Is it possible to merge specific reverse incremental backups inside of the chain to reduce restore points /storage space

[dasadm]

We have a case at work, where the reverse backup-chain of one job reaches as far back as the year 2020 (definitely not my recommendation, so many data loss possibilities, it's not even funny).

There is only the one full-backup for this backup job (which is running right now) and there is not even enough storage space for more full-backups.

So the idea would be (if possible) to reduce the restore points to only one per week for the years 2020 and 2021 (I couldn't change their minds for less restore point with a safer backup strategy.)
to enable more space for one or two full-backups until we can work out another strategy as well as more storage space. (If that even reduces storage space in the first place.)

I only know of cases where you can remove or forget incremental restore point in a forward incremental backup-chain, but couldn't find anything regarding reverse incremental backups.

To be a bit clearer, let's say we have a week in the year 2020 with backups from monday to friday.
All reverse incremantal backups from monday to thursday are supposed to be merged with the friday incremental restore point.
Also only the friday restore point shall be left as a restore point option while recovery of data.

Does anyone know if this would be possible?

In use is Veeam backup & Replication 12

[HannesK]

Hello,
and welcome to the forums.

There is no way to merge existing restore points. If you have no space, then reverse incremental is easy: just delete old files. After that, the configuration database will be out of sync. Rescan should fix that, but you can also double check with support

Best regards,
Hannes
PS: just curious... how many restore points do your chains have? should be up to 1000 if you say 2020?

[dasadm]

Hello Hannes,

thank you for the nice welcome! Quite the fast response. I thought an answer would take a lot more time.

While I knew that you can easily delete the oldest backup from reverse backups,
I was not aware that you can delete reverse incrementals inside the chain without disrupting the chain.

But with just deleting the restore point the changed data from these specific points would be lost, right? I don't think, that I can convince them to do this solution....

I will do some trial runs in my lab at home to see how it works out.


PS: And yes, there are quite a lot of restore points .... 932 in total

[BackupBytesTim]

To elaborate on that a bit, while it would cause the database to be out of sync until a rescan is performed, you can safely just delete the file containing a restore point in a reverse incremental backup, without affecting the ability to recover newer restore points. However, yes, any data that was backed up in that restore point will be lost.

So if you have a file that was created, a backup taken, it will be in that backup. If you do another backup, new restore point, if the file wasn't change it won't be included in the new incremental backup file. But the "full backup", the VBK file, which gets modified with every new incremental backup on your daily schedule (or however often it is) will always contain the most recent version of every file, regardless of whether it was backed up in any particular previous restore point. So you can always recover the entire computer system to the way it looked at the time of the last backup, even if you delete every VIB file in the reverse-incremental chain. (Forward incremental chains are the opposite of that, where the VBK file contains the oldest available version of every file, new versions will be in the VIB files with a forward incremental chain.)

So if your concern is with having all the individual versions of the files, then you wouldn't want to delete those VIB files, but if your main concern is just with being able to recover the entire computer system to the latest version, then you should be good to just delete those old VIB files, technically you should rescan the repository to ensure consistency in the database, but my understanding is it will still perform new backups fine even if you don't do the rescan, it'll just give a warning message.

I could be wrong about some of that, we don't use reverse incremental backups hardly at all and it's been a long time since I've had to do a repository rescan manually for anything, but my explanation of the deleting your old VIB files should be good. Perhaps Hannes can confirm that to be accurate, just to be safe.

[HannesK]

Hello,

I was not aware that you can delete reverse incrementals inside the chain without disrupting the chain.

It looks like there is a misunderstanding. Deleting a VRB file breaks everything older than the deleted VRB file.

Deleting restore points is only for cases where you cannot run backup jobs anymore because the repository is full. For all other situations, the way to go is to reduce the retention in the backup job settings.

Best regards,
Hannes

[BackupBytesTim]

For that sort of scenario though, if a particular VIB or VRB file is unusually large due to a significant increase in data on the computer, but was temporary, the data was removed shortly after a backup was performed, that single restore point would be using a large amount of data. So there's definitely a good reason to be able to delete individual restore points. I've had that situation myself, with Veeam of course I've had to delete the entire chain or at least a collection of restore points. Though with other software, such as Acronis, I'd just delete the particular version that was large. Acronis, of course, would also allow me to remove only the particular file or files that took up the extra space, rather than the entire "restore point". That's definitely something Veeam could see some improvement with if you want to add that as a feature request.

[BackupBytesTim]

If the recommended way to remove restore points is by reducing the retention policy, how's that work if they drive is full? In my experience Veeam deletes old restore points only after creating a new one, so if there's not enough room in a repository or a connected drive where the backup is stored, then old ones won't get deleted. I know this is a problem I have regularly on some computers regularly.

[HannesK]

Veeam deletes old restore points only after creating a new one

that's why I described the workaround above. Applying retention without running a job is an open feature request (it was discussed many times already on the forums)

Please don't de-rail topics. The customer had a clear question about merging restore points. This topic is not about "how to delete data from full disks" (where we already have many other topics). Thanks for your understanding