HP StoreAll OS 6.5.1 Release Notes
Previously, these disk drives were identified as part of a rare data miscompare issue that was
identified in firmware stress testing in a lab environment. The recommendation was to update to
the minimum supported firmware version. The issue did not occur in a customer environment.
• The FMT has been updated to discover all attached disk drives and report their current firmware
versions. Previously, the FMT displayed an error message that a disk drive was unsupported or
unknown if the firmware for that disk drive was not included in the FMT.
• When WebDAV is enabled and the internal WebDAV files in the root of the file system are
scanned, all .lock and .pag files in the .DAV and webdav directories are now automatically
excluded from the scan. Previously, you had to manually exclude these files.
• Updated the FMT and StoreAll OS Software to ensure only firmware version 4.6.247.5 and driver
version 4.6.247.7-1 is available for the NC550, NC552, and BL460c G7 LOM network interface
cards when used in a StoreAll system. The firmware is in the FMT; the driver is in the StoreAll OS
Software. This combination of driver and firmware versions fixes an issue that could cause loss
of network connectivity or, on rare occasions, a kernel halt or panic after loading or unloading
the be2net driver. As a result, the following log entry displayed in dmesg or /var/log/
messages:
be2net 0000:xx:00.0: Out of MCCQ wrbs
• The StoreAll OS Software has been corrected so that Red Hat Linux/CentOS 6.x clients do not
experience delays when writing to a file.
• Changed the health monitoring check for StoreAll clients to occur once per hour. Previously, it
occurred every 10 seconds and this was causing the node running the active Fusion manager to
be flooded with client health monitor messages.
• Updated the FMT to fix an issue with the firmware smart component for hard disk drives
ST31000424SS, ST32000444SS, ST2000NM0001, and ST1000NM0001. Previously, the
firmware update would fail with error codes 0x1011 and 0x1032 when attempting to update
these disk drives.
• Updated the Red Hat kernel to prevent loss of connectivity from clients to NFS shares after an
incomplete manual failover or failback operation (due to the node not being powered down first).
This issue only occurred on 10 GbE systems configured for mode 6 bonding, which HP does not
recommend for 10 GbE, and which are considered uncommon router configurations. Previously,
this issue required manual intervention to resolve.
• The HP StoreAll Storage Installation Guide has been updated to notify users that, before beginning
the installation process, they should verify that the NTP and AD servers are synchronized. This
synchronization ensures that the cluster nodes do not receive conflicting time updates, which could
lead to errors (such as file system creation failures).
• Per customer request, the procedure for renaming a cluster has been added to the Administrator
Guide for all platforms. The content is available in the "Maintaining the system" chapter.
• Replaced the "quiet" parameter in the crashkernel boot entry (in grub.conf) with the "verbose"
parameter. This ensures that more detail is available to HP Support via the serial console when
creating a crash dump.
• During a node failover, a message is now displayed instructing the user to configure the GUI
browser to support cookies. Without cookies supported, the Events database could fill beyond its
limit and consume all available Java heap memory because multiple GUI login requests were
being received per second, and each request was triggering an event to be created.
• Red Hat Linux 5.8, which is included in StoreAll OS Software 6.5, includes a kernel fix that
resolves the issue of kswapd utilization being stuck at 100% when there is a large number of
processes in prune_dcache.
6.5 9