Hi all,
we have a running Gitlab version v15.11.13 (Chart version: 6.11.13) in Kubernetes 1.21.7. We are currently testing our backup and restore procedure. Usually we would use Velero but this doesn’t seem to be working with Gitlab. So we use the builtin backup / restore functionality from the toolbox. Backup is scheduled to run every day and this works just fine. But we are facing some problems with our restore. We have a second test K8s cluster where we test the restore. We do restore in a different namespace gitlab-test, while on the prod cluster we use gitlab. But I don’t think this plays any role. We simply deploy a fresh Gitlab Helm chart using the same secrets and settings as from the production, except that we disable the daily backup and we use a hostname suffix of staging2. The restore works fine but once we log in on the restored gitlab instance, we see that e.g. avatar pictures no more exist and we get a “500 error” when trying to access artifacts. The link for the avatar picture looks just fine. It shows the updated URL. Do we need to apply some additional magic to make this work? Do we miss an option to actually backup the avatar pics and artifacts?
Any help is highly appreciated.
Cheers,
Oliver