HP P6000 EVA Release Notes (XCS 11001100) (5697-2516, March 2013)

The DR group and members affected by the hardware failure will be deleted during
the hardware failure resolution process. The unaffected DR group members will be
accessible to hosts for read/write access after removal from the failed DR group. If a
DR group failover was performed prior to the hardware failure recovery process, it is
possible to have LUNS with the same worldwide ID on two different P6000/EVA
systems, which can be presented to the same host. If there are applications residing
on a host that have access to both the source and destination sites that would not
tolerate this situation, access to the LUNs in the DR group affected by the hardware
failure should be prevented. This can be accomplished by either un-presenting the LUNs
to the application’s host or affecting changes to the SAN configuration.
HP P6000 Command View display
• Fixed issue that prevented accurate display of a failed host port status in EVA P6000 Command
View.
Disks (disk groups, virtual disks)
This section describes issues affecting physical and virtual disks or disk groups, which can lead to a
range of behaviors including a loss of redundancy, performance impact, or data unavailability.
• With the 11001000 controller software release, any members of a Continuous
Access DR group that are not directly impacted by a disk group meltdown will be
usable after being removed from the affected DR group and after the associated
DR group is deleted. For instance, if a disk group has a vraid5 meltdown (due to
two non-RSS mirror pair disk members failing), all Vraid1/6 virtual disks data
would still be accessible. In older XCS versions, the Vraid1/6 virtual disks would
also need to be deleted.
• Fixed an issue where when creating a DR group, the destination host access is
assigned as None, though it is specified as Readonly.
• Fixed an issue where too much spare capacity was reserved, returning disk group
spare allocation to pre-XCS 10000000 levels. This change does not require user
input.
• Fixed an issue where communication between SAS I/O modules is disrupted
causing the I/O modules to stop responding.
• Modified read cache routine in which aggressive prefetching causes high drive
latencies and performance degradation.
• Fixed issue with DR group metadata that can result in virtual disk inoperative
conditions when other failures occur on the storage system.
• Fixed issue that results in a Unit Stalled Too Long (USTL) condition when a site
failover occurs with automatic suspend on link down enabled.
• Modified the device side recovery routine to prevent instability on one link from
affecting the other link.
• Fixed issue where the failure of one backend port can cause the other port to also
fail.
• Fixed an issue where metadata volumes were not properly promoted due to read
errors on the source volume.
6 HP P6000 Command View display