So far the Documentation regarding Restic is clear
For Reference: https://velero.io/docs/v1.0.0/restic/
But i miss some Feature/Information.
Let's start with what i have done already:
I've added the annotation for the pods in the deployment:
spec:
template:
metadata:
annotations:
backup.velero.io/backup-volumes: volume-name-to-backup
I triggered an backup:
velero backup create berndklaus-at --include-namespaces berndklaus-at
When performing:
velero restic repo get
I can see a repo was added, there was no repo before performing the velero backup command.
velero backup get
I can see a backup, state: completed
OK, thats fine.
Well now i wont to take a newew backup of the above things, so i fired:
velero backup create berndklaus-at --include-namespaces berndklaus-at
That does not work, cause the backup still exists.
Ok, i delete the backup "berndklaus-at"
But when looking into my spaces, i can still see the restic part is NOT removed.
So i tought, there should be a command like: "velero restic backup delete." But i did not find a way to remove a restic backup.
So there came up a few questions:
What does the restic repo represent?
How can i manage restic backups (delete/recreate/...)?
How can i trigger a restic backup only?
Does a schedule also starts a restic backup?
Please excuse typos and/or bad english.
Thanks for clearification, BR Bernd
@Berndinox good questions!
What does the restic repo represent?
Restic uses "repositories" to organize backup data. Velero creates one per namespace + backup storage location.
How can i manage restic backups (delete/recreate/...)?
In velero, restic backups are always tied to a main "velero backup". When you delete a velero backup, we run the restic forget command which deletes metadata about the restic backup from the repository. Separately, there's a recurring restic prune operation that runs for each repository every 24h by default, that deletes the actual blocks of data. We do this asynchronously because it can be slow.
How can i trigger a restic backup only?
Velero does not support this - all restic backups are tied to a main "velero backup".
Does a schedule also starts a restic backup?
Yes, if you have workload pods annotated!
Ok, thanks alot for clearification!
Closing this issue out.
Most helpful comment
@Berndinox good questions!
Restic uses "repositories" to organize backup data. Velero creates one per namespace + backup storage location.
In velero, restic backups are always tied to a main "velero backup". When you delete a velero backup, we run the
restic forgetcommand which deletes metadata about the restic backup from the repository. Separately, there's a recurringrestic pruneoperation that runs for each repository every 24h by default, that deletes the actual blocks of data. We do this asynchronously because it can be slow.Velero does not support this - all restic backups are tied to a main "velero backup".
Yes, if you have workload pods annotated!