3PAR InForm® OS 2.3.

1 MU4
Release Notes

3PAR Inc.
4209 Technology Dr.
Fremont, CA 94538 USA
QL226-96005
May 2011

3PAR Confidential
Revision Notice
This is the first release of this document.

Changes
The material in this document is for information only and is subject to change without notice. While reasonable efforts have been
made in the preparation of this document to assure its accuracy, 3PAR Inc. assumes no liability resulting from errors or omissions in
this document or from the use of the information contained herein.
3PAR reserves the right to make changes in the product design without reservation and without notification to its users.

Updates to the Documentation at 3PAR Central

Any updates to this document, or to other 3PAR technical documents, can be found by logging on to 3PAR Central’s Document
Control System from 3PAR’s Support page at http://www.3par.com/support/index.html

3PAR Technical Support and Services

Contact your local service provider for technical support and services.

Sales and Ordering Information

For sales and ordering information, contact:
3PAR Inc.
4209 Technology Drive
Fremont, CA 94538 USA
Telephone: 510-413-5999
Fax: 510-413-5699
Email: [email protected]

Reader Comments and Suggestions

Please email your comments and suggestions about this document to [email protected].

Copyrights
© 2011 3PAR Inc. All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted in any
form or by any means, electronic, mechanical, photocopying, recording or otherwise, without the prior written consent of 3PAR Inc.,
4209 Technology Drive, Fremont, CA 94538. By way of exception to the foregoing, the user may print one copy of electronic material
for personal use only.

Trademarks
3PAR, InServ, InForm, InSpire and Serving Information are registered trademarks of 3PAR Inc.
Intel and Pentium are registered trademarks of Intel Corporation.
Linux is a registered trademark of Linus Torvalds.
Microsoft, Windows, and Windows NT are either registered trademarks or trademarks of Microsoft Corporation.
Oracle is a registered trademark of Oracle Corporation.
Sun, Solaris, and Java are trademarks of Sun Microsystems, Inc.
UNIX is a registered trademark of The Open Group.
All other trademarks and registered trademarks are owned by their respective owners.

3PAR Confidential
InForm OS Version 2.3.1 MU4 3PAR InForm OS Release Notes

Table of Contents
1 Introduction
1.1 Audience 1.1

1.2 Organization 1.2

1.3 Related Documentation 1.2

2 3PAR InForm Operating System Release Notes

2.1 Online Upgrade Considerations 2.1

2.2 Supported Platforms 2.3

2.3 Components 2.3

2.4 What’s New in this Release 2.3

2.5 Modifications to the InForm OS 2.4

2.6 Limitations of the InForm OS 2.14

2.6.1 Connectivity Issues with Remote Copy IP Interfaces for 100Mbps Switch Ports 2.14

2.7 Known Issues with the InForm OS 2.15

2.7.1 RPI Resources Exceeded for Fibre Channel Ports (Bug 61125) 2.15
2.7.2 Internally Generated Host ID Returned by Host Explorer (Bug 62212) 2.15
2.7.3 CIM Server Error Passing Maximum Length Virtual Volume Name
(Bug 55445) 2.15
2.7.4 Physical Copy Issue with Larger Destination Volume Size (Bug 55681) 2.16
2.7.5 4Gb/s 2-Port adapter Issue 2.16

Table of Contents iii

3PAR Confidential
 3PAR InForm OS Release Notes InForm OS Version 2.3.1 MU4

2.7.6 Downgrade Issues 2.16

2.7.7 Remote Copy and Recovery Manager Snapshot Interaction Issues 2.18
2.7.8 Firmware Core Dump from FC Adapter by LSI
(Bugs 24178, 30353, 32351) 2.19
2.7.9 Windows 2003 Pop-Up 2.19
2.7.10 Cage Loop Down Alert is not Auto-Recovered (Bug 40274) 2.19
2.7.11 System Manager Could Crash During Rolling Upgrade if Remote Copy
Groups Running (Bug 40867) 2.19
2.7.12 No Explicit Way to Reset the Default Value for Tunelinks (Bug 43579) 2.20
2.7.13 RCFC Port Shutdown Failure (Bug 44520) 2.20
2.7.14 Remote Copy Volumes go Stale Due to Lack of Adm Space (Bug 44963) 2.20
2.7.15 controlport Configuration Gives Warning after Being Taken Offline
(Bug 44989) 2.20
2.7.16 createvvcopy Issues with Halt (Bug 45142) 2.20
2.7.17 Disk Failure Alert has Invalid/Incorrect Magazine Information
(Bug 45224) 2.21
2.7.18 Some Messages Not Clear During upgradecage
(Bug 45229) 2.21
2.7.19 showtask for updatevv Displays Incorrect Completion Time
(Bug 46328) 2.21
2.7.20 Tasks Become Active after a Failed Upgrade (Bug 46418) 2.21
2.7.21 Remote Copy: Possibility of Secondary Volumes going out of Sync
with Primary Volumes with fail_wrt_on_err Policy (Bug 46443) 2.22
2.7.22 CIM Server Core after Upgrade from 2.2.3.116 (MU1) to 2.2.4.160
(MU6) (Bug 50476) 2.22
2.7.23 Remote Copy: Long Waits for Remote Copy Commands to Return Data 2.22

3 3PAR InForm CLI Release Notes

3.1 Installation Notes for the InForm CLI 3.1
3.1.1 Virus Checking 3.1

3.2 Known Issues with the InForm CLI 3.2

iv Table of Contents
3PAR Confidential
InForm OS Version 2.3.1 MU4 3PAR InForm OS Release Notes

1
Introduction

In this chapter
1.1 Audience 1.1

1.2 Organization 1.2

1.3 Related Documentation 1.2

This document provides the release notes for the 3PAR InForm OS (Operating System) and
Command Line Interface (CLI).

You should read this document prior to installing the 2.3.1 Maintenance Update 4 (MU4)
version of this product.

NOTE: These release notes describe the features and bug fixes included in
InForm OS 2.3.1 MU4. Refer to the appropriate release notes for all prior releases.

1.1 Audience
The information in these release notes is intended for use by 3PAR customers, partners, and
3PAR field representatives.

Introduction 1.1
3PAR Confidential
 3PAR InForm OS Release Notes InForm OS Version 2.3.1 MU4

1.2 Organization
This guide is organized as follows:

◆ Chapter 1, Introduction, this chapter.

◆ Chapter 2, 3PAR InForm Operating System Release Notes, provides the InForm
Operating System release notes.

◆ Chapter 3, 3PAR InForm CLI Release Notes, provides the InForm CLI release notes.

1.3 Related Documentation

The following documents also provide information related to InServ storage servers and the
InForm Operating System:

For information about… Read the…

Using the InForm Command Line 3PAR InForm OS CLI Administrator’s Manual
Interface (CLI) to configure and and InForm OS Command Line Interface
administer InServ Storage Servers Reference

Using the InForm Management Console 3PAR InForm Management Console Online
graphical user interface to configure and Help
administer InServ Storage Servers

InServ Storage Server concepts and 3PAR InForm OS Concepts Guide

terminology

Using Remote Copy 3PAR Remote Copy User’s Guide

Determining InServ Storage Server 3PAR InServ E-Class/F-Class Storage Server

hardware specifications, installation Physical Planning Manual
considerations, power requirements,
networking options, and cabling 3PAR InServ S-Class/T-Class Storage Server
Physical Planning Manual

1.2 Organization

3PAR Confidential
InForm OS Version 2.3.1 MU4 3PAR InForm OS Release Notes

2
3PAR InForm Operating
System Release Notes

In this chapter
2.1 Online Upgrade Considerations 2.1

2.2 Supported Platforms 2.3

2.3 Components 2.3

2.4 What’s New in this Release 2.3

2.5 Modifications to the InForm OS 2.4

2.6 Limitations of the InForm OS 2.14

2.7 Known Issues with the InForm OS 2.15

2.1 Online Upgrade Considerations

This feature allows the upgrade of the InForm OS concurrently with I/O activity on the
attached hosts, provided certain conditions are met. For further information and pre-planning
Rolling Upgrades, refer to the latest version of the InForm OS Upgrade Pre-Planning Guide,
320-200158.

3PAR InForm Operating System Release Notes 2.1

3PAR Confidential
 3PAR InForm OS Release Notes InForm OS Version 2.3.1 MU4

NOTE: Installing this Maintenance Update (MU) will likely disconnect the
management interfaces, thereby ending the CLI and InForm Management
Console (IMC) sessions and possibly generating alerts. This is expected behavior. If
the management interfaces are disconnected, reconnect and start new CLI or IMC
sessions.

2.2 Online Upgrade Considerations

3PAR Confidential
InForm OS Version 2.3.1 MU4 3PAR InForm OS Release Notes

2.2 Supported Platforms

The 3PAR Configuration Matrix for InForm OS 2.3.1 contains information about supported
hardware and software platforms. Contact your local sales engineer or service representative
to obtain a copy, or find it online at http://www.3par.com. Use your 3PAR login user ID to search
the Document Control System for the latest version of the Configuration Matrix.

If you are licensed for and using the Thin Persistence feature and you are
upgrading from InForm OS 2.3.1 GA to 2.3.1 MU4, you must install 2.3.1 P10
before installing 2.3.1 MU4.
If you are licensed for and using the Thin Persistence feature and you are
upgrading from InForm OS 2.3.1 MU1 to 2.3.1 MU4, you must install 2.3.1 MU1 P9
before installing 2.3.1 MU4.

2.3 Components

Component Version

OS 2.3.1.330 (MU4)

CLI 2.3.1.330

SNMP Agent 1.4.0

2.4 What’s New in this Release

This release includes support for new 100 and 200 GB Solid State Drives (SSD) on HP 3PAR T-
Class and F-Class Storage Systems.

Supported Platforms 2.3

3PAR Confidential
 3PAR InForm OS Release Notes InForm OS Version 2.3.1 MU4

2.5 Modifications to the InForm OS

The following items have been addressed in this release.

Bug ID Item Description

45951 SCSI-2 LUN When an initiator logged out or was disconnected

reservations not while holding a SCSI-2 reservation on an InServ
properly cleared on LUN, the reservation would remain associated with
Fibre Channel the LUN and had to be manually cleared for access
disconnects. by other initiators.

The SCSI-2 reservation on a LUN is now cleared

when the holder of the reservation logs out or is
disconnected from the array. The InServ also sets a
Power_on_Reset Unit Attention on the LUN
indicating that the reservation has been cleared for
initiators accessing the LUN.

46473 Node goes off-line due When a local host write was terminated with an
to a Remote Copy error while a Remote Copy request was scheduled
request that is not for retransmission, the Remote Copy request was
properly cancelled sometimes not removed from the retransmit queue
from the retransmit before it was canceled. This could result in the
queue following retransmit queue retaining stale entries which
termination of a local could cause controller nodes to go offline at a later
host write operation. stage. The most common situation for a local host
write to be terminated with an error was when a
different controller node would go offline.

Remote Copy requests are now properly handled

through the retransmit queue to account for the
termination of write operations by the local host.

2.4 Modifications to the InForm OS

3PAR Confidential
InForm OS Version 2.3.1 MU4 3PAR InForm OS Release Notes

46815 Using removehost - If the removehost -rvl command was used to

rvl to remove a WWN remove a WWN from a host when the WWN was
from a host can cause visible on multiple ports and a volume was
the next reference to exported to the host, the next reference to the list
the listing of VLUNS of VLUNs associated with that host, (e.g., issuing a
for that host to showvlun command), would cause System
generate a System Manager to fail.
Manager failure.
The listing for VLUNs is now properly updated
following the removal of a WWN when using
removehost with the -rvl option.

49261 Processing of requests Requests that required CIM Server to process large
that require an responses, (e.g., when processing an
extensive response by EnumerateInstances request), could result in a
CIM Server results in a failure condition that caused the process to hang or
process failure. crash.

Processing of a large response no longer causes CIM

Server to fail. However, it should be noted that if
CIM Server tries to generate a response that is too
large for it to process, the transaction with the
client can be terminated without generating an
exception.

50428 CIM Server process A response queue function is serialized per request
occasionally fails when by locking a mutex owner based on the
sending a chunked OperationAggregate object for that request.
response. However, if the response enqueued was the last
response, the OperationAggregate object was
deleted within the method. When the AutoMutex
that was created at the beginning of the method
went out of scope, it held onto a mutex that had
previously been deleted.

CIM Server now handles the responses correctly.

Modifications to the InForm OS 2.5

3PAR Confidential
 3PAR InForm OS Release Notes InForm OS Version 2.3.1 MU4

54071 The TPD_Storage The classname for an association between

Setting associated Current Volume StorageSetting and
to a StorageCapabilities was not previously
TPD_StorageVolume explicit.
was unnecessarily
associating to a StorageSetting that is associated to a
TPD_Storage TPD_StorageVolume has been renamed to
Capabilities in the TPD_VolumeSetting, and its association with
CIM API. TPD_StorageCapabilities has been removed.

54074 Instances of fixed Instances of fixed TPD_StorageSetting

TPD_Storage associated to the StorageCapabilities of a
Setting associated to DynamicStoragePool were not previously made
the Storage available for creation of a volume from that

Capabilities of a DynamicStoragePool.
Dynamic
New instances of fixed TPD_StorageSetting
StoragePool
that are associated to the StorageCapabilities
required for creation
of a DynamicStoragePool have been added so
of a volume from
that a client can use them to create a volume from
Dynamic
that DynamicStoragePool.
StoragePool
in the CIM API.

54726 CMP owner node is The CMP owner node was not properly handling
not properly handling the merging of credit pages when a new credit
the merging of credit dmask overlapped.
pages.
The IO write credit pages merge code path has
been modified, so that when the LDIO flusher
flushes dirty cache pages, it leaves the dirty mask in
the log entry for those credit pages in a consistent
state.

2.6 Modifications to the InForm OS

3PAR Confidential
InForm OS Version 2.3.1 MU4 3PAR InForm OS Release Notes

55017 Clearing of older In some cases, older message references were not
message references being cleared properly when cancelling inter-
not handled properly process messages.
when canceling inter-
process messages. Message references are now properly cleared.

56195 SAN boot on Cisco UCS When the firmware in the 2-port 4G fibre channel
M1 Blade Server brings HBA in the E and F Series systems received certain
port down. incorrectly formatted fibre channel frames, the
frame buffers were not released into the free
buffer pool. When a large number of these invalid
frames were received, the port could run out of
buffer credits causing the port to go off-line.

This situation has been remedied by correctly

releasing the invalid frames.

56754 Non-login iSCSI PDUs If an iSCSI port on an InServ received a non-login-

received during iSCSI related iSCSI PDU from an initiator before a login
login processing can command sequence was completed, the InForm OS
cause a controller kernel could fail due to the accessing of an un-
node to go offline due initialized pointer.
to the accessing of an
un-initialized pointer This has been fixed by processing the PDU correctly
when processing an and sending a Login reject as per the iSCSI protocol.
iSCSI PDU.

57370 Initialized value is An appropriate value was not returned for

required for MaxMediaSize of DiskDrive class in the CIM API.
MaxMediaSize of
DiskDrive class in DiskDrive.MaxMediaSize property now returns
the CIM API. the maximum capacity of the disk drive.

Modifications to the InForm OS 2.7

3PAR Confidential
 3PAR InForm OS Release Notes InForm OS Version 2.3.1 MU4

58086 StorageHardwareID The StorageHardwareID reference returned

reference returned from a CreateStorage
from a HardwareID method was invalid.
CreateStorage
HardwareID method The InstanceID of the reference to
is invalid in the CIM StorageHardwareID returned from a
API. CreateStorageHardwareID method invocation
has been corrected.

58088 Instances of Storage Instances of Storage

Capabilities of Capabilities of DynamicStorage
DynamicStorage Pool were previously not available.
Pool are missing in
the CIM API. Instances of TPD_StorageCapabilities
associated with TPD_DynamicStoragePool are
now available.

58089 InForm OS release The InForm OS MU level was not previously

level description identified.
required for
StorageSystem in The MU level of an InFormOS release has been
the CIM API. added to the Description property of the top level
ComputerSystem.

58215 Improper clearance of In some cases, a rejoined node could be put in a

Link Power Down degraded state due to the improper clearance of a
event causes node to Link Power Down event.
be put in degraded
state after a node Link Power Down events are now cleared to
rejoins. prevent a node from being put in a degraded state
after it rejoins.

2.8 Modifications to the InForm OS

3PAR Confidential
InForm OS Version 2.3.1 MU4 3PAR InForm OS Release Notes

58269 SCSI-3 Persistent In some cases, the InForm OS kernel was incorrectly
Reservation events identifying SCSI-3 Persistent Reservation events
from multiple nodes from different nodes as duplicate events and thus
incorrectly identified omitting some steps in validating and processing
as duplicate events. the events. As a consequence, each node's list of
registered SCSI initiators could be corrupted with
un-initialized values. The resulting inconsistency
triggered an assertion failure on all nodes when an
initiator registration was removed, causing a cluster
failure.

The check is now extended for duplicate events to

include the node identifier, and to ensure that the
initiator list fields are populated correctly for
duplicate events.

58426 Issuing Initiating a ModifyInstance operation on

TPD_StorageHardwa TPD_StorageHardwareIDCollection to
reIDCollection change HostPersonaID resulted in an error.
modifyinstance
returns error code in ModifyInstance operation on
the CIM API. TPD_StorageHardwareIDCollection to
change HostPersonaID no longer generates
error.

58703 Memory leak occurs The InForm OS Kernel could leak memory when
when processing processing some SCSI commands, as memory
certain SCSI commands buffers dynamically allocated for command
received on LUNs to processing were not being freed on command
which no volumes are completion. This leak affected certain commands
exported. received on LUNs to which no Virtual Volume was
exported. If a sufficiently large number of these
commands were received, available memory could
be exhausted and the controller node would go
off-line.

Memory allocated for command processing is now

properly released on command completion.

Modifications to the InForm OS 2.9

3PAR Confidential
 3PAR InForm OS Release Notes InForm OS Version 2.3.1 MU4

58766 Host ID not available It was not previously possible to get the Host ID
from from the StorageHardwareIDCollection class.
StorageHardwareID
Collection class in A new property ID has been added to the
the CIM API. TPD_StorageHardwareIDCollection class and
a new association has been implemented between
TPD_AuthorizedPrivilege and
TPD_StorageHardwareIDCollection.

58796 SPCAllowsNoTarget The SPCAllowsNoTarget and

and SPCAllowsNo SPCAllowsNoInitiators properties in
Initiators ProtocolControllerMaskingCapabilities
improperly set for did not properly represent association with target
associated target and and initiator ports.
initiator ports in the
CIM API. The SPCAllowsNoTarget and
SPCAllowsNoInitiators properties in
ProtocolControllerMaskingCapabilities
have been set to TRUE indicating that an SPC can
have no associated target port or initiator port.

58852 Operator intervention During a power or multi-node failure recovery, the

is required after a operator was required to issue a setsysmgr
power fail or multi- force_iderecovery in the following situations.
node failure recovery.
■ When all but one node had saved data without
initiating the proper handshaking with the missing
node.
■ When the system was in the process of creating a
volume (base or snapshot) at the time when the
save occurred.

Operator intervention is no longer required in

these situations, and the system recovers the data
automatically.

2.10 Modifications to the InForm OS

3PAR Confidential
InForm OS Version 2.3.1 MU4 3PAR InForm OS Release Notes

58886 Virtual volume size or In some cases, the size for a virtual volume and
snapspace reclamation snapspace resulting from a reclaim operation were
provided by System not being properly evaluated and processed by the
Manager is not metadata checker.
properly evaluated.
For a virtual volume, if the metadata view of size is
smaller than the actual size, the meta data checker
now fails the volume. Also when checking
snapspace due to a reclaim operation, if the
metadata view cannot account for all the holes, the
meta data checker fails the volume.

58939 Operator intervention During a powerfail (or multi-node panic) recovery,

is required after a if the system encountered problems paging in
power fail or multi- volume metadata, the recovery caused all the
node panic recovery. nodes to panic. In this case, when the system came
back on-line, it required operator intervention in
the form of a setsysmgr force_iderecovery
command to proceed with the recovery.

If the metadata page fails due to disk

unavailability, the system now automatically
proceeds with recovery and operator intervention
is no longer required.

Modifications to the InForm OS 2.11

3PAR Confidential
 3PAR InForm OS Release Notes InForm OS Version 2.3.1 MU4

59114 Multiple disk drives If the firmware on the 4-port 4G fibre channel HBA
are marked failed (3PAR FC044X) received a fatal error while IO was
after a path goes ongoing to disk drives via the HBA, the firmware
down due to HBA would not respond for several seconds while the
fwcore dumps. transmit and receive lasers were still active. This
could lead to the disk drives becoming
unresponsive for several minutes resulting in the
drives being marked failed by the InServ. If disk
cages were connected to multiple ports of the HBA,
this could result in multiple drives getting marked
failed leading to virtual volumes getting marked
unavailable.

This has been resolved by having the firmware turn

off the transmit and receive lasers when it gets a
fatal error thus allowing the drives to detect the
link failure immediately and recover without
becoming unresponsive.

59274 System Manager A System Manager restart occasionally ran into

restart encounters problems following an online upgrade.
problems following an
online upgrade. System Manager restarts are no longer affected by
online upgrades.

60141 The system goes off When attempting to rebalance the load after a
line when a controller controller node rejoins a cluster, out of resource
node rejoins a cluster conditions could cause the system to go off line.
due to an inefficiency
in rebalancing when Resources are now evaluated prior to initiating the
encountering load rebalancing process after a controller node
insufficient resources. rejoins a cluster.

2.12 Modifications to the InForm OS

3PAR Confidential
InForm OS Version 2.3.1 MU4 3PAR InForm OS Release Notes

60638 Out of memory When System Manager attempted to create a copy

condition caused by of itself to handle a showspace request and the
improper parsing of copy creation failed, the generated event was not
events generated by parsed correctly and resulted in repeated failures
System Manager. by the event filter. This caused events to be queued
in memory indefinitely, resulting in an out of
memory situation, especially in configurations
using Remote Copy.

The events generated by System Manager are now

parsed correctly and out of memory problems are
avoided.

61212 A Remote Copy A Remote Copy miscompare between the primary

comparison error and secondary volumes was occurring after a node
occurs between the went down while IO was being received which
primary and secondary spanned two logical disks with different owners.
volumes after a node
goes down. Remote Copy requests are now properly issued
after a backup node completes the IO to the base
volume on the primary.

61315 SCSI-2 reservations on The Logical Unit Reset task management command
addressed LUN not received on 4G fibre channel ports was not properly
properly released by releasing SCSI-2 reservations for addressed LUNs.
Logical Unit Reset Task
Management Handling of the Logical Unit Reset task
command. management function has been modified to ensure
that SCSI-2 reservations on an addressed LUN are
released appropriately.

Modifications to the InForm OS 2.13

3PAR Confidential
 3PAR InForm OS Release Notes InForm OS Version 2.3.1 MU4

2.6 Limitations of the InForm OS

2.6.1 Connectivity Issues with Remote Copy IP Interfaces for 100Mbps Switch
Ports
3PAR has observed connectivity issues with Remote Copy IP interfaces connected to 100Mbps
switch ports that can cause the Remote Copy links to negotiate their speed and duplex settings
incorrectly. This issue has the potential to cause the Remote Copy IP links, as well as the
connected switch ports, to stop working (e.g., states of offline, down, unknown, loss_sync). In
many cases, the 3PAR controller-node has to be rebooted to clear this condition.

In some configurations, depending on switch type and InForm OS version, selecting Auto-
negotiation has overcome the problem, while in others, forcing the ports to 100/Full has
provided relief. In all cases, utilizing a 1Gbps switch port has alleviated all negotiation issues.

To resolve this issue, 3PAR requires all Remote Copy IP ports to be connected to a 1Gbps switch
port in order to alleviate the observed connectivity and negotiation related problems.

This reduction in functionality is described in Product_Alert_247 200011-

RevA_RCIP_Link_Speed.

2.14 Limitations of the InForm OS

3PAR Confidential
InForm OS Version 2.3.1 MU4 3PAR InForm OS Release Notes

2.7 Known Issues with the InForm OS

2.7.1 RPI Resources Exceeded for Fibre Channel Ports (Bug 61125)
Each 4G fibre channel port on the InServ has 511 Remote Port Indicator (RPI) resources. One
resource is used when an initiator logs in. When the initiator logs out or gets disconnected, the
RPI is not released. If the initiator logs back in with the same fibre channel D_ID, the RPI gets
reused, but if it logs in with a different D_ID, the original RPI is lost. After 511 such logins, the
resource is completely exhausted and all further logins are blocked.

The resource counts in use on the 4G fibre channel ports as indicated are output to the debug
eventlog every hour. Here is an example that shows that the port has 0x1c7 (455) RPI resources
available.

2011-05-03 11:22:34.44 CEST 3 580016 Notification Debug

Notification undefined Port 3:5:4: Max xri: 0x800, iocb: 0x8ce, rpi:
0x200; Avail xri: 0x7fe, iocb: 0x8ce rpi: 0x1c7; HBQ count r0: 1983,
r1: 127 cmds: 2

 Workaround - Resetting the InServ fibre channel port using the controlport rst
<n:s:p> command restores all of the resources of the port.

2.7.2 Internally Generated Host ID Returned by Host Explorer (Bug 62212)

The host ID shown by the CLI showhost -agent command incorrectly shows the internally
generated host ID from the InServ and not the one reported by the Host Explorer running on
the host.

 Workaround - Consult Host Explorer for the appropriate ID.

2.7.3 CIM Server Error Passing Maximum Length Virtual Volume Name
(Bug 55445)
The virtual volume buffer in apilib is 1 byte shorter when it should be 32 bytes including 1
byte for the null terminator. A 31 byte long virtual volume name causes cimserver to crash.

 Workaround - Use virtual volume names of 30 bytes long or shorter to avoid this problem.
Use the setvv –name <new_name> vvname command to rename a virtual volumes over
31 bytes long.

Known Issues with the InForm OS 2.15

3PAR Confidential
 3PAR InForm OS Release Notes InForm OS Version 2.3.1 MU4

2.7.4 Physical Copy Issue with Larger Destination Volume Size (Bug 55681)
The PblkRemain field describes how many physical blocks still remain to be copied from the
parent volume during the physical copy process. The number of blocks should always be
positive. This bug occurs when using a larger size destination volume than parent volume. The
value of PblkRemain is calculated using the size of the parent volume rather than the size of
the destination volume.

 Workaround - Use parent and destination volumes of the same size when creating
physical copies.

2.7.5 4Gb/s 2-Port adapter Issue

The 4Gb/s 2-port Fibre Channel Adapter firmware does not detect a loss of sync when the host
reboots and has older adapters (ex: QLA2342) that do not turn off the laser.

 Workaround - None.

2.7.6 Downgrade Issues

The following are downgrade issues for InForm OS release 2.3.1 MU2.

2.7.6.1 Issues Preventing Downgrades

These issues will prevent an InServ Storage Server from downgrading from InForm OS release
2.3.1 MU2 to 2.2.4. (These issues do not apply to a downgrade from 2.3.1 MU2 to 2.3.1.)

■ Dual SD Space - Thin Provisioned virtual volumes can have two Snap Data (SD) volumes in
2.3.1 and 2.3.1 MU2. This is not a supported feature in previous InForm releases. Prior to a
downgrade, dual SD TPVVs must be converted to single SD TPVVs by removing all of the
snapshots for those TPVVs.

■ Raid Multi-Parity (MP) - Raid MP was not supported in InForm OS releases prior to 2.3.1,
therefore any Raid MP Logical Disks (LDs) must be tuned to a supported RAID type or
removed prior to downgrade.

■ SD Reclaim Space - SD Reclamation is not supported by InForm OS releases prior to 2.3.1.

Before a downgrade, SD volumes that have its space reclaimed must be removed. If that SD
is used by snapshots only, then all of its snapshots must be deleted so that SD can be freed.
If that SD space is used by the TPVV itself, then that TPVV would need to be copied over to
a new single SD TPVV. After the copy operation completed, the original TPVV can be
removed.

2.16 Known Issues with the InForm OS

3PAR Confidential
InForm OS Version 2.3.1 MU4 3PAR InForm OS Release Notes

■ 256 Read/Write Snapshots - In InForm OS 2.3.1 and 2.3.1 MU2, the number of read/write
snapshots was increased to 256 per virtual volume. Previous release supported a smaller
maximum number of read/write snapshots per virtual volume. A downgrade will be
prevented if any virtual volume has more read/write snapshots than the maximum
supported on the previous release. When enough snapshots have been removed to meet
this criterion the downgrade is allowed.

■ Multiple Read/Write Snapshots per Read-Only Snapshot - In InForm OS 2.3.1 and 2.3.1
MU2, the creation of multiple read/write snapshots per read-only snapshot is permitted.
Previous releases supported a single read/write snapshot per read-only snapshot. A
downgrade will be prevented if any read-only snapshot has more than one read/write
snapshot. Once enough read/write snapshots have been removed to meet this criteria, the
downgrade is allowed.

■ Synchronous Long Distance Remote Copy - If there are Remote Copy groups
configured with multiple targets, those groups must be removed or reduced to a single
target before a downgrade is allowed.

■ 16 TB Support - Virtual volumes larger than the maximum size supported in a previous
release must be removed before a downgrade is allowed.

■ Multiple Remote Copy Links per Node - InForm OS release 2.3.1 and 2.3.1 MU2 allow
multiple Remote Copy links to be configured on a individual node. Previous releases only
allowed a single link per node. Before downgrading, ensure that additional Remote Copy
links are dismissed to ensure a valid configuration for the downgrade. A check is run prior
to downgrade to validate the existing Remote Copy link configuration.

2.7.6.2 Issues Causing a Change to the InForm Configuration

The following issue causes a change to the InForm configuration when downgrading from
InForm OS release 2.3.1 or 2.3.1 MU2.

■ Host Personas - If a system running InForm OS release 2.3.1 MU2 is downgraded, host
persona fields are cleared and the old port persona settings resume effect. However, these
settings are not kept up-to-date nor are they converted back on downgrade, so they will
not reflect any configuration changes made since loading 2.3.1 MU2. On a system which
only ran 2.3.1 MU2, there will be no port persona settings to revert to.

Prior to the downgrade, a check is run to ensure the system can run in the downgraded
state. If host persona specific configurations are detected, a message is issued. This
condition will need to be resolved before the downgrade is permitted. One example of this

Known Issues with the InForm OS 2.17

3PAR Confidential
 3PAR InForm OS Release Notes InForm OS Version 2.3.1 MU4

would be different host persona settings sharing a port which can not be supported via
port personas.

2.7.7 Remote Copy and Recovery Manager Snapshot Interaction Issues

When a non-zero scheduled synchronization period is specified for an Asynchronous Periodic
Remote Copy volume group, any application coordinated snapshots taken on the same
Remote Copy group by Recovery Manager may conflict with the periodic snapshots taken by
Remote Copy. This can result in Remote Copy (RC) snapshots that cannot be deleted and/or are
not coordinated with the application.

 Workaround:

Remote Copy can support the mixture of Recovery Manager snapshots and Remote Copy
snapshots (outside of Recovery Manager) if the following procedure is followed:

Inside the Recovery Manager Windows scheduler script:

1) Set the Remote Copy period to 0.

2) Verify Remote Copy SyncStatus as "Synced".

3) Execute Recovery Manager snapshot creation for Remote Copy.

4) Wait for SyncStatus to be completed and Recovery Manager to take a snapshot of

secondary volumes.

5) Recovery Manger will monitor the SyncStatus to be "Synced" every 120 seconds as
default then issue snapshot creation of secondary volumes. Wait until these
snapshots complete.

6) Reset Remote Copy period to the sync period it was at prior to setting to 0 in step 1.

WARNING: Steps 2 and 4 are critical to ensure Recover Manager-created

snapshots and Remote Copy-created snapshots do not interfere with each other.

NOTE: See Product Alert 247-200010 in DCS for a more detailed explanation.

2.18 Known Issues with the InForm OS

3PAR Confidential
InForm OS Version 2.3.1 MU4 3PAR InForm OS Release Notes

2.7.8 Firmware Core Dump from FC Adapter by LSI

(Bugs 24178, 30353, 32351)
Under certain error conditions, the firmware on a Fibre Channel Adapter from LSI may crash
resulting in a firmware core dump. The driver for the adapter resets the port pair after
collecting the firmware core dump.

 Workaround - There is no workaround for this and no user action is required. After the
core dump is generated, the port pair is reset and comes online automatically after a few
seconds.

2.7.9 Windows 2003 Pop-Up

Client machines running the Windows 2003 operating systems receive an unknown device
pop-up from Windows Plug and Play security checking when connecting to a
2.3.1 MU2 InServ Storage Server system.

 Workaround - To avoid this device pop-up message you can install NULL driver for 3PAR
SCSI Enclosure Service device located on DCS.

2.7.10 Cage Loop Down Alert is not Auto-Recovered (Bug 40274)

There are certain conditions during which a drive cage loop missing alert may not be auto-
resolved even though the drive cage loop comes back.

 Workaround - Once it has been verified that the alert is indeed invalid, check showcage -
d and showpd to verify the loop is present. The alert can be removed by removealert or
setalert.

2.7.11 System Manager Could Crash During Rolling Upgrade if Remote Copy
Groups Running (Bug 40867)
When creating snapshots during the shutdown process there is a known race condition that
may result in a system manager crash. Therefore, the user is no longer permitted to manually
take snapshots during this process. However, this is not the case with snapshots that are taken
when a synchronous remote copy group is stopped.

During a rolling upgrade synchronous groups are normally not stopped. However, if there is a
target failure during the process any associated groups will automatically be stopped, and if
the group is synchronous a snapshot for each volume in the group will be taken. This leaves
the system open to hitting the race condition described above.

 Workaround - Stop all remote copy groups before starting a rolling upgrade.

Known Issues with the InForm OS 2.19

3PAR Confidential
 3PAR InForm OS Release Notes InForm OS Version 2.3.1 MU4

2.7.12 No Explicit Way to Reset the Default Value for Tunelinks (Bug 43579)
setrcopytarget tunelinks command can be used to adjust the performance value for the
target's links using the <bandwidth> and <latency> specifiers. After using this command,
there is no explicit way for the user to reset the tuning parameters back to its default values.

 Workaround - Reset the parameter to the default value by reissuing tunelinks with
bandwidth=128000 (or higher) and latency=100.

2.7.13 RCFC Port Shutdown Failure (Bug 44520)

A problem exists during the shutdown of RCFC ports when line errors are present on the Fibre
Channel port. It is possible that the port may not shutdown correctly due to the connection
being restarted from the network multiple times due to delayed or duplicated packets being
received.

 Workaround - None.

2.7.14 Remote Copy Volumes go Stale Due to Lack of Adm Space (Bug 44963)
The default grow value of 64 GB might be insufficient for volumes which are in Remote Copy
groups. The reason is Remote Copy relies on taking snapshots during the initial sync phase and
also during the resync periods. The snapshots end up using administrator space depending on
the IOPS and throughput load on the Remote Copy volumes.

 Workaround - Use setcpg -sdgs <increment_value> <volume_name> to increase the

default grow increment. The recommended value is 96 GB.

2.7.15 controlport Configuration Gives Warning after Being Taken Offline

(Bug 44989)
Using the -f option with controlport config, the system responded with a confirmation
request, or refused to force the mode change.

 Workaround - If the -f flag was used and the system still prompts for confirmation,
respond to the prompt. If there are other messages, they need to be resolved (e.g. take
port offline), or possibly reissue the command.

2.20 Known Issues with the InForm OS

3PAR Confidential
InForm OS Version 2.3.1 MU4 3PAR InForm OS Release Notes

2.7.16 createvvcopy Issues with Halt (Bug 45142)

When a virtual volume copy is halted, the detailed status of the target virtual volume indicates
copy_failed. The detailed status of the target virtual volume continues to show this state even
if the virtual volume copy is restarted and the virtual volume copy is going on.

 Workaround - The virtual volume detailed status is cleared once the restarted virtual
volume copy completes. No user intervention is required.

2.7.17 Disk Failure Alert has Invalid/Incorrect Magazine Information

(Bug 45224)
A disk drive failure alert may not correctly identify its corresponding drive magazine location.
When a disk is considered "missing" by the system, the alert will identify the Drive Magazine
as "N/A" (Not Available). For example, "Message : Magazine N/A, Physical Disk 104 Failed."

 Workaround - The Physical Disk drive number is correctly displayed in the alert and in the
InServ Storage Server's event log. Examine the CLI output of the showpd and showpd -s
commands to determine the cage/magazine location of the failed disk.

2.7.18 Some Messages Not Clear During upgradecage

(Bug 45229)
Messages displayed during an upgrade of a drive cage are not clear if the drive cage is actually
being upgraded.

 Workaround - Run the showcage and showcage -d command to see if a firmware upgrade
is still required. Verify there are no other instances of upgradecage running, then run
upgradecage again.

NOTE: This was observed when multiple upgradecage commands were run. Do
not run multiple upgradecage commands.

2.7.19 showtask for updatevv Displays Incorrect Completion Time

(Bug 46328)
showtask for updatevv displays the incorrect completion time.

 Workaround - In these cases updatevv did complete for all virtual volumes in the set,
however the completion time is incorrect. The actual completion time can found by
gathering the creation time of the new snapshots.

Known Issues with the InForm OS 2.21

3PAR Confidential
 3PAR InForm OS Release Notes InForm OS Version 2.3.1 MU4

2.7.20 Tasks Become Active after a Failed Upgrade (Bug 46418)

There are certain situations after a failed upgrade when tasks that were scheduled and
suspended become active.

 Workaround - Following the failed upgrade using the showsched command to determine
which scheduled tasks need to be suspended again. Use the setsched -suspend
<schedule_name> command.

2.7.21 Remote Copy: Possibility of Secondary Volumes going out of Sync

with Primary Volumes with fail_wrt_on_err Policy (Bug 46443)
The intent of a synchronous Remote Copy group with the fail_wrt_on_err policy is to fail any
host I/O that cannot be successfully mirrored to the secondary cluster. There are a number of
circumstances where the host write can get a failure when fail_wrt_on_err policy is enabled: 1)
a remote copy target failure, 2) failure of the secondary cluster or, 3) a shutdown of the
primary cluster.

There is a known issue that can occur with any I/O that has been aborted as a result of an
unsuccessful write to the secondary cluster with fail_wrt_on_err policy; that is, data may
get written successfully to the primary cluster and not to the secondary cluster and host I/O
fails with errors. This will lead to a primary volume going out of sync with respect to secondary
volume.

 Work around:

1 When the host I/O is completed with errors, start a full sync of the remote copy group
having that VLUN.

2 Use the no_fail_wrt_on_err policy, which is the default policy.

2.7.22 CIM Server Core after Upgrade from 2.2.3.116 (MU1) to 2.2.4.160
(MU6) (Bug 50476)
After the CIM Server starts, it sends SLP service requests in multicast to advertise availability of
the SMI-S service capability, including the host and port. If the host address or port field is
missing from the multicast message, the problem can occur.

 Workaround - None.

2.22 Known Issues with the InForm OS

3PAR Confidential
InForm OS Version 2.3.1 MU4 3PAR InForm OS Release Notes

2.7.23 Remote Copy: Long Waits for Remote Copy Commands to Return Data
Commands can take quite some time to return on a busy system. For this reason, the secondary
Remote Copy system cannot be used as a standalone system. Commands such as showrcopy
and showvv have been observed to take between 2 and 8 minutes to return data. Remote
Copy operations are not impacted, however during the time of a DR restore, where multiple
resyncs to different targets could be taking place, commands can take a significant amount of
time to execute.

Known Issues with the InForm OS 2.23

3PAR Confidential
 3PAR InForm OS Release Notes InForm OS Version 2.3.1 MU4

2.24 Known Issues with the InForm OS

3PAR Confidential
InForm OS Version 2.3.1 MU4 3PAR InForm OS Release Notes

3
3PAR InForm CLI Release Notes

In this chapter
3.1 Installation Notes for the InForm CLI 3.1

3.2 Known Issues with the InForm CLI 3.2

3.1 Installation Notes for the InForm CLI

3.1.1 Virus Checking
Virus checking is known to slow down and potentially cause problems with installation of the
InForm CLI. Before attempting an install or uninstall of the InForm CLI, shut down all running
applications and disable virus-checking software.

3PAR InForm CLI Release Notes 3.1

3PAR Confidential
 3PAR InForm OS Release Notes InForm OS Version 2.3.1 MU4

3.2 Known Issues with the InForm CLI

Bug ID Item Description

48776 CLI online help for The CLI online help for the createsv command, in the
createsv states A Notes section, incorrectly states that a maximum of
maximum of 2048 2048 virtual copies per volume are allowed.
virtual copies per
volume are allowed. This sentence should read A maximum of 500 virtual
copies per volume are allowed.

44872 Passing a virtual volume Use of createsv –i with virtual volume sets is not
set to the createsv –i currently supported.
command results in an
error.

44937 Misleading command When using the createvvcopy command with virtual
output when passing volume sets, if both sets are empty, the output of the
empty virtual volume sets command erroneously references the
to the createvvcopy creategroupvvcopy command.
command.
Workaround: Pass virtual volume sets that are not
empty to the createvvcopy command.

45413 Integer arguments to CLI The CLI shell is a TCL-based shell, as such it parses
commands with leading 0 integers with leading 0 (zero) as octal numbers, just
(zeros) are treated as octal like many other languages and shells (bash, perl, etc).
numbers. Consequently, integer arguments passed to the CLI
commands are also treated as octal numbers if they
have a leading 0.

Workaround: If you do not wish for integers to be

treated as octal integers, do not use leading zeros.

3.2 Known Issues with the InForm CLI

3PAR Confidential
InForm OS Version 2.3.1 MU4 3PAR InForm OS Release Notes

46372 Drive cage displays do not The showcage -sfp command shows the sfp state as
correctly report degraded OK, Degraded, Failed or Unknown. For an
SFPs. unqualified SFP, the state should be Degraded but is
shown as OK.

Workaround: If showcage –sfp shows an SFP as OK,

try using showcage –sfp –d <cage_name>. If the
Qualified field shows No, then it is an unqualified SFP
and its state should be deemed as Degraded.

47270 Discrepancy between There are two issues associated with this bug.
showsys -space and
showvv -r data for base 1 Base volume space is double counted.
volumes.
This problem has already been fixed in 2.3.1 GA and
also back ported to 2.2.4 MU6 (bug 31892 and 45225).

2 The raw used space is calculated wrong for RAID 5 and

RAID 6 Logical Disks.

In this case, the showcpg –r and showsys –space

command output will display wrong information.

Bug 48234 is opened to track this problem.

Known Issues with the InForm CLI 3.3

3PAR Confidential
 3PAR InForm OS Release Notes InForm OS Version 2.3.1 MU4

3.4 Known Issues with the InForm CLI

3PAR Confidential