Locating orphaned backup data

[RubinCompServ]

Around 6-9 months ago, my VBR seems to have gone through a period where it removed old backups from the configuration but didn't actually delete the corresponding backup files on the target. As a result, I have a sizeable amount of old backup files taking up space on my repo and not associated with any backup job. Does VeeamONE have a report that can tell me which files are "orphaned" backup files so that I can go through and delete them?

[RomanK]

Hello David,

Not exactly the same as requested, but still could help with the task Backups on Repository report.
There you could see the Parent Jobs and Latest Restore Point columns and find old backups or jobs.

Thanks

[RubinCompServ]

@RomanK,

I finally got my VeeamONE working properly (a separate issue) and was able to run this report. Unfortunately, it does the opposite of what I'm looking for: I'm looking for backup (VBK & VIB) files that are not associated with a backup anymore. In other words, these are files that were once associated with a backup but VBR thinks that it aged them out and removed them but, for whatever reason, they never actually got deleted from the actual repository. Rescanning the repo does not bring them in, but they still exist and count against the customer's quota, which is problematic when I try to justify why someone needs >200TB of backup space with only 100TB of servers and a very basic backup job (they had around 50TB used from VBKs from 2 years ago, despite the backup job being a basic Forever Forward with 14 days' retention).

[ginnyh]

Hey There,

I was just wondering if there was any solution to this issue as we are experiencing the same problem

Thanks

[RubinCompServ]

Sadly not. VeeamONE seems to be rather lacking in the Veeam Reports arena.

[jorgedlcruz]

Hello,
We do not have a current solution for this as what we report and monitor is pulled from VBR database. We never look, explore, or discover repositories files ourselves.

I agree that the repository scan and orphaned backup files hunter, sounds very nice as a feature, but presents so many challenges from the VeeamONE perspective that I do not foresee tackle the issue soon.

Perhaps VBR can include this repo scan vs database scan as a daily task, adding @Dima P. to see what he thinks.

Thank you

[Dima P.]

Possibly you can re-import such backups to a Veeam B&R and review all the imported backups one by one?

[RubinCompServ]

The problem is that there are some VBK/VIB files out there that, for some reason, don't show up as part of a backup, don't get imported when the repo is rescanned, but DO count against the configured quota. It could be that this is just a bug in VBR that it doesn't pick up every VBK/VIB in a scanned repo; I've opened tickets for that in the past and the engineers have all said the equivalent of, "Maybe there's something wrong with the file which is why VBR doesn't pick it up. Just manually delete it and you'll be fine." This works great as a one-off but doesn't solve the problem of those "rogue" files that exist for any customer that hasn't done their math.

[Dima P.]

I'd look into backup file properties via B&R console to get the full list of associated files within this backup / backup chain. Then delete files which are not listed in the UI from the file system (do not go full delete, copy those to another folder, make sure that backup chain is fully operational and only then delete).

[RubinCompServ]

@Dima P.,

I have upwards of 500 backup jobs to go through. To run a manual audit of each backup job against what's on the actual repositories would be a herculean task.

[mjs66]

I'm have this exact same issue. I've found a way to list the files associated with a backup job, but have not found any means via powershell to get a full list of files on a repository so they can be compared.

[RubinCompServ]

It seems to be that this should be a (relatively) simple item to script (famous last words, amirite?). VBR (or the VBR db) is aware of all files that are associated with all backup jobs. It also has the ability to read all of the files in a given repo. As such, there should be a way to query that list of files (not just a repo rescan) and do a diff against the list of files that are associated with backup jobs on that repo. Doing so will give you a list of files that are present and shouldn't be, and will also give you a list of files that aren't present but should be (which means that there are missing backup points).

[DanielJ]

This might be the most persistent Veeam bug ever. Check the script I wrote in the thread below. It needs work since it's not aware of modern per-vm chains, but it could be a starting point.

cloud-connect-backup-f43/bug-vbr-can-le ... 63-30.html

[RubinCompServ]

It looks good - I'll have to play with it. Thanks!