Pod stay in ContainerCreating forever if failed to create sandbox

[liqlin2015]

What happened:

Kubelet failed to create sandbox for pod because of timeout when connect to docker engine. See the pod even

Failed create pod sandbox: rpc error: code = DeadlineExceeded desc = context deadline exceeded

Since it's not a permanent env error kubelet should retry to create sanbox. In my case, I just deleted the pod in ContainterCreating status and new pod was successfully created.

What you expected to happen:

In some error cases kubelet should retry to create pod. Maybe mark pod as failed instead of ContainerCreating.

How to reproduce it (as minimally and precisely as possible):

Not easy to reproduce.

Anything else we need to know?:

Environment:

• Kubernetes version (use kubectl version): v1.13.7
• Cloud provider or hardware configuration: None
• OS (e.g: cat /etc/os-release): Ubuntu
• Kernel (e.g. uname -a): 18.04
• Install tools: None
• Network plugin and version (if this is a network-related bug): Calico
• Others:

[liqlin2015]

/sig node

[Pluies]

FWIW I'm running into the same issue, and it appears very similar to issue #79398 and #78125 and, if I understand correctly, should be fixed by #79451.

[liqlin2015]

Close it based this PR #79451