Unable to delete pod stuck in terminating state

[BeyondEvil]

Not sure if it's a bug or if I'm doing something wrong.

But we have a pod in ArgoCD that's been stuck in "terminating" state for two days. Trying to force-delete gives below error.

The pod does indeed not exist in the cluster when checking with kubectl get pods.

What can I do?

[zhaoxiaomin1983]

I met the same issue. How can I do to fix it?

[Babbili]

delete it this way

kubectl delete pod/POD_NAME --force --grace-period=0

[BeyondEvil]

I don't know about @zhaoxiaomin1983 but in my case that wouldn't have worked (I guess), since the pod didn't exist in the cluster - only in ArgoCD.

The pod does indeed not exist in the cluster when checking with kubectl get pods.

[andrii-korotkov-verkada]

Maybe the cache needs to be invalidated. Normally, ArgoCD should get event from Kubernetes about the deletion and update the cache, but it could be slow on medium+ clusters. ArgoCD version 2.13 added some big optimizations, which help with event processing as well at least for medium clusters. I suggest giving it a shot.

[BeyondEvil]

Thanks for the tip. Will try!

[rumstead]

The phantom pods are a known bug :(. Here is a pretty detailed issue - #9226