2.6 Cluster Definition

According to HP specifications, the TruCluster Server V5.1A for Tru64 UNIX "provides highly available and scalable solutions for mission-critical computing environments." The specification (Quick Specs for TruCluster Server) goes on to describe a combination of features from SMP, distributed computing, and fault resilience that provide very high availability and scalability.

So a cluster is a mechanism whereby a closely coupled group of computers can combine their resources and present itself to an application, administrator, or even to the network as a single, highly available machine. The TruCluster Server product attempts to eliminate as many points of failure as possible in order to provide consistent access to a company's critical resources. It also provides administration mechanisms through which cluster administration becomes only slightly more complex than single node administration. In a cluster, product installations, system monitoring, and most other management tasks can be done a single time, rather than once for each member of the cluster. A cluster can appear as a single node in a network if desired.

In the remainder of this chapter, we will introduce you to the cluster components that provide the mechanisms for making the above-mentioned features happen.