CODEWITHSUNDEEP
kubernetes
junerey
junerey

Reputation: 120

Kube Controller Manager CrashLoopBackOff

My kube-controller-managerkeeps staying on CrashLoopBackOff status.

I found this upon looking in the logs of the pod: failed to create listener: failed to listen on 0.0.0.0:10252: listen tcp 0.0.0.0:10252: bind: address already in use

Then I stumbled upon this article who fortunately was able to find a fix for it. Where he killed the process using the port and restarted his kube-controller-manager pod. https://medium.com/@deepeshtripathi/kubernetes-controller-pod-crashloopbackoff-resolved-16aaa1c27cfc

So I did follow the steps he made. When I have tried to get into the master node to find which process is using this port, I can't see anything that uses it.

root@ip:/# netstat -tunlp | grep 1025 
tcp6       0      0 :::10250                :::*                    LISTEN      1598/kubelet    
tcp6       0      0 :::10251                :::*                    LISTEN      7472/kube-scheduler
tcp6       0      0 :::10255                :::*                    LISTEN      1598/kubelet    
tcp6       0      0 :::10256                :::*                    LISTEN      5629/kube-proxy

Is there anyone else know any solution on how to fix this?

Upvotes: 3

Views: 6394

Answers (1)

hoque
hoque

Reputation: 6471

failed to create listener: failed to listen on 0.0.0.0:10252: listen tcp 0.0.0.0:10252: bind: address already in use

According to the error message port 10252 is in use. So need to stop listening on this port. You can do that by running

fuser -k 10252/tcp

Upvotes: 2

Related Questions