I am seeing a bunch of Snapshot Full warning events being generated recently from various volumes on our filers:

A Warning event at 22 Jun 10:03 Alaskan Daylight Time on Snapshot snapmirror_sql_trns_log.snapshot on Storage System netappfiler.contoso.com: The snapshot is 251.89% full (using 37.7 GB of 14.9 GB).

It might be that I am just tired but I am having a hard time really understanding what this means. My read of the documentation is that this Volume's snapshot reserve is completely full and the snapshots have spilled over into the unused space in the portion of the Volume that is allocated for normal use (How Snapshot copies and snapshot reserve use space in a volume).

I have a few questions about this condition:

  1. Is my understanding correct? I am a bit thrown by the terminology NetApp uses as well as the "snapshot is full" (does that mean one snapshot or the snapshot reserve)?
  2. What happens if the user data grows and competes with the snapshot spill space? Does the snapshot occupying it get deleted?
  3. What condition/s generally cause this to happen? It's my understanding that once the snapshot reserve is full or if a new snapshot will fill it, ONTAP just bumps the oldest snapshot/s out the reserve as necessary. In that case, how exactly can you even "fill the snapshot"?
  • Literally "snapshot full", not "snap reserve full"?
    – Basil
    Jun 23 '15 at 1:28
  • @Basil - Yep. Literally "snapshot full". The Alarms are being generated by NetApp OnCommand DataFabric Manager
    – user62491
    Jun 23 '15 at 16:18
  • Ah, DFM makes a difference :) I've only used the version they make for CDOT, not the one you're on.
    – Basil
    Jun 23 '15 at 16:31

You should use snapshot autodelete. Without it, snapshots that exceed the reserve don't get deleted by default.

Snapshot autodelete can be configured to behave several ways. The one I use is to set the trigger to "snap_reserve", meaning it'll delete any snapshots that exceed the reserve. We use snapshots as more convenient backups, but we still take backups and would rather lose a snapshot than the ability to write to our volumes.

The wording in that error message is [edit] from NetApp OnCommand DataFabric Manager, as per your comment. It does indeed mean that the snap reserve is full.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy