Description

containerd/docker service not always starting after flatcar-upgrade

Impact

All nodes upgraded itself, but all of them came back in NotReady state, essentially killing the cluster.

Environment and steps to reproduce

1. Set-up: flatcar running rke
2. Action(s): waiting for nodes to upgrade themself.
3. Error: docker-service and containerd service not running.
4. fix: Running docker ps on the node started the service automatic.
5. observations: docker.service is set to disabled. So this might be expected?

Expected behavior
flatcar would be able to do automatic upgrades without ending up with all workers in NotReady state

Additional information
I'm not sure if the issue here is that the docker service is not enabled, or if there are something else. Running docker commands starts everything up and the node will be Ready. This is probably some torcx magic, but it seams not working as it should.