Many DBAs seem to have difficulty understanding exactly what clustering is. What do we (and Microsoft) mean when we refer to SQL Server 2008 clustering? Here's a good working definition:
Microsoft SQL Server 2008 clustering is a high-availability option designed to increase the uptime of SQL Server 2008 instances. A SQL Server 2008 cluster includes two or more physical servers, called nodes, identically configured. One is designated as the active node, on which a SQL Server 2008 instance is running in production, and the other is an inactive node, on which SQL Server is installed but not running. If the SQL Server 2008 instance on the active node fails, the inactive node becomes the active node and continues SQL Server 2008 production with minimal downtime.
This definition is straightforward and to the point, but it has a lot of implications that are not so clear, and this is where many clustering misunderstandings arise. One of the best ways to more fully understand what clustering can and cannot do is to drill down into the details.
The benefits of using SQL Server 2008 clustering are very specific. Clustering is designed to boost the availability of physical server hardware, the operating system, and the SQL Server services. Should any of these aspects fail, an instance of SQL Server 2008 fails. If any of these three fails, another server in a cluster can automatically be assigned the task of taking over ...