Hibernating a cluster will fail as the Karydia pod/pods cannot be stopped. This is due to changes in Kubernetes regarding leader elections of the controller for endpoints (https://github.com/kubernetes/kubernetes/pull/84084).
Steps to reproduce
Create a cluster with Kubernetes version 1.17.x
Hibernate the cluster (e.g. via the Gardener UI)
Expected behavior
The cluster should go into hibernation mode and be easily woken up afterwards.
"Hibernating control plane" failed: retry failed with context deadline exceeded, last error: retry failed with context deadline exceeded, last error: waiting until there are no running Pods in the shoot cluster... there is still at least one running Pod in the shoot cluster: karydia/karydia
Description
Hibernating a cluster will fail as the Karydia pod/pods cannot be stopped. This is due to changes in Kubernetes regarding leader elections of the controller for endpoints (https://github.com/kubernetes/kubernetes/pull/84084).
Steps to reproduce
1.17.xExpected behavior
The cluster should go into hibernation mode and be easily woken up afterwards.
Logs / console output / screenshots / affected lines of code
Error in Gardener:
Environment
kubectl version): 1.17.2