kubernetes: api-server and controller-manager cant start
The problem is simply that some service is already bound on 6443 to check that out you can use netstat -lutpn | grep 6443 and kill that process and restart kubelet service.
$ netstat -lutpn | grep 6443tcp6 0 0 :::6443 :::* LISTEN 11395/some-service$ kill 11395$ service kubelet restartThis should fix the situation.
With kubernetes this usually happens if the kubernetes is not properly rested and containers are not properly cleaned out.
To do so...
$ kubeadm reset$ docker rm -f $(docker ps -a -q)$ kubeadm init <options> # new initializationWhich would mean the nodes will have to rejoin again.