3

We are currently trying to launch a pod on our kubernetes cluster. The pod get stuck in the Waiting state (ContainerCreating reason) forever.

When getting details about the pod (kubectl describe pod pod01), the only event that's present is:

Events:
  Type     Reason      Age                     From     Message
  ----     ------      ----                    ----     -------
  Warning  FailedSync  2m29s (x439 over 125m)  kubelet  error determining status: rpc error: code = DeadlineExceeded desc = context deadline exceeded

From the various posts I found on the web, it can possibly mean that kubelet timing out while trying to contact etcd.

Is that correct? How could I troubleshoot that further?

Thanks in advance,

Emmanuel

E. Jaep
  • 2,095
  • 1
  • 30
  • 56
  • Have you checked kubelet logs? – Andromeda Mar 29 '23 at 09:54
  • Try checking the kubelet logs of the node where the pods try to schedule ```journalctl -u kubelet```. Then *cordon* the node, making it UNSCHEDULE and try to deploy the Pod again; we expect it on the other nodes to work as you say. https://kubernetes.io/docs/reference/kubectl/cheatsheet/#interacting-with-nodes-and-cluster Finally restart the node in a safe way for your applications, by performing a *drain* first. – glv Mar 29 '23 at 09:56
  • @glv after a cordon/drain/reboot/uncordon, I get quite a lot of " kubelet... transport: loopyWriter.run returning. connection error: desc = "transport is closing"" which, if I'm correct, not to worry about – E. Jaep Mar 29 '23 at 11:27
  • Well, I hope the problem is gone! – glv Mar 29 '23 at 12:08

0 Answers0