-
Notifications
You must be signed in to change notification settings - Fork 0
Description
What happened:
Other controllers, like the PlatformService
DNS
for example, might reconcile Cluster
resources and deploy stuff on them which they need to properly remove when the Cluster is being deleted. This is currently not possible, because the ClusterProvider removes the cluster as soon as the Cluster
resource has a deletion timestamp, thereby preventing other controllers from cleaning up, which again might lead to the Cluster
becoming stuck in deletion, because the other controllers will never remove their finalizers.
What you expected to happen:
The Cluster controller needs to wait with deleting the cluster until all unknown finalizers have been removed from the Cluster
resource.
How to reproduce it (as minimally and precisely as possible):
Run any controller that uses finalizers on Cluster
resources and needs some time to cleanup resources on the cluster when it is being deleted.