After clean Install: Port occupied about ansible-kubernetes-openshift-pi3

[rhuss]

Sorry, the current check for whether Kubernetes is running or not is a bit limited. It used kubectl get nodes and if this fails with error code 1 then it is assumed that no cluster is runnin gand kubeadm init is called again.

I think this should be made more robuts. Any idea ? (maybe we should run kubeadm upgrade plan or so ....)

[dklueh79]

Any solution for completing your kubernetes setup?

[rhuss]

@dklueh79 wdym ? Actually, the current detection works when Kubernetes has been properly installed and nodes are running. So in this case the kubeadm init step is skipped. However when the initial setup didnt work and you are in a half-baked state, the kubectl get nodes fails, but kubeadm init also fails. One should properly do a full reset then.

So you should try a full reset when this error ocurs for you before trying again:

ansible-playbook -i hosts kubernetes-full-reset.yml