Hi All, I have old version of rancher running ( v2.6.3 ). The downstream cluster was running fine, but now is stuck with message cluster agent not connected and I can’t connect to cluster either through UI or by using rke2.yaml on one of the control nodes of downstream cluster. The downstream cluster is 5 node cluster with 3 control plane nodes (N1, N2 and N3) and 3 worker nodes with N3 node acting as both control plane and worker. The downstream cluster has gone into split brain with no etcd communication between N2 and N3 and rke2 service not running on N1. Is there a way to recover connectivity between rancher and downstream cluster without affecting the workloads ? Really appreciate any pointers to recover from this situation.

what distribution and version is your downstream cluster

Thanks. It’s running rke2 with v1.21.9+rke2r1. I can only access N2 node and kubectl is not working. So I am not sure how to find out the version of downstream cluster