Or rename the old cluster/nodes to "name_old" or something, rename the new cluster/nodes, and make sure it works before taking the old stuff all offline.