Hi Bo,
many thanks for your reply and my apologies here, but am still rather confused.
I was testing the PVs restore without the helper added/configured (velero integrated with minio (using AWS plugin) with restic and it was OK:
- I scaled down the app deployment config to zero
- deleted the app pod PVC, which then triggered the auto PV delete
- run velero restore create from taken backup file with including the relevant namespace and PV and PVC only (--include-namespaces testapp --include-resources persistentvolumeclaims,persistentvolumes )
The restore run and completed fine, the PV and PVC were brought back. After scaling the deployment config back to original replicas count, pod was working.
So the main question is, if the restic restore helper is mandatory when using velero with restic. And how to see/test it properly when we add it.
I created the documented config map file to enable the helper, but saw no effect on the restoration process (maybe the restore was too fast).
I added at the end the helper docker image as a plugin via velero CLI directly, the velero pod was restarted and the velero logs were listing image pull, but again no direct visibility that it is active (the number of velero and all restic pod containers is 1 only).
Is .velero directory present on the source (minio dir structure) during restoration process only? I was not able to find it on subdirectory under restic directory structure.
Many thanks for your advices/help
Kind regards,
Peter