HP StorageWorks 9000s NAS 4000s and 9000s Administration Guide - Page 197

Cluster Administration 10 One important feature of the HP StorageWorks NAS server is that it can operate as a single node or as a cluster. This chapter discusses cluster installation and cluster management issues. Some of these topics are discussed or mentioned elsewhere in this guide. The discussion in this chapter is more detailed than other references and addresses the unique administration procedures for operating in a clustered environment. Cluster Overview As introduced in chapter 1, two server heads (nodes) can be connected to each other and deployed as a no single point of failure (NSPOF) dual redundant cluster. The nodes are connected by a crossover cable and are each connected to network switches or hubs. This connection allows communication between the nodes to track the state of each cluster node. Each node sends out periodic messages to the other node; these messages are called heartbeats. If a node stops sending messages, the cluster service will fail over any resources that the node owns to the other node. For example, if the node that owns the Quorum disk is shut down for any reason, its heartbeat will stop. The other node detects the lack of the heartbeat and takes over ownership of the Quorum disk and the cluster. Multi Node Support Beyond Two Nodes The NAS 4000s and 9000s devices may be deployed in multi node clustering beyond two nodes. Refer to the associated Storage Array documentation to determine the number of nodes supported by the array under Windows Storage Server 2003. While the discussion presented in this guide addresses only two nodes, additional nodes may be added into the cluster. Considerations for additional fiber path connections and the private network should be made. In the case of the private network, a hub or switch is required since the cross over cable is no longer applicable. NAS 4000s and 9000s Administration Guide 197