Missed full backups can cause permanent job failure

[filippoAdb]

Case #07216378
Don't think the following problem is due to misconfiguration but only to unfortunate circumstances and behaviors which are not easy to predict.

Incremental backup on Storeonce repository, one job for day, synthetic full on Saturday, 20 restore points.
It happens that Saturday backup job goes late and it is executed at the beginning of Sunday so synthetic full is not generated. I received no warning about the missing full backup.
The backup chains continues to grow and Storeonce has strict limitations, so after some days backup job fails persistently with the following error:
"Unable to allocate processing resources. Error: Failed to dispatch repository file sessions resource request. Requested amount of file sessions (25) is greater than the maximum amount available (24) ".

The only way to recover seems to be an active full backup which is very time consuming.

[Mildur]

Hi Filippo

Thank you for reporting.
The limit of restore points comes from the StoreOnce Appliance (concurrently open file limitation). If you miss a weekly full backup, the chain may get double amount of restore points as planned.
We have requests that we retry "full backups" after a failure instead of just continuing "incrementals" on the next day. Your use case would benefit from such request. Let me discuss it with the team.

Best,
Fabian

[filippoAdb]

It is clear that a skipped full backup will cause the failure of backup jobs soon when you have a StoreOnce repository. Unfortunately a skipped full backup doesn't generate any warnings, so it can happen without being noticed. In my specific case all the backup jobs are scheduled after each other, one of them performs an active full on Saturday evening and it lasts some hours till the beginning of Sunday. So the following backup job executes next day (after midnight) and doesn't perform the synthetic full scheduled on Saturday, but no warning or error is reported. I think there can be other unpredictable circumstances where the full backups are not executed so a warning or retry mechanism should be implemented.

[Mildur]

Understood.
Thank you for the additional feedback.

Best,
Fabian