Data Protection and Snapshot Management
Local protection questions test the difference between space-efficient point-in-time copies and full copies, and the operational rules for linking, restoring and expiring them.
SnapVX fundamentals
- A SnapVX snapshot is taken at storage group level. It is not a device; it records a point-in-time view of every device in the source group.
- Snapshots are space-efficient: only tracks changed on the source after the snapshot consume additional SRP capacity. Untouched tracks are shared with the source.
- Snapshots can be linked to target devices in a separate target storage group, giving a test or reporting host read access to that point in time, for example
symsnap -sid 123 link -snapshot_name PATCH_01 -lnsg TEST_SG. - TTL expires a snapshot automatically; secure snapshots cannot be terminated before their TTL elapses, which is a protection feature and a common surprise.
- A restore reverts source data to the snapshot, so it overwrites current data: verify the source is quiesced first.
- Unisphere snapshot policies automate creation and retention on a schedule (interval, number of snapshots, retention), which is the right answer when the requirement is recurring protection without manual steps.
Clones and VP Snap
- A TimeFinder/Clone is a full physical copy: it consumes capacity equal to the source, but the target is fully independent and does not depend on the source.
- Clones are built with an initial copy plus optional precopy and can be activated at a chosen moment for a consistent target.
- VP Snap is the space-efficient middle ground: a linked target with copy-on-write tracking, useful when the target only needs temporary access.
Worked example
Before a database patch, create a SnapVX snapshot of APP_SG with a seven-day TTL, then link it to TEST_SG on a validation host. The source group is 10 TB and changes about 2% of its data per day. After three days the snapshot has consumed roughly 600 GB of SRP space (10 TB x 2% x 3). If the patch fails, restore from the snapshot; if it succeeds, unlink the targets and terminate the snapshot, or let the TTL expire it. A full clone of the same group would have cost 10 TB up front.
Traps
- Believing a snapshot reserves the full source capacity.
- Linking targets that are members of the source storage group.
- Expecting to delete a secure snapshot early.
- Assuming a TTL rotates or replaces snapshots; it only expires one.
- Restoring without checking that the source can tolerate being overwritten.