following on from the above: is it possible/supported to retrigger the cluster join process from a host that has failed to join the cluster, or does it require a reimage?