Changing quorum_dev_poll paramater in RHEL6 - cluster restart

Hi,
I am making some changes to quorum_dev_poll parameter in RHEL6 on a 2 node cluster and I know this requires restart of cluster.
Question is, can I freeze my managed application service on the node, stop and restart cluster services on that node (cman and rgmanager) to maintain service to my application , and then stop and start the cluster services on second node, or must I stop my application service, stop the cluster services on all nodes at same time, then start them all, before starting my managed service.
I am not clear on whether the cluster can be restarted on one node at a time, to maintain service of my app or whether I need to restart cluster services on all nodes at same time ( I recall needing to do this on all nodes at same time in earlier releases and not independently).

thanks in advance