Control plane node stuck in "provisioned" status as it cannot join the ETCD cluster as a member
TCA: 3.2, TCP: 5.0
The new node is blocked from joining etcd because an older node in the cluster has an expired certificate, breaking peer-to-peer communication. You can verify by connecting to the node directly as capv from TCA-CP (root) and checking the cloud-init-output.log
While connected to the VIP, identify the node the old control plane node that is not a ETCD follower using the below commandssudo -ialias etcdctl='/var/lib/containerd/io.containerd.snapshotter.v1.overlayfs/snapshots/*/fs/usr/local/bin/etcdctl --cert /etc/kubernetes/pki/etcd/peer.crt --key /etc/kubernetes/pki/etcd/peer.key --cacert /etc/kubernetes/pki/etcd/ca.crt'etcdctl -w table member list
Once the node is identified delete the machine, vspheremachines and vspherevm object from the management cluster
1. SSH to the management cluster
2. Delete the old control plane node that is not a ETCD member by deleting the aforementioned object
3. Delete the control plane node that is in provisioned state
This will start a new control plane node deployment process and resolve the issue