Rescue cluster after accidentally master nodes


I have a cluster with rancher v2.4.7 and I erased accidentally the data of my master nodes. I deleted all that was related to docker on the VMs.

Fortunaltely, my worker nodes are still up and they still do their job properly. But my cluster is stuck in “Updating” status.

I managed to delete from the UI the master nodes that weren’t there anymore. But when I try to add a new master node, I got this message :

Failed to delete controlplane node [xx.xx.xx.xx] from cluster: Get https://xx.xx.xx.xx:6443/api/v1/nodes: dial tcp xx.xx.xx.xx:6443: connect: connection refused; waiting for etcd and controlplane nodes to be registered

And the registration process is not completed.

The IPs are the ones of my masternodes that aren’t there anymore, despite that I removed the nodes as advised in
Cluster Stuck in Updating State when Nodes are Missing During Deletion · Issue #19916 · rancher/rancher · GitHub but it seems there is still some traces of my old masternodes somewhere that prevent me from adding new one.

Any idea ?

PS : The new masternode that I’m trying to add has the same IP that the old one.