Hardware manual

Group Administration Volume replication
12–9
After each replication completes, the primary group replaces the failback snapshot to update the failback
baseline. Therefore, the volume data represented by the failback snapshot on the primary group always
matches the volume data represented by the most recent complete replica on the secondary group.
If you failover to the secondary group and write to the recovery volume, you can fa
il back to the primary group
by replicating only the changes made to the recovery volume if the failback snapshot still exists. If the failback
snapshot does not exist, you must replicate the entire volume contents to the primary group to complete the
failback operation.
It is important to allocate sufficient local replication reserve to e
n
sure that replication operations complete and,
optionally, to maintain the failback snapshot.
If there is not enough free local replication reserve to complete a replication operation, one of the follow
ing occurs:
• If you enabled the option to borrow free pool space, and sufficient free pool space is available (at least 10%
free pool space), replic
ation continues. The primary group generates an informational message, specifying that
it is temporarily using free pool space during the replication.
• If you did not enable the option to borrow
free pool spac
e, or if you enabled the option, but there is not enough
free pool space, the primary group cancels the replication and generates an event message, stating that the
replication was cancelled.
If there is not enough free local replication reserve to maintain the failback snapshot, on
e of the following occurs:
• If you enabled the option to borrow free pool space, and free pool space is available, the primary group
gene
rates an informational message specifying that it is temporarily using free pool space.
• If you did not enable the option to borrow free pool spac
e, or if you enabled the option, but there is not enough
free pool space, the primary group deletes the failback snapshot and generates an event message. To reestablish
the failback snapshot, increase the local replication reserve and replicate the volume.
In addition, if you attempt to replicate a recovery volume to
the primary group, and there is insufficient local
replication reserve to store the changes, the primary group generates an event message advising you to increase the
space.
Guidelines for sizing local replication reserve
On the primary group, you specify the value of the local replication reserve and whether to keep the failback
snapshot when you configure a volume for replication. You can also enable the option that allows you to borrow
free pool space if there is not enough local replication reserve. You can later modify these settings.
The local replication reserve size is based on a percenta
ge (5% to 200%) of the volume reserve. For a thin-
provisioned volume, the volume reserve size changes dynamically, based on volume usage; therefore, the local
replication reserve size also changes.
The recommended local replication reserve percentage depend
s on whether you want to keep the failback snapshot.
• No failback snapshot. Specify 100% for the local replication reserve.
• Keep the failback snapshot. If you want to keep the failback snapshot,
specify 200% for the local replication
reserve.