backup and Restore

5. bppfi connects back to the restore request handler (bprd) on the master server to
collect the list of files to be restored from this image. It also connects to tar to pass the
connection to the backup agent (bpbkar).

6. bppfi then starts the backup agent bpbkar and sends it the Backup Selections list.

7. tar consumes the data stream generated by bpbkar and restores the contents to the
clients original file system.

8. When tar completes the restore, it notifies the backup and restore manager (bpbrm) of
the completion. bpbrm notifies the restore request handler, and all processes
associated with the restore end.

Appendix B, Processing Background 241

Process Diagrams: Backup and Restore

Instant Recovery: Browsing from Snapshot

The following diagram shows the major restore components for browsing directly from an
Instant Recovery snapshot (non-database). NetBackup mounts all relevant snapshots so
that you can browse directly from the snapshots, instead of relying on catalog entries from
the .f files.

Master server Client

1 Browsing
2 for Restore
bplist Backup, Archive,
bprd
Restore
3

bpdbm
5
6 bpcd Snapshot of
4 client data

NetBackup
Catalog

1. User interface invokes bplist to display files during browse for restore.

2. bplist sends request to bprd with browsing criteria, such as time, backup id, and file
path.

3. bprd issues catalog queries Q_IMAGE_LIST, Q_IMAGE_BY_BACKUPID,

Q_IMAGE_LIST_FILES to obtain file details.

4. bpdbm searches the catalog. If it is a backup for Instant Recovery with a partial .f file
entry created by a non-database Instant Recovery backup, it starts bpcd and sends the
fragments that contain snapshot names, mount point, and mount device to bpcd.

5. bpcd decodes the fragments and mounts the snapshot(s) specified in the fragments.

6. A bpcd request BPCD_READ_DIR is then issued by bpdbm to browse the mounted

snapshot(s) and return information to bpdbm. bpdbm treats the .f file entries from
bpcd the same way as the entries from the .f file.

242 NetBackup Advanced Client System Administrators Guide

 Process Diagrams: Backup and Restore

Instant Recovery: File Promotion

The following diagram shows the major components in a restore using file promotion

from an Instant Recovery image.

Instant Recovery: File Promotion

Master server Client

Restore 1
Request
bprd Backup, Archive,
Restore
2
6
bpbrm
3
bppfi

bpbkar tar
7
5

Snapshot of Client file

client data system

1. User submits the file promotion request.

2. bprd starts bpbrm and sends Backup Selections list and associated snapshot
information.

3. bpbrm starts bppfi on the client and passes the Backup Selections and snapshot
information to it.

4. bppfi mounts the snapshot, starts bpbkar, and sends Backup Selections to bpbkar.

5. bpbkar goes through each of the files, promoting each file (if possible) from the
snapshot to the clients primary file system.

6. If promotion fails, bpbrm starts tar.

7. bpbkar sends a backup image of the file to tar, and tar performs a regular restore.

Appendix B, Processing Background 243

Process Diagrams: Backup and Restore

Instant Recovery: Rollback

The following diagram shows the major components in a restore using Point in Time
Rollback from an Instant Recovery image.

Instant Recovery: Rollback

Master server Client

Restore 1
Request
bprd Backup, Archive,
Restore
2

3
bpbrm bppfi

4
5

bpfis Snapshot Client file

of client system
data

1. User submits a rollback request.

2. bprd starts bpbrm and sends the backup id and associated snapshot information.

3. bpbrm starts bppfi on the client and passes the backup id and snapshot information to
bppfi.

4. bppfi uses bpfis to rollback the snapshot.

5. bpfis unmounts the primary file system, rolls back the snapshot, and remounts the file
system.

244 NetBackup Advanced Client System Administrators Guide

 Process Diagrams: Backup and Restore

FlashBackup: Backup Process Overview

The following figure shows the major NetBackup components involved in
FlashBackup/FlashBackup-Windows backups.

Master server
Backup Client
Request
Admin bprd
Console
Backup
Request
ifrfr
bpdbm
bpsched

Media server Policy

Ba

info
cku
pi
nfo

bpcd

Policy
info icy bpcd
Pol
info Policy
bptm info
(parent) bpbrm Backup
in fo
Backup Bac bpbkar
kup
image ima bptm e
ge
(child) Backup imag

Tape or Optical Client

disk

During the backup, NetBackup performs the following tasks:

If Perform snapshot backups is configured: bpbkar uses the snapshot method
configured for the raw partition entry in the policys Backup Selections list.

Appendix B, Processing Background 245

Process Diagrams: Backup and Restore

If Perform snapshot backups is not configured: by default, bpbkar turns on the

FlashBackup raw-partition snapshot mechanism (nbu_snap on Solaris or
VxFS_Snapshot on HP). When a block is updated in the partition, the previous
contents of the block are written to the cache partition.
Creates a file system map that represents files contained in the raw partition.
If the backup type is a file incremental, bpbkar creates a bit map that associates blocks
belonging to selected files with their physical addresses in the partition. Note that
block level incrementals are not supported.
Sends the backup information to bpbrm, which in turn directs the backup information
to bpdbm. The backup information is composed of:
File system map name
Bit map name (if incremental)
Individual file names
Raw-partition name
Transmits the backup image to bptm. The backup image is composed of:
File system map data blocks
Bit map data blocks
Raw-partition data blocks (raw-partition extents, if this is an offhost backup)
If Perform snapshot backups is not configured: bpbkar turns off the FlashBackup
snapshot mechanism.

246 NetBackup Advanced Client System Administrators Guide

 Process Diagrams: Backup and Restore

FlashBackup: Restore Process Overview

The following figure shows the major components involved in restoring files from
FlashBackup/FlashBackup-Windows images.

Master server Client

Restore Restore
Backup, Archive, request request
bprd Backup, Archive,
Restore
Restore
Backup
info
bpdbm
sfr

Media server ap
m
em
yst
s
le
Fi fo bpcd
r in
fe
ans
tr
le
Fi

bptm
(parent) bpbrm bpcd
D at
ab
lock
s
bptm
(child)
Data blocks tar

Tape or Optical
Client
disk

Appendix B, Processing Background 247

Process Diagrams: Backup and Restore

During a restore, the user browses for and selects files from FlashBackup images in the
same manner as standard backup images. The difference is that bprd calls the single file
restore (sfr) program when processing FlashBackup images. During a restore, the sfr
program performs the following tasks:
Retrieves information about the backup from bpdbm. The backup information is
composed of:
File system map name and its media location.
Bit map name and its media location.
Raw-partition name and its media location.
Using the backup information, sfr retrieves the file system map data by directing
bptm to the location of the file system map on the tape and then reading from the
tape. A similar procedure is followed to position bptm to obtain the bit map data
when restoring from incremental backup images and to obtain the raw-partition data
when restoring the raw partition.
Then, using the information contained in the file system map, sfr directs bptm to the
location of the individual file data blocks on the tape.
bptm then reads the file data blocks from the tape and writes them to the clients tar
program. Finally, tar writes the file data to the client disk.

248 NetBackup Advanced Client System Administrators Guide

 How Copy-on-Write Works

How Copy-on-Write Works

A copy-on-write type of snapshot is a detailed account of data as it existed at a certain
moment. Unlike a mirror, a copy-on-write is not really a copy of the data, but a specialized
record of it.
The copy-on-write process works as follows: when a snapshot is required, any unfinished
transactions or changes to the source data are allowed to complete, but new changes are
temporarily stalled. The source is momentarily idled (made quiescent). Once the
copy-on-write is activated, new transactions or changes (writes) to the source data are
allowed to take place. However, the copy-on-write process briefly intercepts or holds the
first write request that is issued for any particular block of data. While holding those
requests, it copies to cache the blocks that will be affected by those writes, and keeps a
record of the cached blocks. In other words, it reads each source block that is about to
change for the first time, copies the blocks current data to cache, and records the location
and identity of the cached blocks. Then the intercepted writes are allowed to take place in
the source blocks. (See figure Copy-on-write process on page 250.)

Appendix B, Processing Background 249

How Copy-on-Write Works

Copy-on-write process

Source data
1 Image of source data is
frozen; copy-on-write is
s0 s1 s2 s3 s4 s5 s6 s7 s8 s9 s10 activated

Writes delayed
2 New write requests to s4,
s7, s8 are held by
copy-on-write process
(see arrows)
s0 s1 s2 s3 s4 s5 s6 s7 s8 s9 s10

Result of copy-on-write
process

3 Copy-on-write process writes

cache contents of blocks s4, s7, and s8 to
c0 c1 c2 c3 c4 cache. These blocks are written to
cache only once, no matter how
many times they are changed in the
source during the snapshot.
Record of cached blocks:
4 Copy-on-write process
s4 = c0
keeps a record of the
s7 = c1 writes to cache
s8 = c2

Source data, modified after 5 Write requests are now

copy-on-write process allowed to take place

s0 s1 s2 s3 s4 s5 s6 s7 s8 s9 s10

The immediate results of the copy-on-write are the following: a cached copy of those
portions of the source that were about to change at a certain moment (see step 3 above),
and a record of where those cached portions (blocks) are stored ( 4 ).
The copy-on-write does not produce a copy of the source; it creates cached copies of the
blocks that have changed and a record of their location. The backup process refers to the
source data or cached data as directed by the copy-on-write process (see next diagram).

250 NetBackup Advanced Client System Administrators Guide

 How Copy-on-Write Works

Backing up a copy-on-write snapshot

Source Backup of copy-on-write

Source data, modified after
snapshot:
copy-on-write process
1 Backup reads
source data from
s0 s1 s2 s3 s4 s5 s6 s7 s8 s9 s10 s0, s1, s2, s3

Cache
2 At s4, copy-on-write
tells backup to read
c0 instead of s4
c0 c1 c2 c3 c4

3 Next, the backup

reads s5 and s6
from the source.
s0 s1 s2 s3 s4 s5 s6 s7 s8 s9 s10

Cache
4 At s7 and s8,
copy-on-write tells
c0 c1 c2 c3 c4 backup to read c1,
c2 instead of s7, s8.

5 Backup continues
reading source or
cache, as directed
by copy-on-write.

6 When backup
completes, backup
data is identical to
original source.

Backup image
s0 s1 s2 s3 c0 s5 s6 c1 c2 s9 s10

As shown in the above diagram, an accurate backup image is obtained by combining the
unchanged portions of the data with the cache. When a backup of the snapshot begins, the
backup application copies the source data 1 until it comes to a block that changed after
the copy-on-write process was activated. The copy-on-write tells the backup to skip that

Appendix B, Processing Background 251

How Copy-on-Write Works

changed block and read in its place the cached (original) copy 2 . The backup application
continues copying source data 3 until it comes to another changed block. Cache is read
again 4 as the copy-on-write process dictates. The backup, when finished, is an exact
copy of the source as it existed the moment the copy-on-write was activated.

252 NetBackup Advanced Client System Administrators Guide

Managing nbu_snap (Solaris) C

This appendix describes commands for managing backups that use the nbu_snap method.

Note This appendix applies to Solaris systems only.

253

Cache for nbu_snap

Cache for nbu_snap

The nbu_snap method is used by NetBackup to create a copy-on-write snapshot in the
following circumstances (Solaris systems only):
If the nbu_snap method is configured for the back up on the Snapshot Options dialog.
If the nbu_snap method was chosen by NetBackups auto-selection mechanism when
the backup was initiated.
If the client is in a FlashBackup policy and no snapshot method is configured
(Perform snapshot backups on the Policy display is not selected). This is called
FlashBackup default mode, and will be discontinued in a future release.
In all cases, a raw partition must be specified as cache for the snapshot. The same cache
can be used by any number of concurrent nbu_snap backups, as long as the cache is large
enough to hold copies of all file system blocks that were changed by user activity while
the snapshots were active.
In the Administration Console, the raw partition cache can be specified in any of three
places (see How to Enter the Cache on page 136).

Determining Cache Size

The size needed for the cache partition depends on the amount of user write activity that
occurs while the snapshot is active, not on the size of the file system. The more user
activity, the larger the cache must be. You can use the snaplist and snapcachelist
commands described in this chapter to determine the size of the cache required for that
activity. For a procedure, see Sizing the Cache Partition on page 135.
If a cache is too small and overflows, all snapshots using the cache become unreadable
and the backups that are reading the snapshots will fail.

Terminating nbu_snap
NetBackup ordinarily removes snapshots after the Advanced Client backup completes.
However, as a result of certain kinds of system failures, such as a system crash or
abnormal termination of the backup, the snapshot may not have been removed.
Use the snapoff command to terminate an nbu_snap snapshot that was not terminated
by the backup job (see snapoff on page 258). For more information on terminating
snapshots, see Removing a Snapshot on page 206.

254 NetBackup Advanced Client System Administrators Guide

 nbu_snap Commands

nbu_snap Commands
The following commands relate to the nbu_snap snapshot method.

snapon
snapon starts an nbu_snap snapshot (copy-on-write).
Execute this command as root:
/usr/openv/netbackup/bin/driver/snapon snapshot_source cache
where snapshot_source is the partition on which the clients file system (the file system to
be snapped) is mounted, and cache is the raw partition to be used as copy-on-write
cache.
Example 1:
/usr/openv/netbackup/bin/driver/snapon /var /dev/rdsk/c2t0d0s3

Example 2:
/usr/openv/netbackup/bin/driver/snapon /dev/vx/rdsk/omo/tcp1

/dev/vx/rdsk/omo/sncache

Note The snapshot is created on disk, and remains active until it is removed with the
snapoff command or the system is rebooted.

snaplist
snaplist shows the amount of client write activity that occurred during an nbu_snap
snapshot. Information is displayed for all snapshots that are currently active.
Execute this command as root:
/usr/openv/netbackup/bin/driver/snaplist

Information is displayed in the following form:

id ident size cached minblk err time

9 6730 2097152 2560 0 0 05/29/03 07:52:18

device = /dev/vx/rdsk/omaha/tcp1
cache = /dev/vx/rdsk/omaha/sncache
16 7857 2097152 128 0 0 05/29/03 12:16:00
device = /dev/vx/rdsk/omaha/vol01
cache = /dev/vx/rdsk/omaha/sncache
17 7908 20971520 4224 0 0 05/29/03 12:17:38
device = /dev/vx/rdsk/omaha/vol03
cache = /dev/vx/rdsk/zetab/cache

Appendix C, Managing nbu_snap (Solaris) 255

nbu_snap Commands

If no snapshots are currently active, no output is returned.

Description:
id
The snapshot ID. This is used by the snapoff command to terminate the snapshot.
ident
A unique numeric identifier of the snapshot (this is the pid of the process that created
the snapshot).
size
The size of the clients snapshot source in 512-byte blocks. This is the partition on
which the clients file system (the file system being backed up) is mounted. Note: this
size value is not a reliable guide as to the size of cache needed for the snapshot; the
user write activity during the snapshot is what determines the size of the cache
needed. See the cached column of this output.
cached
The number of 512-byte blocks in the client file system that were changed by user
activity while the snapshot was active. Prior to being changed, these blocks were
copied to the cache partition. The more blocks cached as a result of user activity, the
larger the cache partition required. However, additional overheadnot shown in this
cached valueis required in the cache.

Note To see the total space used in a particular cache partition, use the snapcachelist
command.

minblk
In the partition on which the file system is mounted, minblk shows the lowest
numbered block that is currently being monitored for write activity while the
snapshot is active. minblk is used by FlashBackup policies only.
err
An error code; 0 indicates no error.
If a snapshot has encountered an error, the err will be non-zero and the snapshot will
be inaccessible. It can be terminated using snapoff and the snapshot ID. Error codes
are identified in /usr/include/sys/errno.h. Also, error messages may be found
in /var/adm/messages.
time
The time at which the snapshot was started.
device

256 NetBackup Advanced Client System Administrators Guide

 nbu_snap Commands

The raw partition containing the clients file system data to back up (snapshot source).
cache
The raw partition used as cache by the copy-on-write snapshot process.

Note Make sure this partition is large enough to store all the blocks likely to be changed
by user activity during the backup. To determine the total space used in a particular
cache partition by all active snapshots, use the snapcachelist command.

snapcachelist
snapcachelist displays information about all partitions currently in use as nbu_snap
caches. This command shows the extent to which the caches are full.
Execute this command as root:
/usr/openv/netbackup/bin/driver/snapcachelist

If no snapshots are currently active, no output is returned.

Output is of the form:
device free busy

/dev/rdsk/c0t4d0s0 238528 264472

Description:
device
The raw partition being used as cache.
free
The number of 512-byte blocks unused in the cache partition.
busy
The number of 512-byte blocks in the client data that changed while the snapshot was
active. Prior to being changed, these blocks were copied to the cache partition by the
nbu_snap copy-on-write process. For each cache device listed, busy shows the total
space that was used in the cache. You can use this value as a sizing guide when setting
up raw partitions for nbu_snap cache.
When a cache is full, any additional change to the client data will cause the
copy-on-write to fail and the snapshot will no longer be readable or writable. Reads or
writes to the client data will continue (that is, user activity will be unaffected). The
failed snapshot, however, will not be terminated automatically and must be
terminated using snapoff.

Appendix C, Managing nbu_snap (Solaris) 257

nbu_snap Commands

Note snaplist and snapcachelist can also be used to monitor an nbu_snap

snapshot that was started by a NetBackup policy. Once the backup completes,
NetBackup removes the snapshot. As a result, the snaplist and snapcachelist
commands no longer return any output.

snapstat
snapstat displays diagnostic information about the snap driver.
Execute this command as root:
/usr/openv/netbackup/bin/driver/snapstat

snapoff
snapoff terminates an nbu_snap snapshot.
Execute this command as root:
/usr/openv/netbackup/bin/driver/snapoff snap1 ... snapn
where snapx is the numeric id of each copy-on-write process to be terminated. (Use
snaplist to display the id of active snapshots.)
If snapoff is successful, a message of the following form will be displayed:
snapshot 1 disabled

snapshot 2 disabled

...

snapshot n disabled

If snapoff fails, an explanatory message is displayed. Error codes are identified in

/usr/include/sys/errno.h.

Caution Do not terminate a snapshot while the backup is active, because corruption of
the backup image may result.

258 NetBackup Advanced Client System Administrators Guide

Index

Numerics and split mirror 11

3pc.conf file 60, 75 and VVR 147

creating 75 configuring 96

description 60 disk config requirements 154

-x option 48 disk group naming 170

4.5 server introduction 10

administration from 5.1 console 24, 84 requirements 96

restriction 119

A
restrictions 96

a=wwpn
testing setup 144, 148

lun 62
alternate world-wide name 63

abnormal termination 206, 254

Any_available storage unit 87, 100, 203

access conflicts (arrays) 173, 174, 175

APP_NBU_VVR 147

activate block-level restore 185

arbitrated loop 27

Active Directory 102

archives 182

and FlashBackup 102

array

active third-party copy device 47

multi-ported 17
actual device file name 105
Array Integration Option (see Advanced
address (in 3pc.conf) 62
Client)
address mapping mode 53
arrays
administration
access conflicts 173
of 4.5 server from 5.1 24, 84
and VxVM 168
Administration Console
as third-party copy devices 79
for 4.5 server 24, 84
create config files 160
Advanced Client
list of supported 152
access web info xix
restrictions 168
installing 34
view state of 163
products included in 2
associating primary to mirror 156
Advanced client 4

auto snapshot selection 89, 91, 98, 123

Advanced Snapshot Options button 89

automatic backup 182

ALL_LOCAL_DRIVES entry 24, 87, 100,

105, 125, 178 B

Allow multiple data streams 100, 175 background on processing 219
alternate client backup
defined 26 agent 26
notes on 96 automatic 182
alternate client backup 3, 9, 104 image 246
and disk arrays 155 local 7, 223
and FlashSnap 144 logs 200, 202

259
 manual 182
bppfi log on client 200

of copy-on-write 251
bpps 35

offhost
bprd 35

configuration 86
bprd log 201

prerequisites 178
bprdreq 35

SAN configuration for 43

bprestore log 201

raw partition 178

bpSALinfo

SCSI offhost 16
address info in 3pc.conf 62

techniques (overview) 4
bpsched 227, 229, 231, 233, 235, 237

types supported 104

bptm 228, 229, 230, 231, 233, 235, 246, 248

user-directed 182
bptm log 69, 200, 201

backup agent 2
bptpcinfo command 48, 49, 55, 75, 160

Backup Policy Configuration wizard 87,

BusinessCopy method 91, 151, 182, 203

100, 125

C
backup scripts 95

c=client 62

Backup Selections list 87

cache

ALL_LOCAL_DRIVES entry 24, 87, 100,

diagram of 250, 251

105, 125, 178

flushing for quiesce 222

block vs. character device 178

overflow 254

data types supported 21

partition 108, 246

directives 108

partition out of space 205

FlashBackup 105

requirements 134

symbolic link 99

size 254

Backup, Archive, Restore 187

specifying raw partition for 134

BCV 153

cache object 127, 141

best practices 172

CACHE= directive 106, 108, 137

BLIB 84

Change Policy dialog 90, 95

block device file (vs character) 134, 178

Chaparral 68

block level incremental backup 84

character device file (vs block) 134, 178

block level incrementals 246

checking logs 200

block-level restore 4, 185

checkpoint

how to activate 185

removing (VxFS) 208

restriction 119, 185

VxFS file system 138

bp.conf file 200

client data

bpbkar 226, 228, 229, 230, 231, 233, 235, 240,

prerequisites for offhost backup 178

243

client list 105

process 205

client name (in 3pc.conf) 62

bpbkar log 200

client option (on bptpcinfo) 48, 75

bpbrm log 201

client path (in 3pc.conf) 62

bpbrm log on server 200

clients

bpcd 227, 229, 231, 237

installing software 35

bpdbm 248

clone

bpdgclone command 212

removing 208, 211

bpfis 213, 236, 237, 244

VxVM disk group 170, 211

examples 215

clustering 35

optional parameters 215

Common Serial Number Mode 156

bpfis command 193, 195, 206

concurrent access conflicts (arrays) 173, 175

bpfis log 200

configuration

bpmoverinfo command 77

260 NetBackup Advanced Client System Administrators Guide

 auto method 89
directives (Backup Selections list) 108
backup selections 105
directories for troubleshooting 200
client list 105
disk
files 77
access conflicts 173
FlashBackup 103
SCSI vs. IDE 203
prerequisites 84
visibility 9, 203
procedure 85, 90, 122
disk array
robot in Media Manager 113
as third party copy device 79
supported data types 102
choose snapshot method 151
copy manager (see third-party copy) create config files 160
copy-on-write supported 152
cache 250 disk group 168
compared to mirror 5 VxVM 172
defined 26 disk group clone 170, 211
for Storage Checkpoint 120 DISK keyword 70
how backed up 251 disk layout for Storage Checkpoint 118
overview 5, 249 disk restore
Core Frozen Image Services (see Advanced caution 172, 195
Client) disk snapshot
cp command 193 restoring from 193
cross mount points (disabled) 100 disk storage units 87, 178
Crossroads 68 DLL patch for VxFS 24
customer support 203 DMP (Dynamic Multipathing) 156, 171, 173
dr keyword 72
D
data consistency (quiesce) 220
E
data mover (see third-party copy)
E4 target (see identification descriptor)
Database Edition for Oracle 24
EMC Symmetrix disk groups
database files (Windows) 102
creating 157
data-property-name entry 57
END keyword 71
DB2 118
exceptions to exclude list 102
DCO plexes 144
exclude list 102, 105
deinstallation
Extended Copy command 3, 9, 30, 60
clients 41, 42 Extended Frozen Image Services (see
server 40 Advanced Client)
deleting a clone 211 extent 229, 231, 233, 235
deport disk group 144 defined 27
deport snapshot 194, 196
F
deporting disk group 145
failover 35

device
fast mirror resynch 142

ID 62

FastResync 29, 124, 144

path 62

FastResync (defined) 27

recognition 54

feature/requirements table 18

serialization 25, 178, 203

features of FlashBackup 102

Device Configuration wizard 84

fibre channel

device file name 105

defined 27
device file name (FlashBackup) 105
types supported 27
directives
file pathname (max length) 88, 99
for scripts 95
file promotion 4, 185

Index 261
 restore-process diagram 243, 244
hr keyword 70, 72

file system map 246

file systems

I/O components
defined 27

supported 102

quiescing 222

i=iddesc

files list (see Backup Selections list)

i=reserve value 71

FIXED address mapping mode 53

IBC messages 147

FlashBackup 2, 254

IBC receive timeout (seconds) 94

actual device name 105

IBC send timeout (seconds) 94

and VxFS_Checkpoint 138

IDE vs. SCSI 203

backup selections 105

identification descriptor 64

device file name 105

importing disk group 145, 146

features 102

include list 105

files to back up 105

individual file restore from raw 103

restoring 183

inetd 162

FlashSnap 24

Inline Tape Copies (Vault) 25

deporting the disk 144

inquiry page code 83 25, 178, 203

FlashSnap method 11, 91

insf command 54

preparing for alternate client backup 144

install script 34

restoring from disk image 194, 196

installation

flush file system 220, 222

clients 35

format command 168

deinstallation 40

fsck (after raw partition restore of vxfs

directories 34

system) 184

of Advanced Client 34

fsckptadm 30

reinstallation 35

fsckptadm command 30, 120, 208

upgrade (patch) 35

full-sized instant snapshot 92

Installation Options menu 34

full-sized instant snapshots 128

installing Advanced Client 34

G instance number 163

get_license_key 34
Instant Recovery

GID 96
defined 27

glossary of terms 26
restore-process diagram 240, 243, 244

Glossary. See NetBackup Help.

selecting in policy 123

group of storage units 87

instant recovery 4

instant snapshot 127, 141

H
instant snapshots 92, 128

hardware-level disk restore 172, 195

ioscan command 49, 54

hardware-level restore 195

HBA drivers 59
J
Hitachi 68, 91
JBOD array 54

HOMRCF 91
jbpSA 187

HORCM_CMD 160
jnbSA 85, 90, 95, 103, 122

HORCM_DEV 160

K
HORCM_INST 161

Keep snapshot after backup 93, 206

HORCM_MON 159

restoring from image 193

HORCMINST 163

kernel messages 201

horcmstart command 163

keywords in mover.conf 70, 71

HP 68, 108

8 devices per target 54

262 NetBackup Advanced Client System Administrators Guide

 L fast resynch 142
l=lun 63
overview 6
label option of format command 168
mirror (VxVM snapshot) 22, 29, 141
labeling secondary disks 168
mirror association 156
left over snapshot 206, 254
mirrors
license keys, installing 34
preparing 126
Limit jobs per policy 100
mklogdir.bat 201
limitations (also see Release Notes) 24
mover.conf file 60, 65, 73, 77
links (in Backup Selections list) 99
72
Local Host
creating 77
network configuration for 223 DISK 70
Local Host backup method dr keyword 72
network configuration for 7 END 71
lock file system 220, 222 examples 78
log (bpbrm) 200 how determine entries for 65
log (bppfi) 200 hr keyword 70, 72
logging i=reserve value 71
directories to create 200 order of searching 74
logical volume (as raw partition) 134 policy_name 79
logs 200, 201, 202 TAPE 71
creating for Windows 201 multi-device system (VxFS) 138
for SCSI reserve/release 69 Multiple Copies (Vault) 25
loop (Fibre Channel) 27 multiple data streams 100, 175
ls command 30, 120 configuring 100
LUN 55, 63 multiplexing 25, 178
LVM 102 multi-ported disk array 17
M N
managing snapshots 213
name entry in st.conf 57
manual backup 182
NAS
manuals xv
policy configuration 111
mapping 9
NAS_Snapshot method 4, 91, 123, 149, 184
defined 27 NBU_CACHE 127
information 229, 231, 233, 235 nbu_snap method 91, 134, 246, 254
Maximum jobs per client 100 NDMP 4
Maximum multiplexing per drive 100 access web info xx
maximum pathname length 88, 99 NDMP snapshot 149
Maximum Snapshots (Instant Recovery NDMP V4 123
only) 93, 114, 129 NetBackup for NDMP
MDS 138 configuration on UNIX 111
Media multiplexing 100 NetBackup Media Server
media server (see NetBackup Media Server) and storage units 87, 178
messages file 201 network diagram of 14, 46, 224
method process overview 18, 228
selecting offhost backup 86 selecting 86
selecting snapshot 90 NetBackup Media Server method
mirror 6 defined 9, 28
compared to copy-on-write 7 Network Attached Storage 9, 86
defined 28 NEW_STREAM directive 108

Index 263
 no-data Storage Checkpoint 119 point-in-time snapshots 4
policy

O
how to select type of 86

Offhost and SAN Data Movement Services

storage unit 87, 100

(see Advanced Client)

Policy dialog 85, 103

offhost backup 86, 246

policy_name (on mover.conf file) 73, 77, 79,

and multiplexing 178

203

NAS 9

poll 160

overview 9

port name

prerequisites for 178

using SANPoint Control 55

raw partition 178

primary to mirror association 156

SCSI connections 16

primary vs. alternate client

type of disk (SCSI vs. IDE) 203

processing background 219

without SAN 16

promotion, file 4, 185

online_util log 200

operating system changes 57

Q
operating system patches 24
query snapshot 193, 195, 206

Oracle 24, 220

quiesce 220, 249

overwriting

R
raw partition restores 183

RAID 5 141, 174

P RAID Manager daemons 163

P=clientpath 62
raw partition 105, 108

p=devpath 62
as snapshot source 99

page code 83 25, 178, 203

block vs. character device 178

page code inquiry 62, 64

defined 28

pairdisplay listing 164

not supported with

pairresync command 209

VxFS_Checkpoint 138

partitions
restore 183

Windows 102
specifying for cache 134

partner host 161

raw partition restore
partner service 161
fsck needed after vxfs restore 184

passive third-party copy device 46

recognizing devices 54

passthru device paths 49, 54, 77

recovery procedure 206, 254

patch for VxFS 24

Registry 102

patches required 24
and FlashBackup 102

pathname length 88, 99

rem_drv sg command 55

Perform block level incremental backups 84

remote snapshot (see alternate client

Perform snapshot backups 254

backup)
re FlashBackup 245
removing

performance
clone 211

increasing tape 204

snapshots 206, 254

peripherals (latest info on web) xix, xx

replicated host 13

Persistent Frozen Image (see Instant

replication

Recovery) testing setup for alternate client

PFI_BLI_RESTORE file (for block-level backup 148

restore) 185
replication (for alternate client backup) 147

physical device (as raw partition) 134

requirements for NetBackup 24

platform requirements 24
reserve/release

platforms supported 33
dr keyword 72

264 NetBackup Advanced Client System Administrators Guide

 hr keyword 70
sd.conf 53

introduction 69
second world-wide name 63

third-party reservation 69, 70, 71

serial numbers 25, 62, 178, 203

restore 183 serialization 25, 178, 203

and fsck 184 sfr program 248

arrays 195 sg.conf 54, 57

block-level 4 sg.install script 55

block-level restore 185 sg.links 54, 57

file promotion 4, 185 sgscan command 49, 54, 68

from disk image 193 Shadow Copy Service (see VSS)

from FlashBackup 183 ShadowImage method 91, 151, 182, 203

hardware-level 172, 195 shared disks 9

logs 201, 202 SIGHUP 162

Oracle files 185 single file restore (sfr) 248

overwrite option 183 SIZE_DATA_BUFFERS 204

raw partitions 183 snap

re. device file 183 removing 207

restrictions 102
snapcachelist command 257
restrictions (also see Release Notes) 24
snapctl 102
resyncfromreplica option 195, 197
driver log 201
resynchronization of mirror 142
overview 254
resynchronize
snapctl driver 106
at end of backup 172 snaplist command 207, 255
disks 172, 208 snapoff command 207, 258
Resynchronize mirror in background 94 snapon command 255
RMAN 28 snapshot
rollback auto selection 89, 91, 98, 123

restore-process diagram 243, 244 back up to local storage 7

root, specifying as snapshot source 100 backup to local storage 223

RSM Database 102 configuration 90, 122

copy-on-write vs mirror (how to

S
choose) 5

s=sn 62
defined 29

SAN 25, 178

ID 193, 195, 206

defined 28
methods

not required for offhost backup 16

matching to array 152
SANPoint Control (SPC) 45, 55
mirror (creating) 144, 145
scripts
mirror (VxVM) 22, 29, 141
running before/after backup 95
mirror, defined 6
SCSI Extended Copy command 60
on Solaris client, troubleshooting 205
SCSI Inquiry Page Code 83 25, 178, 203
overview 4
SCSI offhost backup 16
pre-processing for 220
SCSI reserve/release 69, 70
removing 206, 254
dr keyword 72
restoring from disk 193, 195
hr keyword 70, 72
selecting method of 90
i=reserve value 71
source
SCSI serialization 25, 178, 203
defined 29
SCSI target 55
for symbolic link 99
SCSI vs. IDE 203
types supported 21

Index 265
 volume (creating) 144, 146
synchronize disks 208

VxVM instant 92
Synchronize mirror before backup 94

Snapshot Options dialog 89, 90, 98

system requirements for NetBackup 24

Snapshot Volume
System State 102

defined 29
and FlashBackup 102

snapshots 4
system-protected files and FlashBackup 102

snapstat command 258

T
software required (table) 18

tape
software upgrade 35

increasing performance 204

Solaris

TAPE keyword 71

version requirements 24

tape reserve 69

space-optimized instant snapshots 92

tape-config-list 57

space-optimized snapshot 127, 141

target (HP limit of 8 devices) 54

SPC 55

technical overview 245

Spectra Logic 68

Terminal Services Database 102

split mirror backup (alternate client) 11

terminate bprd 35

st.conf 53, 57

termination 206, 254

stack quiescing 222

terminology 26

standard (primary) disk 153

third-mirror (vxvm) 141

Standard policy

third-party copy 9

restoring 183

and multiplexing 25, 178

Storage Checkpoint 123

and storage unit 87, 100

basic features of 120

configuration files for 77

defined 30

defined 26

determining disk usage 30, 120

device configuration 60, 86

disk layout 118

network diagram of 15, 47, 225

storage checkpoint 92

process overview 232

Storage Checkpoints 24

third-party copy device

storage checkpoints 102

active 47

storage devices 84

and storage units 87, 178

storage unit 73, 77, 87, 100

defined 30

restrictions for data movers 87, 178

FIXED mode 53

storage_unit_name version of mover.conf

how determine passthru path for 65, 68

file 80, 203

not on same network as media server 48,

streams, allow multiple data 108

75

Sun disk array 68

passive 46

support web site (VERITAS) xix, xx

SCSI reserve 69

supported

Third-Party Copy Device offhost

data types 102

backup 104

supported platforms 33

third-party reservation 70, 71

switched fabric 27

TimeFinder method 91, 151, 182, 203

SYMAPI 91

timeout setting (for mover) 72

symbcv command (EMC) 157

to keyword 72

symbolic links 99

troubleshooting 200

symdg command (EMC) 157

directories to create 200

symld command (EMC) 157

type

Symmetrix disk groups

selecting for policy 86

creating 157

types of backups supported 104

symmir command 158, 172, 209

266 NetBackup Advanced Client System Administrators Guide

 U restoring 184

UFS file system 30, 91, 134

VxFS multi-device file system 138

UID 96
VxFS_Checkpoint method 92, 102, 123, 138,

umount clone 208

185, 186

umount command 194, 208

VxFS_Snapshot method 92, 106, 140, 246

UNC for Windows pathnames 112, 114

vxibc command 148

uninstalling NetBackup 40
vxmake 128

Universal Naming Convention (UNC) 112,

VxMS 27, 30

114
vxprint command 207

unmounting snapshot 194

vxrecover command 195

unquiesce 221
vxsnap 129

UNSET 110
VxVM 174

UNSET directive 108

and RAID 5 141

UNSET_ALL 110
clone of disk group 170, 211

UNSET_ALL directive 109

configuration for arrays 168

upgrade software 35
disk group 168, 172

Use alternate client 86

instant snapshots 92, 128

user-directed
mirror 22, 141

archives 182
required version 23, 91, 97, 141

backup 182
restrictions re arrays 168

Volume Manager 22, 24, 141

volume name restriction

vendors latest info on) xx

Instant Recovery
verbose mode (bptpcinfo) 75

VERBOSE setting for logs 200

volume name restriction 119,

VERITAS Federated Mapping Services 30

126, 141

VERITAS Volume Manager 92, 141

volumes (arrays supported) 152

VERITAS Volume Replication 124

vxvm method 92, 119, 124, 141, 169, 175

vfm.conf file 215

VxVM mirrors

volume, defined 30
preparing for Instant Recovery 126

volumes
vxvol command 141, 144

multi-device 138

W
VSP method

W=wwpn 63

description 91

w=wwpn 63

using with/without Advanced Client 91

web access to product info xx

VSS method 92

web access to recent Advanced Client

VVR 22, 24, 119, 124

info xix

VVR method 13, 92, 148

Windows

preparing for alternate client backup 147

OS partitions 102

vxassist 128

System database files 102

vxassist command 141, 144, 145, 207

wizard

vxdg command 145, 146, 194, 196

Backup Policy Configuration, use of 100

vxdg list command 211

world-wide port name 62, 63

vxdisk command 194

VxFS clone, removing 208

X

VxFS file system 24, 91, 102, 134

-x option (bptpcinfo) 48, 75

patch for library routines 24

Index 267