HP StorageWorks 3000/5000 Enterprise Virtual Array Release Notes (VCS 3.110) (5697-6378, August 2007)

Fixes
The following is
sues have been resolved in VCS 3.110:
General issues
• Addressed an issue where a single disk d rive failure could lead to RAID1 inoperative state.
• The fault detection logic on the backend that prevents unnecessary disk and port failures is
enhanced.
• A power cycle of a device Fibre Channel loop switch may, in some instances, lead to a failure of
the connected port, leading to a Single Port on Fibre (SPOF) condition.
• During regeneration of a failed disk drive, error monitoring thresholds are suspended to remove
the possibility of soft errors leading to a dual disk failure within a single redundant storage
set (RSS).
• Addressed an issue which may occur when background allocation of virtual disk space is in
progress and a write is received for space which is not yet mapped leads to the inability to
completely delete a LUN.
• Corrected an issue where a high number of writes on a Data Replication tunnel may lead to
missing data within the tunnel which results in a resource allocation issue.
• Detailed information has been added to Controller Event lo gs for loop recovery events (Event
0x063b0008) to assist support personnel.
• Previously added support for the 30-10022-01 RoHS compliant loop switch has been removed
due to an issue caused by modified loop timing.
• Detailed information indicating when and how array initialization occurs has been added to the
Controller Event Log.
• A progress indicator has been added to allow the display of d evice ungrouping operations within
HP Command View EVA.
• Within an HP Continuous Access EVA environment fault isolation between source and the
destination EVAs has been enhanced.
• Anissuewhichcouldleadtodatainconsistencybetweensourceanddestinationarrayshas
been addressed.
• An issue which lead to stalling of reconstruction or merging operations of VRAID5 disk groups
has been addressed.
• A disk drive, in some instances, may simultaneously induce a persistent condition on both the
Fibre Channel loops, blocking the communication with all the other disk drives sharing the same
dual Fibre Channel loops. This issue is addressed.
• An issue which could prevent leveling and migration process from completing has been addressed.
• An issue has been addressed which could cause unre coverable errors during a VRA ID1
reconstruct process leading to an inoperative state of the disk group.
• An issue where scheduling of work twice in a background process could lead to inconsistent unit
state has been addressed.
• An issue where an incorrect unit state could be caused by an erroneous flag set at startup has
been addressed.
• An issue which caused host time-outs in HP Continuous Access EVA during s napshot creation
has been addressed.
• An issue which prevented access to vir tual disks by HP OpenVMS hosts after a site failover in
HP Continuous Access EVA configuration has been addressed.
Con
troller crash issues
• Management commands which are used to gather VDisk status data could lead to a controller
res
tart due to a resource allocation issue. This issue has been addressed.
• A snapshot creation operation that led to a overcommit state within the array may led to a
controller restart in some instances. This issue has been addressed.
4