HP Storevirtual 4000 Instruction Manual

HP LeftHand Storage Release Notes
HP Part Number: AX696-96225
Published: November 2012
Edition: 8

© Copyright 2009, 2012 Hewlett-Packard Development Company, L.P.

Version:
•SAN/iQ software version 10.0.00.1888
•HP LeftHand Failover Manager for VMware vSphere and HP LeftHand Failover Manager for
Microsoft Hyper-V software version 10.0.00.1888
•HP LeftHand Centralized Management Console version 10.0.00.1888
•HP LeftHand Command-Line Interface (CLI) version 10.0.0.1574
•HP LeftHand Application Integration Solution Pack (was HP P4000 Windows Solution Pack) version
10.0.00.1460
•HP LeftHand Application Aware Snapshot Manager (was VSS) version 10.0.00.1488
•HP LeftHand DSM for Microsoft MPIO version 10.0.00.1486 (installs Microsoft certified DSM
driver version 10.0.00.1485)
•SRA for VMware SRM 5.0.1/5.1 version 10.0.00.433
•HP LeftHand Recovery Manager 1.1.0.123
Updated software components in this release
•SAN/iQ software
•HP LeftHand Centralized Management Console
•HP LeftHand Command-Line Interface
•HP LeftHand Application Integration Solution Pack
HP LeftHand Application Aware Snapshot Manager◦
◦HP LeftHand DSM for Microsoft MPIO
◦HP LeftHand Storage Replication Adapter Software for VMware vCenter SRM
◦HP LeftHand Recovery Manager
Supported upgrade paths
Direct upgrades to version 10.0 can be performed from the following released versions:
•Version 9.5
•Version 9.0
If you are upgrading to version 10.0 from version 8.x (8.5, 8.1, or 8.0), Online Upgrades will first
upgrade to version 9.5 and then upgrade to version 10.0.
If you are upgrading to version 10.0 from version 7.0 SP1, you must manually upgrade first to version
8.0. From version 8.0, Online Upgrades will first upgrade to version 9.5 and then upgrade to version
10.0.
Description
These release notes provide information about upgrading to this release, and about enhancements
and fixes in this release. These release notes are current as of the published date.
Update recommendation
HP recommends that all customers running version 9.5 or earlier upgrade to version 10.0.
Supersedes
Version 10.0 supersedes versions 9.5, 9.0, 8.5, 8.1 and 8.0.
Updated software components in this release 3

Platforms supported for this release
Upgrades to version 10.0 software are available for the following platforms:
•HP StoreVirtual 4130
•HP StoreVirtual 4330
•HP StoreVirtual 4330 FC
•HP P4900 G2
•HP P4800 G2
•HP StoreVirtual VSA for VMware vSphere
•HP StoreVirtual VSA for Microsoft Hyper-V
•HP P4500 G2
•HP P4300 G2
•HP P4000 VSA
Languages
Version 10.0 supports the following languages:
•English
•Simplified Chinese
•Japanese
•Korean
•Spanish
NOTE: The Configuration Interface and the on-system CLI are not localized.
Enhancements
The 10.0 release is a comprehensive release that includes enhancements in the following areas:
•Fibre Channel support in the HP LeftHand Storage environment, including a new Fibre
Channel-capable storage system, the HP StoreVirtual 4330 FC
•Active Directory integration to allow administrators to manage user authentication to HP LeftHand
Storage via the Microsoft Windows Active Directory framework
•Windows Server 2012 support
•Online Upgrade enhancements including the following:
Patch classifications to inform users whether a patch is critical, recommended, or optional◦
◦Patch release notes delivered via Online Upgrades
◦Expanded support for component firmware upgrades including disk firmware
◦Customer notifications delivered via Online Upgrades
◦Support for using a proxy server to access the Online Upgrade ftp sites
•Quorum Alarms now proactively warn of system manager misconfigurations that put data
availability at risk. Previously the Best Practice Analyzer flagged this information in Systems running
managers.
4 Platforms supported for this release

•The SAN Status Page now has an Active Tasks Status Pane, which allows you to monitor the
progress of snapshot creation and deletion, volume resynchronization and restriping, and RAID
rebuilding
•Performance Enhancements including the following:
Multi-threading capabilities allow for improved application performance◦
◦Internal data transmission improvements impact internal SAN processes, such as significantly
enhancing volume resynchronization and restriping speeds, and reducing the impact to
overall system performance resulting from scheduled snapshot activity
◦StoreVirtual VSA adds a second virtual CPU
•Release of HP StoreVirtual 4330 and 4330 FC
Rebranding to HP StoreVirtual and LeftHand OS
Over the next year HP is transitioning to new product naming in our product, documentation, and
solution materials. StoreVirtual is the new name for the LeftHand P4000 product family and LeftHand
OS is the new name for the SAN/iQ operating system.
Fixes
Released version 9.5 patch sets and patches, which are included in version 10.0 software.
Version 9.5—Patch set 04
Centralized Management Console (CMC)
•The CMC reported “export support bundle successful” when the storage system failed to export
all log files.
•The CMC notifications window did not display the most recent patches at the top of the list.
SAN/iQ operating system
•Eliminated the requirement for manually adding the RAID controller SAAP license when a P410
RAID controller is replaced.
•iSCSI sessions could be disconnected during patch upgrades.
•Upgrades and patches did not install when a storage system was added to a management group
without the required SAN/iQ partition.
•Setting the network interface speed to auto failed if it was previously set to a specific speed.
•After upgrading storage systems (nodes) in a management group from 8.5 to 9.5, if you attempt
to delete the 8.5 storage systems (nodes), it is possible under rare conditions that the 8.5 storage
systems (nodes) could cause the 9.5 managers to stop. As a result, all volumes in a management
group would go offline for up to a minute while the managers restart.
•Memory leak occurred with SNMP version 9.5 monitoring. The leaks occurred specifically when
the SNMPWALK or the SNMPGET command was invoked.
•Running concurrent CLI commands on the same storage system caused the management gateway
service to restart. For best performance, HP recommends that a maximum of two CLI commands
be sent to a storage system (node) an any time. If you need to send more commands, HP
recommends balancing the command load by sending commands to different storage systems
(nodes) in the management group instead of to only one storage system (node).
Fixes 5

NOTE: The patch fixes the problem that caused the storage system (node) to fail when multiple
commands are sent. However, even with the patch installed, sending multiple commands to one
storage system (node) could cause the storage system to respond more slowly or time out. It is the
script programmer’s responsibility to handle non-fatal errors returned by the CLI, including timeouts.
This is not a CLI limitation, but is a normal script programming best practice to ensure the operation
requested succeeds.
•An issue with the management gateway service created high and critical memory utilization alerts
that resulted in initiators being unable to communicate with the storage system for a short period
of time.
Version 9.5—Patch Set 03
Centralized Management Console (CMC)
•Addressed various localization issues that affected how text was displayed in the CMC.
•The CMC did not obtain the updated status of the cache battery backup unit (BBU) after a cache
BBU patch was applied.
•Clicking Restore Defaults on the Upgrades tab of the Preferences window did not automatically
populate the default path in the Download Directory box.
•The CMC did not notify the user that Patch 10118 was available for upgrade if no other upgrades
or patches were needed for the management group.
•Multiple log entries were created on the FTP download site for an empty download file. This empty
file is used to create a single log entry on the FTP site to track activity.
•When editing a volume, the reported volume size in the CMC differed from the size provided
during volume creation.
•Added text to the Edit Dialog window to inform users that they should not use Network RAID 5
or Network RAID 6 volumes in a multi-site cluster.
•During a remote copy operation, discovery of other storage systems (nodes) became slower when
there were 20 or more management groups in the configuration, which prevented user input in
the CMC for a long period of time.
•The value of Network Utilization statistics in the Performance Monitor displayed more than 100%
for systems using 10 GbE network interfaces. This was due to the CMC defaulting to the speed
for 1 GbE network interfaces.
•When using the 9.5 CMC, application-managed snapshots could not be created for associated
volumes on 8.5 and 9.0 management groups.
•Passwords stored in an unencrypted format in the global database caused alarms to be unavailable.
SAN/iQ operating system
•Data could become unavailable during an upgrade to version 9.5 due to a store crash.
•False warning alarms were reported after upgrading the remote management group to a later
version of SAN/iQ than the primary management group.
•HP Insight Remote Support cases were prevented from being automatically generated.
•Improved reliability of SNMP services on the storage system.
•If a read workload occurred on a RAID 5 or RAID 6 volume in a five-node management group
with one overloaded node, all volumes went offline if another storage system (node) in the
management group went offline.
6 Fixes

•The version of a store operation request was not logged prior to the request causing an upgrade
to fail. Logging the version will aid in future troubleshooting.
•Enabled the RPC load balancing policy to be changed to round robin from the default of linear
when upgrading from 9.0 to 9.5. Round robin is better for configurations that require higher
sequential read workloads and are not prone to network saturation issues.
•VSA configurations sent false warning alarms for high latency values.
•During a heavy load of small I/Os, a store process on multiple storage systems could fail in close
succession, causing unreplicated and two-way replicated volumes to be unavailable while the
store process restarts.
•A problem occurred if an initiator terminated an iSCSI session while internal login processing
was in progress.
•Changed the performance parameters for midline SAS drives to prevent a storage system (node)
from becoming overloaded when there are higher I/O loads.
•The createRemoteSnapshot CLI command did not work when the primary management group
for the remote copy was running an earlier version of SAN/iQ and the remote management
group was running a later version of SAN/iQ.
•On rare occasions, the creation of an application-managed snapshot (either manual or scheduled)
would hang while the SAN/iQ operating system connected to the application server, which
caused subsequent scheduled snapshots to hang. Now, if there is a problem with snapshot creation,
the action will fail instead.
Version 9.5—Patch Set 02
Centralized Management Console (CMC)
•A repaired storage system could not be added as a site for an existing cluster. A message about
the cluster changing from a multi-site to a standard cluster, which typically appears when deleting
a site and the cluster would contain only one site after the deletion, was displayed.
•Attempting to edit a site that included a ghost storage system caused the Edit Site dialog box to
become unresponsive.
•When email is not configured for a storage system, the warning icon no longer displays incorrectly
next to the storage system icon. The warning icon correctly displays next to the Email icon under
Events storage system (node) in the Navigation pane.
•After adding a Failover Manager (FOM) to a management group, the license registration dialog
box is no longer displayed. License registration is only required when adding storage systems.
•Clicking OK on the Edit Volume dialog box, even when making changes unrelated to size, caused
the volume size to increase or decrease automatically.
•In a management group with 700 or more servers or volumes, the CMC login no longer causes
slow performance.
•For non-English systems that use the comma as the separator instead of decimal point, the Edit
Volume dialog box displayed the Reported Size as 10 times larger than it was.
•Enhanced the Online Upgrades feature to:
◦Display a notifications window in the CMC, which ensures you are notified of any updates
to the system, such as patches, customer advisories, or updated release notes. The Notifications
Fixes 7

window displays automatically when the CMC is first opened. You can also access it by
select SAN Status Page in the navigation pane.
◦Identify the type of patch (urgent, recommended, informational) in the notifications window
so you can quickly determine the importance of the patch.
◦Improve download capability by fixing the limitation of concurrent FTP connections. Increasing
the number of concurrent FTP connections ensures that you can obtain updates in a timely
manner and avoid connectivity issues.
•The Cancel button was not localized in the Spanish CMC.
•Added localized content for the Online Upgrades notifications window and for the FOM message
that displays on the Welcome window of the Management Group, Clusters, and Volumes wizard.
•The storage system may become unstable if, when entering a server address for email notification,
the address contains an invalid character (such as a forward or backward slash (/, \) or a colon
(:).
•The firmware version for both P410 RAID controllers in a P4900 G2 will not display correctly.
Application Aware Snapshot Manager
•Application-managed snapshots fail if vCenter is running under a domain account
SAN/iQ operating system
•Application-managed snapshots fail when CHAP is enabled.
•Application-managed snapshots fail when the host IQN contains uppercase letters.
•The storage server process may encounter a race condition and restart with the Storage Server
Down alarm during heavy system CPU load conditions.
•Performance issues could occur randomly on reads with large (2 TB or greater) volumes.
•The Maximum Transmission Unit (MTU) of the 10 GbE interfaces on a P4460sb G2 (P4800 server
blade) is incorrectly limited to 4,074 bytes. This patch will enable the MTU to be set to 9000
bytes.
•Some SNMP trap event severities do not match to corresponding CMC event severities.
•Notification emails and traps for email configuration events contain substituted text for "@x" in
the event message, where xis one of several alphanumeric characters.
•Updated Slot logs are generated successfully and support log bundles can now be emailed to HP
Support.
•Enable the configuration of additional network ports for SAN management traffic and keep
management traffic separate from SAN data traffic (for P4800 G2).
•The iSCSI session auto rebalancer may get stuck, causing data unavailability because iSCSI
sessions to a volume cannot be established. This may also result in a Storage Server Down alarm
if the storage server process restarts when encountering this condition.
•There was no time zone listed for GMT+01:00-BST-Europe/London.
Version 9.5—Patch Set 01
Centralized Management Console (CMC)
•Unable to log in to a management group after scheduling a snapshot on a volume that contains
special characters in its name.
•The Volume Failover/Failback wizard cannot be accessed due to a failed iSCSI session.
8 Fixes

•Wording changes were made in the "Getting Started" wizard.
•Users can obtain software patches for currently installed software after a new upgrade has been
posted to the FTP site.
Application Aware Snapshot Manager
•Memory resource allocation during a managed snapshot is not completely returned to the system
for reuse.
•A help file shortcut for the HP P4000 Application Aware Snapshot Manager Authentication
Console was broken.
•An application-managed snapshot is not created if the management group name has multibyte
characters and these management group credentials are provided in the Authentication Console.
•Application-managed snapshots will fail if theVMware ESX server has a NFS volume mounted.
Prerequisites
Patch 10118
HP LeftHand Centralized Management Console
•For the CMC complete installation, 35 MB disk space and 50 MB RAM during runtime
•Minimum monitor resolution for the CMC is 1024 x 600
HP LeftHand Application Integration Solution Pack
•HP LeftHand Application Aware Snapshot Manager (was VSS Provider)
Windows
◦Windows Server 2003 R2 SP2
◦Windows Server 2008
◦Windows Server 2008 R2
◦Windows Server 2012 Datacenter
VMware
◦ESX Server 4.x
◦ESX Server 5.x
•HP LeftHand DSM for Microsoft MPIO
Windows Server 2003 R2 SP2◦
◦Windows Server 2008
◦Windows Server 2008 R2
◦Windows Server 2012 Datacenter
HP LeftHand Command-Line Interface (CLI)
•10 MB available space
•Windows Server 2003 R2 SP2
•Windows Server 2008
•Windows Server 2008 R2
•Windows XP SP2
Prerequisites 9

•Windows Vista
•Windows 7
•Windows Server 2012 Datacenter
HP StoreVirtual VSA on Windows Server 2012
•The VSA installer will run on Windows Server 2012. Deploying a VSA for Hyper-V or a Failover
Manager from a Windows Server 2012 system is supported.
•vCenter installation on Windows Server 2012 is currently not supported by VMware. Consequently,
using the VSA installer for Zero-to-VSA deployments is only supported when vCenter is installed
on a Windows operating system other than Windows 2012.
HP LeftHand Failover Manager for VMware vSphere
•VMware ESX Server version 4.x or later
•Microsoft .Net 4.5 or later on the installer client
•1 GB RAM
•Bridged connection through the VMware Console, or assigned network in VI Client
•15 GB of disk space
HP LeftHand Failover Manager for Microsoft Hyper-V
•Windows Server 2012 Datacenter
•Windows Server 2008 versions
2008 R2 Standard◦
◦2008 R2 Enterprise
◦2008 R2 Datacenter
◦2008 R2 Server Core
◦2008 SP2, except for Core, which is not supported
•15 GB available drive space on Windows server for the Failover Manager
•Hyper-V role installed and running
•At least 1 GB of memory reserved for the Failover Manager
•A single virtual CPU with at least 2000 MHz reserved
Version requirements
To upgrade to version 10.0 SAN/iQ software, the storage systems must be at version 7.0.01 SP1 or
later. If the storage systems are at a lower version, please contact Customer Support for access to the
interim upgrade files. See “HP technical support” (page 22) for customer support contact information.
If you are upgrading to version 10.0 from version 7.0 SP1, you must manually upgrade first to version
8.0. From version 8.0, Online Upgrades will first upgrade to version 9.5 and then upgrade to version
10.0.
Upgrading the SAN
Automatic Online Upgrades are supported in version 9.x or later. See the HP LeftHand SAN Solution
Upgrade for more information.
NOTE: Patches 10136 and 10137, which contain a firmware update for the RAID controller on
specific platforms, are automatically applied during the 10.0 upgrade.
10 Version requirements

Upgrades and Manager Quorum Alarms
In version 10.0, the notifications for optimum manager configuration in a management group move
from the Best Practice Summary to Alarms. Therefore, when you upgrade to version 10.0, you might
receive alarms about any manager configurations that do not meet the recommended best practice.
Compatibility/interoperability
For a complete list of supported SAN/iQ software releases and application software compatibility,
including for the Failover Manager, Microsoft Windows, and VMware Server, see the HP P4000 SAN
Solutions Compatibility Matrix, which is located at:
http://www.hp.com/go/LeftHandcompatibility
Mixed SAN/iQ software versions in management groups
HP LeftHand Storage management groups can accommodate storage systems that are running different
versions of the SAN/iQ software. However, you can only add storage systems that are running a
greater software version to a management group that is running an earlier software version. For
example, you can add a storage system running the version 10.0 software to a management group
currently running the version 9.0 software. You cannot add a storage system running 9.0 software to
a management group that is running 10.0 software.
If you partially upgrade a HP LeftHand Storage management group, the management group will
continue to run on the earlier software version until the entire cluster is upgraded.
Workarounds
HP LeftHand Command-Line Interface
•A disassociated Active Directory user can still execute CLI commands
After a user is removed from an Active Directory group, the user can still successfully execute CLI
commands. To resolve this issue, perform one of the following actions:
◦Reboot the entire management group.
◦Create and run a script containing five getnsminfo commands from five different authorized
users on the storage system that the user is logged in to. This will flush the user out of the
cache.
•You cannot use the CLI to delete an SNMP community string that has no IP mask
You cannot delete a community string using the CLI command deleteSnmpCommunityString
if the community string does not contain an IP mask. To resolve this issue, use one of the following
options:
◦Use the deleteSnmpCommunityString command and do not specify the IP Mask.
◦Use the ClearSnmpSettings command instead.
•The CLI command setadinfo fails with an incorrect message when an incorrect base DN is given
If the CLI command setadinfo fails, it may be because the Active Directory server or the base
DN is incorrect or do not match. However, the error message that currently displays indicates
that the "CLIQ credentials failed." If this error message displays, verify that the Active
Directory server and/or the base DN is listed correctly.
Compatibility/interoperability 11

HP LeftHand Centralized Management Console
•The CMC does not show results if you try to convert volumes to Network RAID-5 to free up space
in the cluster
You cannot convert volumes from Network RAID–10 to Network RAID–5 as a way to save space.
If you attempt this, it will cause volumes to hang and the only fix is to delete volumes.
•Export Events fails when the management group name contains certain special characters
When exporting events to a file, the default file name is the same as the management group
name. If the management group name contains either an asterisk (*) or a question mark (?), the
export will fail. To resolve this issue, delete the special character from the suggested file name.
•The CMC Performance Dashboard is unable to monitor statistics for Fibre Channel Boot from SAN
volumes
Instead of numbers, N/A is displayed. Statistics for non-boot volumes are shown correctly.
HP LeftHand DSM for Microsoft MPIO
•Rebooting an SSD node during I/O workloads causes a 20 second client side latency when the
DSM for MPIO is installed
If a storage system with DSM for MPIO installed is rebooted while I/O activity is occurring, it will
cause a 20 second delay on client applications and possibly cause application failure.
•HP LeftHand DSM for Microsoft MPIO driver info is not displayed under Storage controllers
in the Windows Device Manager
Beginning with release 10.0, the DSM for MPIO driver is listed under the System devices section
of the Device Manager. In previous versions, the DSM for MPIO was listed under the Storage
controllers section of the Device Manager.
•Changing the Fibre Channel load balancing policy for the DSM for MPIO fails and disables path
verification
If you try to change the Fibre Channel load balancing policy from Round Robin With Subset
to Fail Over Only, the change will fail. Additionally, it will disable path verification. You must
manually reset path verification on the LeftHand P4000 disk Multi-Path Disk Device Properties
dialog box. On the MPIO tab, click Details. On the DSM Details dialog box, select Path Verify
Enabled.
•Disks are reported as offline in the Windows Disk Manager after uninstalling the DSM for MPIO
In very rare circumstances, after removing the DSM for MPIO and rebooting the server, some
volumes may display as offline. To resolve this issue, open the Windows Disk Manager, right-click
the offline disk and select Online Repeat for each offline disk.
•Windows Server 2012 NIC teaming with DSM for MPIO does not recover if NICs are disabled and
then enabled
The Microsoft Windows Server 2012 NIC teaming interface does not support the DSM for MPIO.
NIC teaming on iSCSI without MPIO does work and is supported by Microsoft in certain
configurations.
•Upgrading to Windows Server 2012 from Windows 2008 R2 renders DSM for MPIO and disks
inoperable
After upgrading to Windows Server 2012 from Windows Server 2008 on a system with DSM
for MPIO installed, the DSM for MPIO is no longer functional and the volumes no longer display
12 Workarounds

in the Disk Manager. You must re-install and configure the DSM for MPIO after the operating
system upgrade.
•DSM for MPIO I/O may stall
With the DSM for MPIO installed on configurations with a large number of volumes and iSCSI
connections, when a maintenance activity like an upgrade/patch installation, or any other activity
that is related to volumes like snapshots or restriping, occurs on the HP LeftHand Storage, there
can be delayed I/O. Depending on the application that is using the volume, this delay may cause
the application to experience data unavailability or to report other errors.
Notifications and alerts
•SNMP clients and multi-byte characters
SNMP clients must be configured to use UTF-8 encoding to properly display multi-byte characters.
•The email alarm does not clear when a DNS server is fixed or comes back online
The following warning may display when the system attempts to send an email alert:
EID_GLOBAL_CONFIG_MISCONFIGURED: The 'Email' configuration status is
'Misconfigured'. If DNS and the email server are unavailable briefly, the email configuration
is deleted. When DNS and the email server become available again, this message displays
because the email configuration no longer exists. You must reset the email configuration to stop
this warning message and begin receiving email alerts again.
Failover Manager
•Unable to upgrade a management group to 10.0 that is running a version 8.1 Failover Manager
When using a version 10.0 CMC and logging into a Failover Manager running version 8.1, the
following error message is displayed: HAL device is missing. If this error message displays,
you will be unable to upgrade the management group that contains this Failover Manager. To
resolve this issue:
1. Remove the 8.1 Failover Manager.
2. Upgrade the CMC version on the management group.
3. Add a 10.0 Failover Manager to the management group.
•HP LeftHand Failover Manager for Microsoft Hyper-V installer hangs on Windows Server 2012
When installing the Hyper-V Failover Manager on a server running Windows Server 2012, the
installation hangs. To resolve this issue, you must install .NET 3.5 before installing the Failover
Manager. Normally the missing files will be downloaded from Windows Update, but if access
to Windows Update is not possible, enabling .NET Framework 3.5 will fail unless an alternate
source for the missing files has been specified. If Windows Server 2012 is installed in Server
Core installation mode, you must use the command line (DISM.exe) to install .NET 3.5. If Windows
Server 2012 is installed as a full installation (in Server with a GUI mode), you can use the command
line or Windows Server Manager→Add Roles and Features to install .NET 3.5.
NOTE: You will need to change the DVD drive specification.
•A degraded FOM can cause management group outage due to manager timeouts
If a Failover Manager becomes unresponsive and cannot participate in quorum, the management
group cannot respond to commands. However, the management group does not recognize the
problem with the Failover Manager, and the management group becomes unresponsive.
This issue can be caused if the Failover Manager is installed on the HP LeftHand Storage instead
of network hardware other than storage systems. Installing the Failover Manager on the HP
LeftHand Storage is not supported.
Workarounds 13

Fibre Channel
•HP-UX servers cannot see SAN/iQ Fibre Channel LUNs
Fibre Channel on SAN/iQ does not work with HPUX and is not supported at this time.
•Windows Server 2012 Hyper-V cluster shows disk errors on the active node during scheduled
application-managed snapshots when using Fibre Channel
The snapshot process quiesces the volume, then takes the snapshot, then mounts the snapshot
using Fibre Channel, then unmounts the snapshot. Unmounting the snapshot breaks the Fibre
Channel connection, and this sometimes generates MPIO and disk errors in the Windows Event
log. You can ignore these errors. The snapshot completed successfully.
Hardware and platforms
•When upgrading to 10 GbE for storage systems running SAN/iQ version 9.5 or later, the RAM
must be installed
When upgrading to 10 GbE on a P4000 G2 platform running SAN/iQ version 9.5 or later,
ensure that you install the RAM as instructed in the HP StorageWorks P4000 G2 10 GbE SFP+
Hardware Installation Instructions. Without the RAM installed, there will not be sufficient memory
for the 9.5 store to start.
•A P4900 controller firmware mismatch event does not generate an SNMP trap
If the P410 controllers in a P4900 G2 storage system have different firmware versions, the firmware
version mismatch generates an event. However, that event does not generate an SNMP trap. This
situation could occur if a controller is replaced in a storage system and the replacement controller
is at a different firmware version.
•Emulex 10 GbE card offers a new port speed setting of 9 Gb with SAN/iQ 10.0
The Emulex 10 GbE card now supports a 9 Gb port speed setting with the P4800 G2.
The 9 Gb port speed setting may display as an option for the 10 GbE card when used on other
storage system models besides the P4800 G2. The 9 Gb port speed setting should only be used
with the P4800 G2.
•When splitting the P4800 G2 10 GbE NICS into a custom configuration for iSCSI and management,
the TCP speed/duplex reports Unknown
On P4800 G2 storage systems running the SAN/iQ software versions listed in the table, if the
Ethernet ports are configured with a speed other than the supported speeds listed, the network
controller Speed/Method and Duplex/Method are reported as Unknown when viewed on the
TCP Status tab of the storage system Network category.
Supported NIC configurationsSAN/iQ software version
100 Mb, 9.9 Gb, 10 Gb9.5.00 Patch Set 2 or greater
100 Mb, 1 Gb, 9 Gb, 9.9 Gb, 10 Gb10.0.00
Additionally, an event notification is received for "NIC Speed slow." This is a reporting issue
only: it does not affect actual functionality or the bandwidth allocated to network ports. To resolve
this issue, configure the network bandwidth of the P4800 G2 Ethernet ports to supported speeds.
14 Workarounds

•Undefined and Unknown values are showing for Usage Remaining and Wear Status of drives
under SmartSSD Wear Gauge in System Management Homepage for P4900 G2 SSDs
The System Management Homepage displays incorrect values for the SSD status. The correct
status is viewable on the ADU. The CMC also reports the correct status in the following locations:
◦The Hardware Information report in the storage system Diagnostics category
◦The Disk Setup tab in the storage system Storage category
An event email is also generated with the SSD status.
•The CMC reports a drive status as Active and Normal, but the Drive Status is degraded in iLO
The iLO reports the HDD status as Degraded, but the CMC shows the drive as Active on the
Storage category Disk Setup tab. The Hardware Information report shows the drive status as
Normal and in the Diagnostics tests, the Disk Smart Health test passes. This issue is due to failed
communication with the drive carrier. The ADU and iLO logs show warning message
“Authentication on this drive carrier failed due to a communication error. This may not be an
authentic HP carrier. I/O operations to the device will not be affected, but the Smart Array
controller will not control the LEDs to this device." However the problem was not communicated
to the CMC.
•The raw capacity and usable disk space are reported in different units of measurement for an HP
StoreVirtual storage system
Due to the different units of measurement between Terabytes (TB) and Tebibytes (TiB), the raw
capacity reported in an HP StoreVirtual storage system may appear to be lower than expected. One
TB of data is 10^12 bytes where a TiB is 2^40 bytes, a difference of approximately 100 GB. The
CMC reports values in TiB and HDD manufacturers typically report their size in TB. For example,
a 3 TB HDD would be reported as having a size of 2793.97 GB by the CMC. For an HP
StoreVirtual 4330 storage system with 8 1 TB HDDs, a raw capacity of 7 TB would be expected
when using RAID 5 (where one drive is used for data parity), but the CMC would report a raw
capacity of 6519.26 GB. This is the expected behavior, and it is simply two different methods
of representing the same number of bytes.
Installation and upgrades
•A CMC on a Windows 2008 Server is unable to automatically discover storage systems if the
firewall is on
Disable the Windows Firewall on the computer on which the CMC is installed.
•To ensure desired system settings are retained after upgrading from version 8.5, use Copy
Configuration before upgrading
Storage system settings are not retained after upgrading from version 8.5 to version 10.0. In the
case of storage systems in a management group, one storage system in the group serves as the
source for the settings that are applied to all the storage systems being upgraded. Therefore, use
the Copy Configuration function in a version 8.5 or earlier CMC to apply the desired settings to
all the systems that will be upgraded.
The storage system settings that may be retained after upgrading to version 10.0 include settings
for the following features:
◦Alerts, including Alert emails
◦SNMP general settings and traps
◦Remote log files
Workarounds 15

•Blank windows appear when installing the CMC on an HP-UX 11 server
If blank windows appear during an installation of the CMC and you are using the HP-UX Window
Manager (SXM/CDE), check the default font. If it is white, change it to black and begin the installer
again.
•The CMC install and uninstall wizards contain special characters such as question marks
The following unsupported configurations may cause the CMC install and/or uninstall wizards
to display special characters instead of the expected text:
◦The system OS language is set to a different language than the system locale language (for
example, the OS is set to Chinese and the system locale is set to Japanese).
◦The system locale is set to an unsupported language, such as French.
•Upgrading the CMC using Online Upgrades on a 64-bit Red Hat Enterprise Linux 6.x server
To upgrade the CMC using Online Upgrades on a 64-bit RHEL 6.x server, install the following
two packages from 32-bit RHEL 6.x:
◦glibc
◦libstdc++
After installing these packages, try installing the CMC again.
•The HP LeftHand Centralized Management Console is not upgraded when using Online Upgrades
on an HPUX 11.11 server
To resolve this issue, use the HPUX CMC installer, CMC_<build
number>_Installer_HPUX.bin that is already downloaded in the HP\P4000\UI\downloads
folder. Or you can download the CMC installer from the following website:
http://www.hp.com/go/LeftHandDownloads
•The CMC installation wizard shows special characters when the location is set to Korean on a Red
Hat Enterprise Linux 6.1server
When using the CMC installer on a RedHat Linux 6.1 server and the location is set to Korea,
Korean characters are not displayed correctly. Special characters are displayed instead.
•Online Upgrades proxy type is not documented
If you are using a proxy server, the Online Upgrade feature in the CMC only supports a SOCKS
proxy server. The default SOCKS proxy port is 1080. Update the proxy details on the Preferences
→Upgrades window.
•Clicking "View Notifications" causes the CMC to become unresponsive while downloading the
Upgrade files via Online Upgrades
If you close the View Notifications window while the upgrade files are downloading, and then
you attempt to open the View Notifications window again, the CMC stops responding. After all
the upgrade files finish downloading, the CMC becomes responsive again.
•The "Exchange storage system" and "Swap storage system" options are labeled with the same
text in the Japanese version of the CMC
This issue will be corrected in a future release.
16 Workarounds

•The current Japanese translation gives a wrong meaning for the word “Complete” in the CMC
Installation Wizard.
This issue will be corrected in a future release.
•On Windows 2008 Server, if proxy is configured, the VSA installer cannot connect to the ESX
Server
If proxy settings are configured on the Windows 2008 Server on which you are attempting to
run the VSA installer, the installer is unable to connect to the ESX Server. Before running the VSA
installer, you must disable any proxy settings on the Windows 2008 Server.
iSCSI
•CHAP name in multi-byte characters requires the server to be configured with the IQN name
You must create the server first and then edit that server to add the CHAP name:
1. Create a new server and select CHAP not required, and enter the IQN.
2. Next, select CHAP required and enter the CHAP information.
3. Click OK. The server is created with both CHAP and the IQN information.
You can also edit an existing server to add the IQN if necessary:
1. Log off the target in the iSCSI initiator.
2. In the CMC, navigate to the server. Right-click and select Edit Server.
3. In the Authentication section, select CHAP not required and enter the IQN in the Initiator
Node Name field.
4. Select CHAP required and click OK.
5. Reconnect the volume in the iSCSI initiator.
•iSCSI sessions fail if server configuration is changed
Due to a code enhancement, if any of the following items for the iSCSI server configuration are
changed, all server connections will be disconnected:
◦iSCSI initiator node name
◦CHAP required or not required
◦CHAP name, target secret, and initiator secret (if CHAP is required)
This change ensures that the connections be re-established as soon as changes are implemented.
Workarounds 17

Networking
•In rare situations, creating a bond may result in an event notice
Due to timing issues in the software, after you create a new bond, the event notice NIC
Motherboard:Port2=down may appear. You can safely ignore this event. To verify that the
bond was created correctly, check the status of the bond by navigating to the storage system
Network category and select the TCP/IP tab.
HP LeftHand Recovery Manager
•The CLI must be installed on the same system where the HP LeftHand Recovery Manager is installed
for auto-mount policy-related commands to work properly
To use related CLI commands, ensure that the CLI is installed on the same system where the
Recovery Manager is installed.
•The Auto-Mount Policy window in the Recovery Manager does not indicate if the policy is added
for a volume set
The Recovery Manager can only provide volume set information on snapshots, not volumes.
•The Windows Application Event log continuously fills with error messages if the auto-mount service
failed to mount a snapshot when the server is not found in the management group
1. In the Recovery Manager, pause the Auto-Mount policy until the server is added to the
management group.
2. In the CMC, add a server to the management group that contains the snapshot.
3. In the Recovery Manager, start the policy.
•The Recovery Manager documentation for version 10.0 does not list the current supported versions
of SAN/iQ and Windows
The Recovery Manager documentation (Online Help and Deployment Guide) was not updated
for 10.0. Therefore, the "Supported version of Microsoft Windows Server" and "Supported
versions of SAN/iQ" lists do not include Microsoft Windows Server 2012 Datacenter and SAN/iQ
version 10.0, respectively. However, both of these versions are supported with Recovery Manager.
The Recovery Manager documentation will be updated in the next release.
Remote Copy
•While cross–version Remote Copy is occurring from 10.0 to 8.5 management groups, if copying
halts because the cluster is full, the CMC displays the "See Alarms Window" error message,
although the 8.5 management group does not have the Alarms window
The version 9.x and greater Alarms and Events feature is not compatible with pre-9.x versions of
software. Review the Alarms window for version 9.x or later management groups. If you are
operating in a mixed SAN environment with management groups earlier than version 9.x, also
review the Alerts window.
Snapshots
•Application-managed snapshots will not work if the Windows server running VirtualCenter Server
is not in the domain
The following workarounds apply to all supported versions of Microsoft Windows. The Windows
server computer name and full computer name must be the same for the workarounds to work.
Do one of the following:
◦Create a domain and join the management computer to the domain
◦Change the Pegasus server service logon:
18 Workarounds

1. Open the service property page.
2. Select logon type.
3. Change the logon type from Local Service to the account that has administrative
credentials.
•An error occurs when an application-managed snapshot is created during NIC failover on an
application server
If an application-managed snapshot is created while a NIC failover is in progress on an application
server, the following error may display in the Windows event log:
App-Managed SS Failed: Could not get list of volumes on server
Wait until the NIC failover has completed. Application-managed snapshots should then resume
successfully.
•Intermediate VMware snapshots are not deleted when a timeout error occurs
The process of creating an application-managed snapshot for a VMware volume involves creating
snapshots for all virtual machines on the volume, creating the SAN/iQ snapshot, and deleting
the intermediate virtual machine snapshots. If a timeout error occurs when the snapshot is being
created (this can occur if a virtual machine is busy), the snapshot will still be created, but the
intermediate virtual machine snapshots will not be deleted.
You can determine if this issue occurred by reviewing the event log. Error messages such as
Failed to create application-managed snapshotor Failed to create
snapshots for one or more VMs will appear. Also, the Application Aware Snapshot
Manager will not indicate that snapshots were deleted. To resolve the issue, you must delete the
intermediate snapshots manually.
•When creating an application-managed snapshot of a cluster shared volume, an error message
displays on the passive server
If you have a Hyper-V configuration with cluster shared volumes, you may see the following error
when an application aware managed snapshot is taken on one of the cluster shared volumes.
Cluster Shared Volume 'Volume3' ('CSV3') is no longer directly
accessible from this cluster node. I/O access will be redirected to
the storage device over the network through the node that owns the
volume. This may result in degraded performance. If redirected access
is turned on for this volume, please turn it off. If redirected access
is turned off, please troubleshoot this node's connectivity to the
storage device and I/O will resume to a healthy state once connectivity
to the storage device is reestablished.
The error indicates that the volume is no longer accessible, but the volume is in a passive state
on that particular node (and active on the other node). The snapshot is taken successfully. You
can ignore this message.
•The CLI command GETSERVERCAPABILITIES fails on Windows Server Core
Microsoft does not support .NET on Windows Server Core. With the release of version 9.5, which
included .NET, the Application Aware Snapshot Manager is no longer supported on Windows
2008 Server Core or Windows 2008 Server R2 Core. The Application Aware Snapshot Manager
documentation (online help and deployment guide) incorrectly list support and installation
instructions for Server Core. This information will be removed in a future release.
•The Application Aware Snapshot Manager is causing VSS error messages on the Windows cluster
server that does not own the volume
If you have a failover cluster configuration using Windows Server 2012 servers and an
application-managed snapshot is taken on a volume, the following error message may display
in the Windows Application log on the cluster server that does not own the volume:
Workarounds 19

Volume Shadow Copy Service error: Unexpected error querying for the
IVssWriterCallback interface. hr = 0x80070005, Access is denied.
This message can be ignored. The snapshot was take successfully on the cluster server where the
volume resides.
•Scheduled application-managed snapshots fail intermittently if volumes have active I/O
This issue occurs if there are cluster shared volumes in the configuration and the I/O load is set
to 8K OLTP. The failures generate error messages indicating that the application-managed snapshot
has failed due to a connection timing out or due to some other conflict. The error messages include
the following examples:
◦The snapshot schedule 'xxxx' status in cluster 'xxxx' is 'Failed': 'App-Managed SS Completed
with Warning: VSS Provider: Point in time snapshot is created. The system could not quiesce.'
◦The snapshot schedule ‘xxxx’ status in cluster 'xxxx' is 'Failed': 'App-Managed SS Failed:
Another app-managed snapshot is in progress.'
◦The snapshot schedule 'xxxx' status in cluster 'xxxx' is 'Failed': 'App-Managed SS Failed:
VSS operation timed out.'
◦The snapshot schedule 'xxxx' status in cluster 'xxxx' is 'Failed': 'App-Managed SS Failed:
Cannot get application server information: connection timed out'.
•SQL Server is not quiesced when taking application-managed snapshots on an SMB cluster volume
With Windows Server 2012, SQL Server is not quiesced when an application-managed snapshot
is taken on an SMB cluster volume. The snapshot appears as if it is a managed snapshot, but
because SQL Server was not quiesced, it is not a true managed snapshot.
•Unable to take application-managed snapshots of Fibre Channel Hyper-V volumes
Version 10.0 requirements for using a Controlling Server IP address have expanded. If you are
using application-managed snapshots, you must assign a controlling server IP address to each
server you create in the CMC. The information below is available by clicking on the link What
IP address do I use? in the New Server and Edit Server windows.
“The IP address to use depends on your application type and environment. For Windows Failover Clusters, there will
typically be more than one LeftHand OS server object associated with the cluster. For this environment, enter a
controlling IP address for each of the LeftHand OS server objects in the Windows Failover Cluster. The table below
describes which IP address should be entered in the controlling server IP address field for the LeftHand OS server
object.”
20 Workarounds