During an outage, when a secondary replicated volume is unavailable, the Availability Suite feature logs changes made to the primary volume. Once replication is restarted the secondary volume is resynchronized with the primary volume.
A failure during the resynchronization might leave the secondary volume in an inconsistent state, which can result in file system corruption of that volume. To avoid this, you can configure the Geographic Edition software to automatically create a compact dependent shadow volume of a secondary replicated volume immediately prior to resynchronization. This is a “fallback snapshot” of the secondary volume, from which the secondary volume can be reconstructed in case there is a resynchronization failure.
If you decide to configure fallback snapshots of your replicated volumes, you will require two additional volumes on each cluster for each replicated volume, as described in Availability Suite Volume Sets. You can enable fallback snapshots automatically, as described in Automatically Enabling Fallback Snapshots, when you first add a device group to a protection group. You can also enable fallback snapshots manually at any time, as described in Manually Enabling Fallback Snapshots. It is much easier to enable fallback snapshots automatically, so if possible, set up fallback snapshots for the volumes in a device group when you first configure it and add it to a protection group.
Once a fallback snapshot is enabled, the Geographic Edition software automatically activates the snapshot when the cluster hosting an Availability Suite replicated volume is switched to secondary mode, and deactivates it when the cluster is switched to primary.