Address stuck rolling update when subnet CIDR do not have usable addresses

gardener / machine-controller-manager

Declarative way of managing machines for Kubernetes cluster

Apache License 2.0

256 stars 117 forks source link

Address stuck rolling update when subnet CIDR do not have usable addresses #924

Open elankath opened 3 months ago

elankath commented 3 months ago

How to categorize this issue?

/area robustness /priority 3

What would you like to be added:

Rolling update can get stuck when the available subnet CIDR's don't have sufficient usable addresses for launching machines. See canary issue 5348. We should aim to avoid or mitigate this situation

Why is this needed: Prevent necessity of manual operator intervention. Let the customer fix their shoot cluster.

gardener-robot commented 3 months ago

@elankath You have mentioned internal references in the public. Please check.