kubelet update stopped worker nodes

Bug #1830829 reported by Xav Paice
12
This bug affects 2 people
Affects Status Importance Assigned to Milestone
Kubernetes Worker Charm
New
Undecided
Unassigned

Bug Description

We found that all the worker nodes on the cloud were status NotReady. I tried restarting one, which fixed it, using 'snap restart kubelet'. Here's the output after:

ubuntu@juju-1b688f-somecloud-k8s-5:/var/log/juju$ kubectl get nodes
NAME STATUS ROLES AGE VERSION
juju-1b688f-somecloud-k8s-10 NotReady <none> 29d v1.13.5
juju-1b688f-somecloud-k8s-11 Ready <none> 29d v1.13.6
juju-1b688f-somecloud-k8s-12 NotReady <none> 29d v1.13.5
juju-1b688f-somecloud-k8s-7 NotReady <none> 29d v1.13.5
juju-1b688f-somecloud-k8s-8 NotReady <none> 29d v1.13.5
juju-1b688f-somecloud-k8s-9 NotReady <none> 29d v1.13.5

Restarted all the kubelets:
ubuntu@juju-1b688f-somecloud-k8s-5:/var/log/juju$ kubectl get nodes
NAME STATUS ROLES AGE VERSION
juju-1b688f-somecloud-k8s-10 Ready <none> 29d v1.13.6
juju-1b688f-somecloud-k8s-11 Ready <none> 29d v1.13.6
juju-1b688f-somecloud-k8s-12 Ready <none> 29d v1.13.6
juju-1b688f-somecloud-k8s-7 Ready <none> 29d v1.13.6
juju-1b688f-somecloud-k8s-8 Ready <none> 29d v1.13.6
juju-1b688f-somecloud-k8s-9 Ready <none> 29d v1.13.6

Before the restart, the logs were full of this:

May 29 03:50:53 juju-1b688f-somecloud-k8s-10 kubelet.daemon[30866]: I0529 03:50:53.872978 30866 kubelet_node_status.go:279] Setting node annotation to enable volume controller attach/detach
May 29 03:51:03 juju-1b688f-somecloud-k8s-10 kubelet.daemon[30866]: I0529 03:51:03.889760 30866 kubelet_node_status.go:279] Setting node annotation to enable volume controller attach/detach
May 29 03:51:13 juju-1b688f-somecloud-k8s-10 kubelet.daemon[30866]: I0529 03:51:13.907342 30866 kubelet_node_status.go:279] Setting node annotation to enable volume controller attach/detach
May 29 03:51:23 juju-1b688f-somecloud-k8s-10 kubelet.daemon[30866]: I0529 03:51:23.924960 30866 kubelet_node_status.go:279] Setting node annotation to enable volume controller attach/detach
May 29 03:51:33 juju-1b688f-somecloud-k8s-10 kubelet.daemon[30866]: I0529 03:51:33.944682 30866 kubelet_node_status.go:279] Setting node annotation to enable volume controller attach/detach

Restarted kubelet:

May 29 03:51:40 juju-1b688f-somecloud-k8s-10 systemd[1]: Stopping Service for snap application kubelet.daemon...
May 29 03:51:40 juju-1b688f-somecloud-k8s-10 systemd[1]: Stopped Service for snap application kubelet.daemon.
May 29 03:51:40 juju-1b688f-somecloud-k8s-10 systemd[1]: Started Service for snap application kubelet.daemon.
May 29 03:51:40 juju-1b688f-somecloud-k8s-10 kubelet.daemon[21661]: I0529 03:51:40.371776 21661 controller.go:101] kubelet config controller: starting controller
May 29 03:51:40 juju-1b688f-somecloud-k8s-10 kubelet.daemon[21661]: I0529 03:51:40.371996 21661 controller.go:267] kubelet config controller: ensuring filesystem is set up correctly
May 29 03:51:40 juju-1b688f-somecloud-k8s-10 kubelet.daemon[21661]: I0529 03:51:40.372017 21661 fsstore.go:59] kubelet config controller: initializing config checkpoints directory "/root/cdk/kubelet/dynamic-conf
May 29 03:51:40 juju-1b688f-somecloud-k8s-10 kubelet.daemon[21661]: I0529 03:51:40.378500 21661 server.go:407] Version: v1.13.6
May 29 03:51:40 juju-1b688f-somecloud-k8s-10 kubelet.daemon[21661]: W0529 03:51:40.378650 21661 plugins.go:118] WARNING: openstack built-in cloud provider is now deprecated. Please use 'external' cloud provider¤
May 29 03:51:41 juju-1b688f-somecloud-k8s-10 kubelet.daemon[21661]: I0529 03:51:41.407589 21661 controller.go:207] kubelet config controller: local source is assigned, will not start remote config source informe
May 29 03:51:41 juju-1b688f-somecloud-k8s-10 kubelet.daemon[21661]: I0529 03:51:41.407780 21661 controller.go:197] kubelet config controller: starting status sync loop
May 29 03:51:41 juju-1b688f-somecloud-k8s-10 kubelet.daemon[21661]: I0529 03:51:41.407809 21661 status.go:145] kubelet config controller: updating Node.Status.Config
May 29 03:51:41 juju-1b688f-somecloud-k8s-10 kubelet.daemon[21661]: I0529 03:51:41.407809 21661 controller.go:226] kubelet config controller: starting Node informer
May 29 03:51:41 juju-1b688f-somecloud-k8s-10 kubelet.daemon[21661]: I0529 03:51:41.407828 21661 controller.go:231] kubelet config controller: starting Kubelet config sync loop
May 29 03:51:41 juju-1b688f-somecloud-k8s-10 kubelet.daemon[21661]: I0529 03:51:41.426329 21661 watch.go:89] kubelet config controller: initial Node watch event
... and so on.

The snap refresh policy is mon5, I can't see anything in 'snap changes' other than the service restart, but I do know that there was a recent release in snaps and we just hit past a last Monday in the month.

To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.