Any process upon which the Cluster Administrator is dependent can cause the Cluster Administrator to hang.

If any resource managed by the Cluster Administrator hangs, and/or the Cluster service or Cluster Administrator hangs, you need to stop the Cluster service to cause failover to the other Cluster node.

To force the Cluster service to stop, Use the Processes tab of Task Manager and and write down the PID of Cluster.exe. Use Kill.exe from Supplement 4 of the NT Server Resource Kit, or from the Windows 2000 Server Resource Kit, to 'kill' the PID, causing the Cluster to failover to the other node.