Retention not working

[Bjoern_ch]

Hi veeam team,

I have 2 repos which both have the same issue.
Retention is not applied properly. In the logs there are no errors at all. Support cannot find anything. And they tell me to be patient or to reduce the number of restore points.
But the thing is: We try to cut off only the oldest restore point. We adjust the retention before the retention task starts. Meaning if the oldest restore point is 500 days old we set the retention to 499 days. That is the smallest chunk possible. You cannot reduce it anymore.
With both repos retention tasks were running for up to 5 days but the tasks still did not finish. If I can create a restore point every 24 hours then my expectation is that I can remove old restore points every 24 hours. I mean, otherwise, you cannot offer a 24h rpo with it.
Ah, and you need to disable the backup jobs while retention is running otherwise the backup job will kill the retention task. Right now, I have the option that retention is infinite or that my customers have no backups for some weeks (support estimated a time frame of 17 days to finish for the retention task for one of the repos)

So, has anybody any idea how to solve this? (without waiting weeks for the task to finish)
Thanks!

Case # 07906161
Bjoern

[karsten123]

are they object storage? maybe some sort of block generation?

[edh]

Hi,

What S3 Backend do you have in your VBO?

[Bjoern_ch]

yes, object storage. the buckets reside on two different on-prem S3 systems. One is ceph, the other is cloudian.
We have many more buckets on both systems which work perfectly fine.
Both buckets are not small. One is 75 TB (mail data only/13000 user). The other is 120 TB (onedrive data only/2000 user)

[edh]

Well we got something similar but we detected earlier.

In our case, we create a repository (bucket) per customer, and change the retencion cleanup from daily to monthly at non-working backup hours.

As deletion is a blocking operation, we try to download everything at night and cleanup during the mornings. Ofcourse your system right know got a lot to cleanup an block operations