Feature Request - Use pre existing Snapshots.

[vmtech123]

I'm keeping immutable snapshots on my production storage, I do hourly for most jobs, but on the replicated site I'm running daily keeping the retention longer.

When I backup from storage snapshot, Veeam will create a Thin clone of the volumes in a new volume group then map it to the Veeam proxy to do the backup. This makes sense as It can't create a "snapshot" using the current volume group as it's set to immutable and will stay behind.

I love the new IBM plugin and it's working great with Veeam.

Now, my question is, why can't we choose to use an EXISITING snapshot to backup from as well?
If I have a snapshot policy with immutable snaps in place already, I could just as easily select the most recent snapshot as the source of my backup without having to create a new one. I can do restores, file level restores and everything else from those snaps already and they show up in the Veeam console.

The only reason I could think of is the SAN is creating these on a policy and also removing them. There would have to be some way that the SAN can't delete the snapshot during the backup if the job ran past the retention policy timeframe. I'd just as well keep my immutable snapshots longer if that had to happen.

Either way, consider this a request as Immutable snaps on production are a pretty common thing to have.

[foggy]

Hi Scott, other considerations apart - what about application consistency? Veeam B&R creates a snapshot after VM snapshot is triggered with all the necessary quiescing stuff done so the backup is performed with AAIP, while policy snapshots on the storage are just crash-consistent.

[vmtech123]

That’s a good point.

I do some application aware jobs, but I guess for crash consistent jobs it would still work. Apart from domain controllers and email servers it would work for me.

The benefit is leaving on volume protection because if I turn that on the veeam snapshots can’t delete due to IO on the volume.

Ideally IBM should allow that to be set on a volume level. I’ll tell them to implement it lol

[rennerstefan]

Hi,

As part of the process we also need to ask VMware about the changed blocks to backup (CBT). And those CBT informations have to be in-line with the storage snapshots we want to read from. If we would use pre-exsisting snapshots there would not be any change block information available as the snapshot does not contain any vmware snapshots and with that no CBT info.
We would have to run a Veeam based changed tracking which would take way longer as we would need to compare our last run with what is in the Snapshot without talking to VMware.
Technically it would possible but doesn't make much sense unless you want to run FULLs all the time.
And yes, restore out of crash consistent snapshot is not an issue as in this case we don't leverage any information about the content of the snap from VMware.

Thanks

[vmtech123]

Ah, that makes sense with the snapshot and CBT. Thanks for the response, it's always interesting to see why something won't work. It doesn't make sense to do this then.

[rennerstefan]

you are welcome, glad we were able to answer your question.
I think its also useful for others so thanks for asking.