HP StorageWorks Enterprise Virtual Array Release Notes (XCS 09534000) (5697-3074, January 2014)

HP Continuous Access EVA

This section describes issues affecting HP Continuous Access EVA implementations of remote

replication, including data replication (DR) groups.

• 1000943954: During a migration from an EVA3000/5000 running VCS 3.xxx to an

EVA4400 or EVA6400/8400, a controller termination (termination code 0x0442011f)

occurred if the target EVAx4000 had existing DR groups with another EVA model

communicating on the same Fibre Channel port. This problem has been corrected in this

release.

• 1000906659: Addressed an issue where a heavy mirrored workload caused by DR operations

led to a controller termination (termination code 0x0442011f).

• 1000946972: Improved preferred port settings to ensure paths are returned consistently to

the desired controller in HP Continuous Access EVA configurations.

• 1000962806: In an HP Continuous Access EVA configuration, a correction was made to

prevent a controller termination when commands are received out of sequence across mirror

ports (termination code 0x0c0e0106).

• 1000936724: Improved support for the creation of DR groups that have an initial log size

greater than 100 GB.

• 1000983747: When setting the Failsafe on link-down/power-up feature, if the feature is

supported on the source DR group but not on the destination DR group, the feature would be

prevented from being used on the source DR group. The fix allows this feature to be set or

cleared on the source DR group even if the destination DR group does not support it.

• 1000985628: The destination member’s world wide LUN IDs were not being changed to

unique world wide LUN IDs. If you recreate this DR group with the same source members, the

destination members would have the same world wide LUN IDs of the existing volumes. The

duplicate naming would prevent the LUNs from being presented on the destination array.

Now, if you delete a DR group that still contains members, the world wide LUN IDs will be

changed to new unique world wide LUN IDs.

• 1000982147: In HP Continuous Access EVA configurations, frame loss on the intersite link

could result in a tunnel stuck timeout due to an error in the exchange retry code. This issue

only applies to XCS 09500000 or later.

• 1000906582: For an EVA controller pair running a DR consistency group, one of the

destination controllers reboots, thus temporarily stopping the data mirroring relationship

between the local controllers. Upon rebooting, the mirroring relationship of the controllers for

the DR group is restored, but in the process, sometimes terminates with

SCID_CA_BAD_PROXY_WRITE_STATE. The fix ensures that the write back cache is

synchronized before writes are processed.

Operating system/solution support

This section describes issues affecting EVA support of operating systems or solutions.

• 1000965616: Corrected erroneous test failures that occurred during Microsoft Windows

2008 Cluster Validation Testing (CVT).

• 1000967593: An issue of host port connectivity instability, which caused a Windows host

to lose access from one path to a LUN, has been fixed.

• L3 GR8 4000190147, 4000192932, 4000195799: In a server-based management

configuration, HP Command View EVA was unable to manage the EVA4400 because LUN

0 was not accessible (this did not affect host I/O). The fix ensures that LUN 0 remains

accessible.