0

I'm trying to upgrade a Kubernetes multi-master cluster from v1.8.5 to v1.9.1 on Ubuntu 1604 LTS.

When I run sudo kubeadm upgrade apply v1.9.1 it gets as far as backing up the old manifest for kube-apiserver.yaml and then displays this:

[upgrade/staticpods] Waiting for the kubelet to restart the component

It then waits for 60 seconds (apparently a hard-coded time-out in kubeadm) before restoring the original state.

There's nothing obvious in the logs. What could be the issue?

swref
  • 111
  • 4

1 Answers1

1

It turned out to be a problem with the single nodeName in the kubeadm-config being hard-coded to the name of one of the other masters.

I edited it with this command:

kubectl edit cm -n kube-system kubeadm-config -o yaml

and changed the nodeName in the MasterConfiguration section to match the name of the master being upgraded, and then it worked.

swref
  • 111
  • 4