6.3.2 HP StoreAll OS Release Notes

• The ibrix_quotamonitor command now synchronizes with the host errors logged after an
upgrade to 6.1.4. Previously, the error message, “Failed to sync with host <IP>”
displayed in the quotamonitor.err file even though the host was up and running and the
quota usage reported was incorrect.
• Performance of snapshot reclamation jobs has been increased, particularly for configurations with
multiple snapshot directories (snaptrees).
• Under certain circumstances, the purge utility was not working correctly during the snapshot
reclamation task, which prevented the snapshot directory (snaptree) from being deleted. The fix
ensures the purge utility works correctly.
• After an installation (and system reboot) using the QR DVD, the output of the date and hwclock
settings did not match. This could cause events in the var/log/messages directory and the
system logs to display an incorrect time stamp, making it difficult to trace events logged. The fix
ensures date and hwclock are synchronized.
• During an upgrade from 5.6 to 6.1.4 or later, the upgrade failed during the export of NFS shares
with the message, "Operation not permitted". The upgrade script was modified to ensure
the export completes successfully.
• The maximum recommended number of CIFS/SMB shares that can be configured per StoreAll
Storage system has been increased to 200 (previously 20) when using the StoreAll OS GUI or
CLI. If you are using the Microsoft Management Console (MMC), up to 60 shares can be managed
from 2 file serving nodes. HP does not recommend using the MMC to manage shares on systems
with more than 2 file serving nodes.
NOTE: The MMC limit is based on client-side MMC timeouts that are beyond HP’s control. In
some environments, other factors may increase latency, which will reduce the number of shares
that can be effectively managed with MMC. If you experience errors managing a large number
of shares with MMC, HP recommends using the StoreAll OS GUI or CLI to create and modify
shares instead because the StoreAll OS does not have the same timeout issues. Be aware that the
time to complete StoreAll OS add and modify operations increases as the number of shares
increase. The amount of the delay depends on the number of shares. In testing, it took
approximately 60 seconds for a modification with 50 shares and approximately 200 seconds
for a modification with 100 shares. The recommendation to use the MMC on systems with a
maximum of 2 file serving nodes is because communication between nodes causes operations to
take longer to complete, and these timeouts are likely to be exceeded.
• Fixed the ibrix_chassis command to ensure SAS HBAs are not reported as failed or missing.
Originally the code only detected HBAs with "SAS HBA" in the name; now it also detects HBAs
with "SCSI storage controller" in the name.
• Fixed the log rolling functionality of Statstool to ensure log files older than 7 days are removed
from the log files in the /usr/local/ibrix/log/statstool directory. This fix ensures errors
or failures are not encountered due to space issues.
• Fixed an issue to ensure that the DNS entries for the node running the active Fusion Manager are
populated in the /etc/resolve.conf file. This in turn, ensures that ibrix_auth command
is successful and the file system can be shared as a CIFS/SMB share.
• Fixed an issue that caused multiple registrations of remote clusters (using either the GUI or CLI) to
hang after the registrations had previously been deleted.
• When using the Adobe Acrobat Save As option to create a new file on a CIFS/SMB share from
a client running MAC 10.7.5, the newly created file is no longer hidden.
• When using the FMT to update SAS switch firmware and the FMT is presented with an infinite
prompt, the FMT would time out and the update would fail. Previously, the user had to manually
6 6.3.2 release