HP XP RAID Manager Software - Manual (Page 361)

HA Failover and failback 361
3. The S-VOL or the link recovers from the failure. Host A issues the
pairsplit –S , paircreate –vl, or pairresync command to update the
P-VOL data by copying all data, or copying differential data only. The
updated P-VOL is fed back to the S-VOL.
CA recovery procedures
Follow these steps to recover CA operations:
1. If an error occurs in writing paired volumes (for example, pair
suspension), the server software using the volumes detects the error
depending on the fence level of the paired volume.
2. Issue pairdisplay to the paired volume or group to get status
information.
3. If necessary, issue the horctakeover command to recover P-VOL write
access if the secondary volume fails and the primary is fenced (write
inhibited).
4. If the primary volume fails, split or suspend the paired volume and use
the secondary volume as the substitute volume.
5. Find out the reason why the pair was split. Repair or recover the failure
and resynchronize your pairs immediately.
Abnormal termination
A CA command can abnormally terminate for many reasons, for example:
• The remote server is down.
• A local server failure.
• A disk array failure.
• The disappearance of the RM instance.
Check the system log file and RM log file to identify the cause.
If a command terminates abnormally because the remote server fails,
recover the remote server, and then reissue the command. If the RM
instance has disappeared, reactivate the RM instance. If you find failures