Synthetic Full queries

[mrhcarlin]

I raised a support ticket for this, but in hindsight perhaps it would have been best asked here because I'm only after some answers to some questions / general advice (even though in fact we are having a technical issue). The ticket case is # 00529614 for reference.

So basically we have a production environment with 13 VMs. These are backed up to a 12TB SAN every night, with an Active Full scheduled for Saturdays. Since the beginning of the year, we've been having sporadic issues with our weekend Active Full job as the jobs are failing due to lack of space. During the week we have ~1.5TB left on the SAN, but when the Active Fulls are being created this seems to be get depleted and so the job fails. Once the next incremental starts on the Monday night, space is freed from what I gather is a result of the retention policy (although my expertise on Veeam is somewhat vague so I could be wrong!), which for the record we have set to 21 and 14 depending on VM importance. As we are a school we store a lot of information - student and staff data - so we need to be able to make restores from 3 weeks prior to the latest backup.

I was looking at Synthetic Fulls (with transform) as an alternative because it only keeps one VBK on the volume at any given time. This would save us an enormous amount of space because the VBKs for our two domain controller backup jobs, of which there are two, create VBKs that are between 600-900GB, and we have about 8 of them! So it seems obvious what's taking up so much space - high retention policy with several large VBKs being stored.

My main question is as follows: if I change this weekend's full backup from an Active Full to Synthetic Full (with transform), will the oldest VBKs be transformed or just the most recent VBKs? If the latter, then I will surely have to delete old VBKs in order for this to work, otherwise our jobs are almost certainly going to fail again.


Thanks for any advice.

[foggy]

I was looking at Synthetic Fulls (with transform) as an alternative because it only keeps one VBK on the volume at any given time. This would save us an enormous amount of space because the VBKs for our two domain controller backup jobs, of which there are two, create VBKs that are between 600-900GB, and we have about 8 of them! So it seems obvious what's taking up so much space - high retention policy with several large VBKs being stored.

Yes, this is expected with forward incremental mode.

My main question is as follows: if I change this weekend's full backup from an Active Full to Synthetic Full (with transform), will the oldest VBKs be transformed or just the most recent VBKs? If the latter, then I will surely have to delete old VBKs in order for this to work, otherwise our jobs are almost certainly going to fail again.

Only the latest backup chain (consisting of the latest VBK and corresponding VIB files) is transformed into reversed incrementals during transformation. Previous restore points will be deleted according to retention, but seems that you need to remove at least some of them manually just to free up some space for the synthetic full to be created (corresponding records will stay in Veeam B&R database until being deleted by retention).

[mrhcarlin]

Thanks. That answers my question exactly. So would removing some of the oldest VBKs/VIBs manually be safe to do? Would there be any ill effects in doing this?

[foggy]

No other effects except for the described above.

[mrhcarlin]

Thanks. What I'll do then is move them to an external storage device (just in case) to free up the SAN and change our weekend job from Active Full to Synthetic Full with transform. Hopefully this will solve our storage issues.


Thanks again for the help!

[foggy]

You're welcome. Feel free to ask if any additional questions arise.