6.3.3 HP StoreAll OS Release Notes (AW549-96082, January 2014)

• When the virtual interface (VIF) IP address is changed after initial installation or after Phone Home
configuration, the change is automatically detected by SNMP traps and Phone Home solution
information. (The Fusion Manager IP address still needs to be rediscovered in HP Systems Insight
Manager (SIM) with the new VIF IP.) Previously, the traps and solution information detected the
old VIF IP address until the Fusion Manager was restarted or Phone Home was reconfigured.
• Firmware for the EG0146FARTR and EG0146FAWHU disk drive models was added to the FMT
for the 9300 and 9320 platforms. This ensures that these disk drives are updated accordingly
(flash recommendation and flash actions) when using the FMT. Previously, the FMT would indicate
these disk drives were not supported because the firmware was not included in the FMT.
• A create/open request on a file or folder on an SMB share no longer causes unused memory to
be consumed by the HP-SMB service. Previously, this caused the node to become unresponsive
until the service was restarted.
• When multiple durable reconnect requests occur simultaneously with file open requests, the HP-SMB
service no longer reports operation failures on some of the reconnected files.
• When Linux static user mapping is enabled with Active Directory (either in the GUI or the CLI), a
warning message now displays to notify users of its impact. When enabled, the UIDs and GIDs
generated by HP-SMB will not be recognized. Instead, IDs assigned in the UNIX attributes for
Active Directory or IDs from name/ID mapping solutions are used. In the GUI, the warning message
will display when the user selects either Enabled with Active Directory or Enabled with ID Mapping
and AD in the Linux Static User Mapping field on the Active Directory window of the Authentication
Wizard. In the CLI, the warning message will display when the user runs the ibrix_cifsconfig
command to enable RFC2307 (the protocol that enables Linux static user mapping with Active
Directory).
• Fixed an issue in which Phase 1 of the ibrix_fsck command incorrectly reported successful
completion but where the system log still displays an error message (ext3_dx_find_entry:
bad entry in directory) upon next access.
• Heavy I/O load on an NFS share (including rename operations) no longer causes the NFS
processes to report the state as Uninterruptible and the system to become unresponsive.
• When failing over a StoreAll server after a large number of files are unlinked, the server will no
longer hang.
• When installing or restoring a node, the netmask IP address is now retained when you select Back
from the “Configuration Summary” window to the “Form a Cluster Step 2” window in the ASCII
UI Wizard.
• After modifying bond0 to enable bond4 on a 9320 platform, the Credential Manager is no longer
uninitialized. Previously, the uninitialization caused segment fault errors to be added to
/var/log/messages every two minutes.
• When mounting two file systems in parallel from a StoreAll Linux client, the second file system is
now mounted successfully and multiple error messages are not logged in the /var/log/messages
directory.
• When installing the StoreAll Linux client on a CentOS 6.4 server, the installation now completes
successfully. Previously, the error message ERROR: portmap must be running before
loading Ibrix displayed during installation.
• When uninstalling and reinstalling the StoreAll client software, the uninstallation script automatically
unregisters the client from the cluster. This prevents an error from occurring during the registration
step in the reinstallation process.
• The ibrix_online_quotacheck command has been modified so that INODE_NOT_FOUND
(-643) errors do not display when using this tool. These errors are not related to, nor do they
affect, the file system or quota. These errors will still be available to field personnel for debugging
purposes.
8 6.3.3 release