kubernetes-master upgrade missing kubeadm config upgrade steps
Closed this issue · 1 comments
/kind bug
What steps did you take and what happened:
Upgrade a cluster using wardroom, from 1.10, 1.11, 1.12, 1.13, 1.14
What did you expect to happen:
kubeadm upgrade orchestration be used to generate local /var/lib/kubernetes/config.yaml and kubelet-config-1.x config map in primary nodes, then download that from the config map on all other nodes
Currently the /var/lib/kubelet/config.yaml doesn't reflect what's in the kubeadm.conf file (specified in ansible inventory), and new kubelet-config-1.x configmaps don't get generated, which in turn makes node adds fail
Anything else you would like to add:
This process should also be executed if only the kubeadm.conf is modified, to incorporate any changes (e.g. alteration of evictionHard limits) in kubelet configs across the cluster.
Environment:
ubuntu 18.04
- Wardroom version:
branch
v1.14.x - OS (e.g. from
/etc/os-release
):
This is addressed by #215