What Can I Do If the Cluster Console Is Unavailable After a Master Node Is Shut Down?

Symptom

After a master node is shut down, the cluster console is unavailable.

Procedure

The Cilium community does not remove the Cilium endpoint from the pod in the Terminating state. As a result, some requests are distributed to the stopped node, and the requests fail. Perform the following operations:

1. Run the following command to delete the pod in the Terminating state:

   kubectl get pods -nkube-system | grep Terminating | awk '{print $1}'|xargs kubectl delete pods -nkube-system

2. Run the following command to check whether any pod malfunctions:

   kubectl get pods -nkube-system

3. After several minutes, the cluster console works properly again.