Hardware manual

13–1
13 Data recovery
If you replicate a volume to a partner (see Chapter 12, Volume replication), you can recover volume data on the
partner. In addition, you might be able to fail over to th
e partner and later fail back to the original group.
About data recovery
Effective data recovery requires a well-planned disaster protection strategy and the regular creation of replicas and
backups. To protect volume data from unrecoverable failure, you can replicate a volume to a group configured as a
replication partner. See Chapter 12, Vo
lume replication.
If the volume becomes unavailable—either temporarily or permanently—you can recover the data from the
pa
rtner.
When volume failure occurs, or if the primary group is unavailable because
of maintenance, it is important to
resume data availability as soon as possible to prevent or limit application downtime.
The method for recovering data depends on the state of
the groups and your specific data recovery requirements.
See Data recovery procedures.
For example, you can clone a replica to crea
te a new volume on the secondary group. The new volume contains the
same data that existed at the time you created the replica; initiators can connect to it in the usual way. Cloning a
replica has no impact on the original volume and the replication configuration. If the original volume is still
available, replication can continue, as usual. See Cloning an inbound replica on pag
e 12-31.
In most situations in which
you must recover data, the primary group is not available because of maintenance or a
failure. In this case, you can temporarily – or permanently – fail over the volume to the secondary group and make
the volume data available to initiators. If the original volume on the primary group becomes available again, you
can fail back the volume to the primary group, returning to the original replication configuration.
You implement failover and failback by using the following
op
erations:
• promote –
Enables you to convert a replica set into a volume and snapshots. The volume contains the data
represented by the most recent complete replica. The snapshots correspond to the remaining replicas.
For example, you can promote an inbound replica set to a recove
ry volume, as part of a failover operation.
• demote – En
ables you to convert a volume into a replica set.
For example, you can demote a volume to a failback r
eplica set, as part of a failback operation.
See Failing over and failing back a volume on pag
e 13-2 for more details.