HP P6000 Continuous Access Implementation Guide (T3680-96431, August 2012)

Planned Failover Procedure
To execute a planned failover:
1. Optionally, move storage management to another management server. For instructions, see
the HP P6000 Command View User Guide.
2. Ensure that all DR groups have resumed and are fully normalized.
Check the DR group status using HP P6000 Command View or HP P6000 Replication Solutions
Manager. If a DR group is merging or a normalization is in progress, wait for the process to
complete.
3. If the write mode is currently set to asynchronous, set it to synchronous.
If you are running XCS 6.000 or later, consider the following:
• You must wait for the log to merge after changing the write mode to synchronous and
then verify that the actual write mode has changed to synchronous.
• Transitioning to synchronous write mode will impact server I/O performance. For
environments running in asynchronous mode on XCS 6.000 or later, it is best to halt all
server processing while data in the write history log is draining, as server I/Os will be
treated in a synchronous replication manner, resulting in a severe performance impact
on server write I/Os.
• When performing a planned site failover, you must put DR groups in synchronous mode
before performing the actual failover activity. Omitting this mode change before performing
the site failover results in performing an unplanned site failover (and the resulting
normalization). Once the site failover has finished, you can change the replication mode
back to asynchronous mode again.
• If the intersite link (or links) are broken and the DR group (or groups) has entered logging
but has not yet been marked as a normalization (as occurs when the log reaches 100%
capacity), a link is re-established and a log merge should be performed before the failover
is performed.
• If the source site requires maintenance, put the DR groups into synchronous mode, wait
for the write history logs to drain, and perform the failover. Once failover occurs, suspend
replication and the destination array will start logging. When the primary array returns,
re-establish a HP P6000 Continuous Access link, and take one of the following actions:
◦ Wait for the log to drain.
◦ Invalidate the log and start a normalization.
4. Properly halt all applications, and then shut down the servers. Ensure that the server has
properly flushed all internally cached data. Failure to do this will result in the loss of data
cached on the server.
5. Fail over the DR groups.
6. Issue operating system commands for resuming host I/O to the new source disks. For
operating-system specifics, see “Resuming host I/O after failover” (page 118).
7. If you plan to operate for an extended time at the remote site, you can change the write mode
to asynchronous after the original primary array is back up and running.
a. If the DR group is currently suspended, resume it and wait for log to finish merging.
b. Once the log has finished merging, set the write mode to asynchronous.
NOTE: Remember to set the write mode back to synchronous and allow the write history
log to completely drain before you fail back to the original source.
Failover and recovery procedures 107