I recently issued the “juju add-unit kubernetes-worker” command on our Charmed cluster and saw a new machine get spun up and eventually enter the active idle state. However I can’t get any new workloads to deploy there and when I issue “kubectl get nodes” the new node isn’t showing in the list. When I issue “juju debug-log” I see the following every couple seconds:
unit-kubernetes-worker-7: 14:52:51 WARNING unit.kubernetes-worker/7.update-status Error from server (NotFound): nodes “juju-eb259e-46” not found
This is the new worker that was added and not showing. Previously I’ve been able to add new workers without a hitch but I can’t seem to figure out what’s going on at the moment.
#Update
When I issue “juju show-machine [id]” on this worker versus one that is working and has pods, the difference is that there is no flannel.1 listed in the network-interfaces for the bad worker. I compared the /etc/cni permissions as well and noticed that the bad worker has only 700 for the etc/cni folders. Changed this and rebooted and still no luck yet.
Thanks for any advice, m