Problem Description:
In regards to an install or reinstalling konvoy, even after cleaning up (./konvoy down) and reinstalling konvoy (./konvoy up), the first control plane gets created successfully, but additional control plane nodes can’t join the cluster
In regards to an install or reinstalling konvoy, even after cleaning up (./konvoy down) and reinstalling konvoy (./konvoy up), the first control plane gets created successfully, but additional control plane nodes can’t join the cluster
Analysis:
This is because of an existing ~/.kube/config file on a control plane. More likely from when a user might have accidentally ran ./konvoy from a cluster node. If that happens, kubeadm tries to use the default kubeconfig (~/.kube/config), rather than the desired kubeconfig for bootstrapping the cluster (/etc/kubernetes/admin.conf).
NOTE: ./konvoy up shouldn’t be run from a cluster node.
Solution:
In the event you encounter the error, ssh into the node where the existing ~/.kube/config file is residing then you can either remove the file (rm ~/.kube/config) or, rename it;
cd ~/.kube/
mv config
Then ./konvoy down to clean possible remnants and make the required changes to cluster.yaml (if there’s any), then, ./konvoy up .