[ws-manager] fix calling dispose multiple times

[sagor999]

Description

Correct fix for sometimes calling dispose multiple times, which can result in incorrect error about backup failed error.

Related Issue(s)

Part of #9475
Fixes #9598
Fixes #8326

How to test

Start and stop workspaces. They should always shutdown correctly now, previously sometimes they would error out with failed backup error due to race condition.

Release Notes

[ws-manager] fix sometimes workspaces fail with backup not found error

Documentation

[csweichel]

@sagor999 why did this even happen? The finalizer map should have prevented that in the first place. Does ws-manager restart?

Also, this really needs tests. There's an actOnPodEvent test which should capture this behaviour in a fixture. I'm somewhat surprised that there are not fixture changes.

[sagor999]

@csweichel I think it is a race issue.
Finalizer map is cleaned up when doFinalize has finished, if after that something triggers actOnPodEvent, it will attempt to run finalizer again. Hence adding annotation to make sure nothing will trigger it attempting to run finalizer.

Let me know if there is a better way to fix this.