Providing SQL Server Failover Support requires that some key components and services work closely together. Clustering requires that two servers and a shared disk system be connected, configured, and managed. Cluster Administrator allows you to define a number of parameters, including failover/failback timing and services, and resource moves between servers.
Microsoft® Windows NT® Clustering Service architecture is designed to work with standard Windows NT Server storage drivers interfacing with SCSI devices, including SCSI over fiber. The storage device, typically a RAID system, must be connected to both nodes in the cluster for failover to work properly. Shared storage devices must be formatted using NTFS.
Identically configured hardware is not required for each node of a cluster; however, there are two exceptions. First, you cannot mix Alpha and Intel architecture processors in the same cluster. Second, you may not be able to mix hardware from different vendors and, in some cases, different models from the same manufacturer. Before attempting to mix hardware, check the Hardware Compatibility List (HCL), found at www.microsoft.com/ntserver, a qualified systems integrator, or the hardware manufacturer.