Hello,
I am testing the snapshot copy locking feature. We're on OnTAP 9.12.1P11 and SnapCenter 5. Here is my understanding of limits I've discovered during testing thus far. Can someone verify whether my analysis is correct and/or provide insights into these issues? I'm starting to think this might only be a viable solution for the VMware plug-in.
- CIFS - you cannot restore an entire volume to an earlier snapshot unless it's the most recent locked snapshot copy. Of course you can restore individual files still via the hidden snapshot directory.
- SnapCenter Plug-In for SQL - any snapshot (locked or not) can be restored from, but the restore method is always single-file restore and can't be a full volume restore, unless it's the most recent locked snapshot. Best practice is to have very large databases on their own volume so the entire volume can be quickly restored, but this wouldn't be doable on anything but the latest locked snapshot. Restoring from any other snapshot would result in a lengthy outage.
- SnapCenter Plug-In for Windows - similar to SQL, I believe this would never be a full volume restore unless it's the most recent locked snapshot. If the volume contained significant amounts of data and we needed to restore from an older snapshot, this could cause a very large outage.
- SnapCenter Plug-In for VMware - I believe functionality is 100% the same with locked snapshots. Restoring a full VM creates a clone behind the scenes that is mounted, with the existing VM powered off and the backed up copy powered on and migrated, followed by unmounting the backup. You could mount a whole datastore snapshot if needed, detach the old datastore and attach the new, turning the snapshot into a separate cloned volume. Same process.
