Backing up VM's with snapshot

[jbsengineer]

I notice that the backups of VM's with snapshots do not contain the snapshot files. Are the snapshots rolled into the parents VMDK inside the backup job? Or does VEEAM ignore snapshots?

[tsightler]

Actually, I'd like to know the answer to this too. It's very unclear what the behavior is in the case of an existing snapshot. I've been meaning to test this but haven't had too much time lately.

[ceebee]

I am being told by Veeam support that the backup and replication will roll the snapshots to the original vmdk files. I have not tested either.
Will be testing it very soon.

[Vitaliy S.]

Hello Josh, Tom and Chet!

Basically, snapshot backup depends on the mode you choose, with Network processing mode (3rd radio button) all the snapshots are backed up/replicated along within your parent VMDK file, as for other modes (VCB and vStorage API) all the snapshots data is consolidated with the data of your parent virtual disk during the backup.

[Gostev]

Vitaly is correct, but it is also important to note here that if your VM has multiple "snapshot trees", only one - currently "active" - will be processed (backed up or replicated) in the Network processing mode.

Snapshot processing is functionality VMware does not provide, which explains why it is only available with our proprietary processing engine. Both vStorage API and VCB always consolidating VM VMDK/snapshot data to provide backup product with the latest VM state.

Thank you.

[ceebee]

Gostev,
Is this true say if the snapshot chain is compromised? What happens if the snapshot chain is compromised and the VM is running off a current snapshot...how does the SAN backup mode handle the older snapshots and its corresponding delta files to commit them back to the original VMDK files on the target?

[Gostev]

I am not sure what do you mean by "compromised snapshot chain"?

The backup process picks up the "effective" VM disk state of the source VM, no matter whether VM runs off a snapshot or not. Every block in the produced VMDK file (which will go into backup file) is exactly identical to effective current virtual disk state, as source VM sees its virtual disk at the moment when backup starts (there is always only one active state, no matter of number of snapshots, and what snapshot VM runs of).

So on the storage level, as backed up VM disk data is retrieved, in scenario you described some of these blocks will be picked up from VMDK file, and other from snapshot file - but this basically comes down to how snapshots work, so nothing specific to backup...