Copy snapshot from snapmirror destination to primary volume

[rocky marten]

Heyy!
I am wondering if anybody has ever had a use case to copy snapshot from a snap-mirror volume (DP) to primary (RW) volume and how did u achieve it?

My use case is as below

We have recently migrated from VNX to AWS FSX ONTAP. Running almost 100+ volumes in prod. Each volume is backed up using snapshots and snapmirror. For a bunch of our volumes snapshots end up consuming a bulk of primary storage. This is becoming a bit of annoyance as we have to constantly clean up snapshots (support nightmare). We have increased some of these volumes but because our current snapshot policy stores almost a years worth of snapshot, I do not see that as a long term solution.

One of the solutions I am working on is to keep most of the snapshots under snap-mirror volume and keep a small percentage (10 days worth) snapshots on primary. I have tried recovery from the snap-mirror using both volume level and file level recovery and it works fine. However it does not give user access to browse a snapshot and choose.

Therefore I am interested in looking at any option which will allow me to copy a snapshot from snap-mirror volume to primary volume. Or even browse snapshots on a snapmirror volume.

Open to hear other solutions as well

Thank you

Copy snapshot from snapmirror destination to primary volume

[outer vortex]

You can clone the volume, break the clone, and SnapMirror back to the source. I believe you can even start a reverse SnapMirror from a specific snapshot copy. Not sure which is preferred but it's possible with ONTAP, but not sure what Amazon allows.

[trim ibex]

Unfortunately, there is no direct method to get a snapshot back onto the original source volume. There are a couple of options: 1) you could provide the user access to the R/O volume via NFS or SMB. You would then gain access to the /snapshots folder on the DP volume. 2) you could create a clone as Paul suggests and then create a new SnapMirror policy back to the source cluster, but this would require additional storage adequate to store the volume as a new DP volume. 3) you could do a snapshot restore to a different volume on the source cluster and give the users NFS or SMB access to the new volume, but again, you would need adequate space on the source cluster to host the full volume. #1 is the most space efficient as you don't need any additional space on the DP or source clusters.