When backing up VMs to disk (Per VM files created), jobs have the option of changing the sequence in which VMs are processed.
I have configured our disk backups to process the largest VMs first, the smallest last.
When our tape jobs run, the VM files are written to tape in a different order.
I can see no logical order being applied. Can you please advise exactly what logic is applied to processing Per VM backups (created with reverse incrementals) when being written to tape.
Is it possible to have the ordering of VMs to tape mirror the ordering of the disk job?
The reason this is important to me is that I run parallel processing to 2x tapes using per VM backups, processing 2x backup jobs in a single tape job:
Can you please explain in detail so that my assumed workings below can be ruled out / confirmed?
I would hope to ensure that all big jobs are written to tape first (using both tapes) assuming that the remaining smaller jobs will better fill the tape.
The imagined negative alternative is that the same job without sized ordering could for example fill the first tape 70% with a mix of VMs, then a large VM which is too big to fit on the remainder of tape 1 would be written to tape 2 (I see no issue with this) but that Veeam then considers tape 1 to be full as it couldnt fit a large job on the tape, hence wasted space on the tape as well as reduced job performance as all writes to tape are now focused on 1 tape instead of 2.