Comment 21 for bug 1856078

The root cause of this issue is in the upstream tiller project. There is a planned stx.4.0 activity to upversion helm to v3 which will remove tiller and prevent this issue from occurring: https://storyboard.openstack.org/#!/story/2007000

Until that story is implemented please use this workaround:

If helm commands fail with the following error then a restart of the tiller pod will be required:
controller-0:~$ helm ls -a
Error: the server could not find the requested resource (get configmaps)
To restart the tiller pod use the following:
controller-0:~$ kubectl get pods --all-namespaces -o wide|grep tiller-deploy
kube-system tiller-deploy-d6b59fcb-g676l 1/1 Running 1 18h fd01:1::3 controller-0 <none> <none>

controller-0:~$ kubectl delete pods -n kube-system tiller-deploy-d6b59fcb-g676l

Confirm tiller was restarted:
controller-0:~$ kubectl get pods --all-namespaces -o wide|grep tiller-deploy