Incremental and full Synthetic backups

[Perry]

How reliable are synthetic backups over active full backups?

For example if there is an interruption or something went wrong during an incremental backup, I assume all incremental backups from that point forward will be corrupt. Since the synthetic full backup is built from those restore points, I imagine that synthetic full backup will also be bad.

So what is stopping a bad incremental from corrupting all forward backups and what kind of integrity checks are preformed so we don't end up with a bad chain of bad restore points?

Some how I feel a little safer with differential backups that don't rely on previous points, but I don't see that offered. Can you put my fears to rest?

[Shestakov]

Hello Perry,
Indeed, corruption in a single increment can cause a corruption of the further backups of the chain, that`s the main reason to schedule active full.
But the best way to make sure the backups are recoverable is Surebackup.
I would also recommend to read this topic. Thanks

[Vitaliy S.]

Hi Perry,

Every time synthetic backup is built it runs an incremental backup and then starts the transform operation. This topic might be helpful as well > How often do I need a full backup?

So what is stopping a bad incremental from corrupting all forward backups and what kind of integrity checks are preformed so we don't end up with a bad chain of bad restore points?

Regular health checks should help > Health Check for Backup Files. Alternatively, consider using SureBackup jobs as Nikita has correctly pointed it out.

Thanks!

[Perry]

If a gen 2 incremental fails. Will the next gen 3 attempted incremental be based off the failed gen 2 or will it be based on the last successful one which would be gen 1? I assume the latter otherwise from the failure point forward everything would be broken.

[Shestakov]

If the backup job fails it performs retries until backup is done successfully.
Backup file corruptions may happen without reference to the backup job and if you don`t use recovery verification the corruption may interfere to the next increments of the chain.
That`s why it`s recommended either use Surebackup or periodic active full backups.

[Vitaliy S.]

Will the next gen 3 attempted incremental be based off the failed gen 2 or will it be based on the last successful one which would be gen 1? I assume the latter otherwise from the failure point forward everything would be broken.

Yes, it's the latter.

[Perry]

Shestakov,
So are you saying there are scenarios in which Veeam will think the backup went fine (thus not notify error) yet the backup could still be bad thus corrupting future incrementals? I imagine that I would at least be notified of a corruption so I can take action.

[Vitaliy S.]

Not exactly. Shestakov is referring to storage-level corruption which can occur unnoticed to a Veeam backup server. In this case, SureBackup job is the recommended solution which can notify you that something has happened to your hardware. If you want to address this, then using backup copy jobs is a way to go. Here is our official blog post for further reading > The 3-2-1-0 Rule to High Availability