In a cluster, each application that fails over independently of other applications should have its data stored on volumes in disk groups exclusive to that application.

In a clustered environment, the disk group is the unit in which storage fails over from one computer to another. Only entire disk groups fail over. Thus, volumes that hold data for applications that are required to fail over should belong to disk groups that hold data for that application only. This allows an application's storage to fail over with the application and have no adverse effects on other applications or their associated storage. The disk groups should also be part of the application's resource group, so that failover can occur.