<p>When a first server node device fails in a cluster system, a client node device transmits failure detection information to a second server node device. When receiving the failure detection information, the second server node device transmits a survival confirmation request to the first server node device. When not receiving a survival confirmation response from the first server node device, the second server node device judges that the first server node device fails and starts switching control of a server node device which executes a work processing. When receiving failure detection information, the second server node device starts switching control even if it further receives failure detection information from another client node device.</p>