Procedures for Recovering Data

Information and steps for the procedures for recovering data are described in:

• Steps for Hosting a Volume for Backup Purposes
• Steps for Failback
• Steps for Switching Roles in a Replication Configuration
  

Steps for Hosting a Volume for Backup Purposes

You may want to temporarily host a volume from the secondary group for backup purposes. This enables you to use a third-party application to back up a stable copy of volume data while the original volume continues to be available to users.

The following procedure assumes that no writes will be made to the volume while it is hosted on the secondary group, so there is no need to replicate any changes to the primary group.

Steps for hosting a volume for backup purposes are shown next and also described in Table 1: Backup Steps:

1. On the primary group, configure replication on the volume, choosing whether to retain the failback snapshot in the local replication reserve, and create a replica. Make sure the replication completes.
2. On the secondary group, temporarily promote the inbound replica set to a recovery volume. After the promote operation, the replica set will disappear from the list of inbound replica sets and the recovery volume will appear in the list of volumes.
   
   See Temporarily Promoting an Inbound Replica Set to a Recovery Volume.
3. Use a backup application to back up the recovery volume.
4. When the backup is complete, on the secondary group, demote the recovery volume to an inbound replica set, returning it to its original state. The recovery volume will disappear from the list of volumes, and the replica set will reappear in the list of inbound replica sets.
   
   See Demoting a Recovery Volume to an Inbound Replica Set.
5. On the primary group, replicate the volume to synchronize the groups.

Table 1: Backup Steps describes the backup steps on the primary and secondary groups.

Steps for Failback

You may want to temporarily host a volume from the secondary group and then fail back to the primary group, returning to the original replication configuration. For example, you may want to do this if the primary group is unavailable because of failure or maintenance or if the original volume is destroyed.

While the volume is hosted on the secondary group, the group will track volume changes. Later when the original volume becomes available and you are ready to fail back, you can replicate to synchronize volume data across the groups.

If the failback snapshot is available when you fail back to the primary group, you may be able to synchronize the groups by replicating only the volume changes, not the entire volume contents. The failback snapshot represents the failback baseline for the volume (that is, the point in time at which the primary and secondary groups are synchronized).

Steps for failback are shown next and also described in Table 2: Failback Steps:

1. On the primary group, configure replication on the volume, choosing whether to retain the failback snapshot in the local replication reserve, and create a replica. Make sure the replication completes.
2. When the volume becomes unavailable (or is manually set offline), on the secondary group, temporarily promote the inbound replica set to a recovery volume. After the promote operation, the replica set will disappear from the list of inbound replica sets and the recovery volume will appear in the list of volumes.
   
   The promote dialog boxes enable you to set up access controls for the recovery volume, if desired, so systems can connect to the recovery volume.
   
   See Temporarily Promoting an Inbound Replica Set to a Recovery Volume.
3. When the original volume becomes available, on the primary group, demote the volume to a failback replica set. The volume will disappear from the list of volumes and reappear as an inbound replica set that is flagged as a failback replica set.
   
   See Demoting a Volume to a Failback Replica Set.
4. To synchronize the groups, on the secondary group, configure the recovery volume to replicate to the primary group and create replicas as needed. As usual, you must allocate local replication reserve and replica reserve. However, you do not have to delegate space. You cannot use manual transfer replication for a recovery volume.

Note: If the failback snapshot does not exist in the primary group, the first replication will be a full volume transfer.

When you are ready to fail back to the primary group, set the recovery volume offline and create a final replica. Ensure that the replication completes.

5. On the secondary group, demote the recovery volume to an inbound replica set, returning it to its original state. The recovery volume will disappear from the list of volumes, and the replica set will reappear in the list of inbound replica sets.
   
   See Demoting a Recovery Volume to an Inbound Replica Set.
6. On the primary group, promote the failback replica set to a volume, returning it to its original state. The failback replica set will disappear from the list of inbound replica sets and the volume will reappear in the list of volumes. If the option to keep the failback snapshot is enabled on the volume, a replica will be created automatically in order to re-establish the failback baseline. Because the volume data is synchronized across the groups, no data will be transferred.
   
   See Promoting a Failback Replica Set to a Volume.

If, during the failback process, an unrecoverable disaster occurs on the primary group, you can make the inbound replica set promotion permanent, disabling the ability to demote the volume to the original inbound replica set. See Making an Inbound Replica Set Promotion Permanent for more information.

Table 2: Failback Steps describes the failback steps on the primary and secondary groups.

Steps for Switching Roles in a Replication Configuration

You may want to switch roles in a replication configuration. For example, if GroupA was replicating a volume to GroupB, you may want to switch the configuration and replicate the same volume from GroupB to GroupA.

If the failback snapshot is available, you may be able to synchronize the groups by replicating only the volume changes, not the entire volume contents. The failback snapshot represents the failback baseline for the volume (that is, the point in time at which the primary and secondary groups are synchronized).

Steps for switching roles are shown next and also described in Table 3: Steps for Switching Roles:

1. On the primary group, configure replication on the volume, choosing whether to retain the failback snapshot in the local replication reserve. When you are ready to start the transition, set the volume offline and create a replica. Make sure the replication completes.
2. On the secondary group, temporarily promote the inbound replica set to a recovery volume. After the promote operation, the replica set will disappear from the list of inbound replica sets and the recovery volume will appear in the list of volumes.
   The promote dialog boxes enable you to set up access controls, if desired, so systems can connect to the recovery volume.
   
   See Temporarily Promoting an Inbound Replica Set to a Recovery Volume.
3. On the primary group, demote the volume to a failback replica set. The volume will disappear from the list of volumes and reappear as an inbound replica set that is flagged as a failback replica set.
   
   See Demoting a Volume to a Failback Replica Set.
4. To synchronize the groups, on the secondary group, configure the recovery volume to replicate to the primary group and create replicas as needed. As usual, you must allocate local replication reserve and replica reserve. You cannot use manual transfer replication for a recovery volume.

Note: If the failback snapshot does not exist in the primary group, the first replication will be a full volume transfer.

5. When you are ready to complete the role change, on the secondary group, convert the recovery volume to a non-recovery volume, making the promotion permanent and disabling the ability to demote the volume to the original inbound replica set. (You can still demote the volume to a failback replica set.) The resulting volume will appear in the list of volumes and will no longer be flagged as a recovery volume.
   
   See Making an Inbound Replica Set Promotion Permanent.
6. On the primary group, convert the failback replica set to an inbound replica set, making the demotion permanent and disabling the ability to promote the replica set to the original volume. (You can still promote the inbound replica set to a recovery volume.) The resulting inbound replica set will appear in the list of inbound replica sets and will no longer be flagged as a failback replica set.
   
   See Making a Volume Demotion Permanent.

Table 2: Failback Steps lists the steps for switching roles on the primary and secondary groups.