HP P6000 Enterprise Virtual Array release notes (XCS 10001000) (5697-1819, March 2012)

Controller termination
This section describes issues that can result in termination of controller operation. A single controller
termination will result in a reduction of redundancy. A dual controller termination indicates a possibility
of a data unavailable condition. A recursive termination will require manual intervention to address.
• Fixed an issue where a controller can experience a processor exception termination (termination
code 0x0442011f) if there is not an active path to the destination drive at the time the I/O is
being processed.
• Resolved issues with the use of a shared internal resource in the array’s management interface
that resulted in various controller termination codes (0x0402011f, 0x0204010a, and
0x0204010c). The shared resource can no longer be used by an incoming management command
while simultaneously being used by the preceding management command.
• Fixed an issue where a controller could experience a controller termination (termination code
0x0205010a) during failover while checking if a virtual disk is already presented to that controller.
• Fixed a controller restart issue (termination code 0x02210100) that occurred when the controller
received a sequence of verify commands.
• Fixed a termination (termination code 0x0c0a0106) that could occur if the controller software
improperly attempts to free a resource used by another LUN.
• Fixed an issue where an interrupted mirrorclone cache flush can result in a controller termination
(termination code 0x021b0100).
• Improved the quantity and accuracy of termination parameters for EVA6400/8400 controller
termination events.
• Fixed an issue where a controller can terminate (termination code 42280102) during a restart
due to the receipt of an unknown completion message type.
• Fixed a controller restart issue (code 0x0442011f) that could occur when unpresenting a LUN.
• Fixed an issue where a rare condition during host port abort processing caused an item being
aborted to be processed twice, resulting in a controller termination (termination code
0x04d3011f).
• Fixed an issue where multiple threads attempting to access a common mirroring structure could
result in a controller termination (termination code 0x0204010a).
• Fixed an issue where a controller operation would terminate (termination code 0x036c01c8)
and wait for user intervention instead of logging an event and continuing to execute after
encountering an error during drive poll operations.
• Fixed an issue where I/Os could get stuck on a pending queue during a host port logout cleanup
and cause the controller to experience a stalled state and eventually execute a Cell State Manager
(CSM) Reset termination (termination code 0x4238011f) to clear the stalled state.
• Fixed an issue where controllers can experience a termination (termination code 0x0775011a)
when get_info commands are called on a snapshot that is in the first stage of deletion.
• Fixed an issue that caused recursive terminations (termination code 0x020d0104) due to mirrored
cache inconsistency, which sometimes involved inoperative LUNs.
• Fixed an issue where a rare condition during a controller replacement or shutdown could allow
the surviving controller to skip some virtual disk failover steps and encounter a Unit Stalled
Too Long or CSM Reset termination.
• Fixed an issue where the controllers run out of buffer resources and terminate (termination code
0x02160102) when flooded with write commands after a LUN becomes inaccessible due to lost
cache data being acknowledged and cleared.
Fixes (XCS 10000000) 5