Once you have gone through the installation process of Cluster Service, the final step is to verify that the installation was indeed a success. There are several methods you can use to verify that the service is up and running on your cluster nodes.

• Open the Registry on the cluster node and verify that the “Cluster” key has been created under the HKEY_LOCAL_MACHINE subtree.

• Under the Administrative Tools, open the Cluster Administrator to verify that the cluster disks are online. If the status of the disks is set to “online”, it means that the Cluster Service has been configured correctly and is working properly.
• Once Cluster Service has been installed on both cluster nodes, use the Cluster Administrator utility to verify that bother nodes are present. If both nodes are listed, the cluster is operational.

Clusters are designed to provide high availability for network resources, applications, and network services. If one cluster node fails, the other node will continue to service client requests. They are not capable of protecting all components however. The availability of a cluster is dependent on the configuration of the network environment.

In order to provide maximum availability of the cluster and its resources, you need to identify all single points of failure. A single point of failure can include both software and hardware. For example, if IP addresses are dynamically assigned to cluster nodes, their availability depends on the availability of a DHCP server. Some of the common points of failure (and their solutions) are outlined below.

DHCP: Assign static IP addresses to all network interfaces
Disks: Implement hardware or software level RAID for the private disks. Use hardware level RAID for the shared disk(s).
Hardware failure: Always ensure there are spare parts on hand.
Power outages: Use UPS devices to provide power in case of power failure.
Router, hubs, etc.: Have spare components on hand. Implement redundant routes.
Network cards: Configure each cluster node with at least two network adapters.
Authentication: Make the cluster nodes domain controllers in an existing domain or in their own domain.

Tags: diana huggins, cluster service, windows 2000 advanced server, scsi, fiber channel