Successful restore does not trigger all the related events

[Zvirovyi]

This is the same bug I've encountered in VM version of this charm.

Steps to reproduce

1. Deploy postgresql-k8s with s3-integrator
2. Create full backup (juju run postgresql-k8s/leader create-backup)
3. Restore it (juju run postgresql-k8s/leader restore backup-id=<your_value>)

Expected behavior

After successful restore, cluster should print message "Restore succeeded" in the debug-log and Patroni on-failure condition should be restored.

Actual behavior

Cluster appears in the "Move restored cluster to another S3 bucket" and it indicates, that restore process is finished. But there is no "Restore succeeded" message in the debug-log and Patroni on-failure condition isn't really restored - juju show-unit postgresql-k8s/0 shows related fields:

overridden-patroni-on-failure-condition-repeat-cause: restore-backup
patroni-on-failure-condition-override: ignore

Overall, it leads to the conclusion, that successful restore does not trigger all its events and does not restore cluster to the fully working state.

Versions

Operating system: Ubuntu 24.04 LTS

Juju CLI: 3.5.3-genericlinux-amd64

Juju agent: 3.5.3

Charm revision: 349

microk8s: MicroK8s v1.30.1 revision 6843

Log output

Juju debug log: k8s_is_restore_finished_backupid.log

Additional context

Although I can surely tell what commit exactly caused this issue in the VM version (see reference at the top), here PITR was implemented with these changes all from the beginning. I think this issue will be critical when working on Timelines management for the PITR.

[github-actions]

https://warthogs.atlassian.net/browse/DPE-5128

[Zvirovyi]

#690 fixed this issue