[SOLVED] Reinstall node in cluster

Got it. Thanks.

Now, what is the correct procedure for rejoining the cluster? I restored the SSH keys and network configuration. I have the corosync configuration and authkey. I probably messed up the first time around, but I wasn't sure whether it would just start working if I restored those and restarted the aforementioned services. I tried joining the cluster manually, but that recognized the host as already being in a cluster, so I removed those files and tried again, but then it said "can't add existing node".
 
Looks like copying corosync.conf to /etc/pve was the missing piece, and the host will automatically rejoin the cluster after restarting the services. I also had to restore SDN configuration, as well as restart the pveproxy service in order to resolve certificate issues.

Here are the final (assuming I don't discover new issues) steps for restoring a cluster node, then:
  • Restore the following files:
    • /etc/network/interfaces
    • /etc/network/interfaces.d/sdn
    • /root/.ssh/id_rsa
    • /root/.ssh/id_rsa.pub
    • /etc/corosync/authkey
    • /etc/corosync/corosync.conf
    • /etc/pve/corosync.conf
  • Reload network interfaces (ifreload -a)
  • Restart services (systemctl restart corosync pve-cluster pveproxy)
 
You must log in or register to reply here.
Top Bottom
Back