HP Surestore Disk Array 12h - And FC60 Manual page 115

Restriction
A possibility of file system corruption exists
when resources are moved back and forth
multiple times between cluster nodes using
Microsoft Cluster Server running under
Windows NT 4.0 (Service Pack 4, 5, or 6.x).
When the cluster resource is moved, the
owning node flushes cache and then releases
the reservation to the volume(s). The other
node then reserves the volume(s). In some
instances, file system reads and writes do not
finish before the release. This results in a
reservation conflict and file system corruption.
Trying to recover using the chkdsk command
causes the file system information to be
corrected in such a way that some files are no
longer accessible.
To avoid losing access to data, do not attempt
to immediately re-open the Array
Management Window after performing a
firmware upgrade. During the upgrade, you
are prompted to close the Array Management
Window. It then takes approximately five
minutes for the upgrade to complete.
Workaround
None. The problem is highly unlikely to occur
during normal failover situations.
After you respond to the prompt to close the
AMW, wait approximately five minutes, then use
SM60 to confirm that the firmware has
successfully upgraded on both controllers before
re-opening the Array Management Window.
115