3PAR InForm OS 2.3.1 MU2 Release Note (325-200142 Rev B, October 2010)

2.18
Known Issues with the InForm OS
3PAR InForm OS Release Notes InForm OS Version 2.3.1 MU2
3PAR Confidential
2.7.19 Tasks Become Active after a Failed Upgrade (Bug 46418)
There are certain situations after a failed upgrade when tasks that were scheduled and
suspended become active.
 Workaround - Following the failed upgrade using the showsched command to determine
which scheduled tasks need to be suspended again. Use the setsched -suspend
<schedule_name> command.
2.7.20 Remote Copy: Possibility of Secondary Volumes going out of Sync
with Primary Volumes with fail_wrt_on_err Policy (Bug 46443)
The intent of a synchronous Remote Copy group with the fail_wrt_on_err policy is to fail any
host I/O that cannot be successfully mirrored to the secondary cluster. There are a number of
circumstances where the host write can get a failure when fail_wrt_on_err policy is enabled: 1)
a remote copy target failure, 2) failure of the secondary cluster or, 3) a shutdown of the
primary cluster.
There is a known issue that can occur with any I/O that has been aborted as a result of an
unsuccessful write to the secondary cluster with fail_wrt_on_err policy; that is, data may
get written successfully to the primary cluster and not to the secondary cluster and host I/O
fails with errors. This will lead to a primary volume going out of sync with respect to secondary
volume.
 Work around:
1 When the host I/O is completed with errors, start a full sync of the remote copy group
having that VLUN.
2 Use the no_fail_wrt_on_err policy, which is the default policy.
2.7.21 Busy Volumes May Lead to Failure of setrcopygroup failover
Command when Executed Against the Periodic Secondary
of a Synchronous Long Distance Remote Copy Group (Bug 46693)
If the setrcopygroup failover command is executed against the periodic secondary of a
Synchronous Long Distance Remote Copy group while the volumes of that group are busy
(such as during volume growth or tuning) or some other transient error is encountered the
command will eventually time out and fail rather than waiting for the condition to clear. The