There should be a method in remote cache manager that allows a client to manually switch to a different cluster. This can be useful after a client has failed over to a secondary cluster, have the client switch back to the primary cluster.