Cluster goes into bad state if a restore fails #1990
Description
Report
We were testing restoring a backup, and for some reason the restore failed (separate issue). I thought perhaps I could delete the restore to abort it, or maybe the cluster would give up and go back to normal on its own. However, it did not. I didn't see any documentation on how to abort a restore from backup. I tried deleting the psmdb resource, which did not work on its own, I had "help it" to delete the statefulset before it would actually finish deleting the psmdb so I could create it again.
More about the problem
...
Steps to reproduce
- Create a cluster with just a replset
- Attempt to restore in some way that fails (e.g. using requireTLS in our case)
- See what you can do with the cluster after restore fails
Versions
- Kubernetes v1.30.12
- Operator v1.20.1
- Database mongodb
Anything else?
CR YAML: https://gist.github.com/dobesv/c2727a9ee382ce80638d61bd0d64ca30