Getting Started with SQL Server 2008 Failover Clustering

SQL Server failover clustering provides high-availability support for an entire SQL Server instance.  SQL Server failover clusters are built on top of Windows Server failover clusters. To create a SQL Server failover cluster, you need to first create the underlying Windows Server failover cluster.

A SQL Server failover cluster, also known as a failover cluster instance, consists of the following:

• One or more Windows Server failover cluster nodes

• A cluster resource group dedicated for the SQL Server failover cluster which contains the following:

  • Network name to access the SQL Server failover cluster

  • IP addresses

  • Shared disks used for the SQL Server database and log storage

• Resource DLLs that control the SQL Server failover behavior

• Check-pointed registry keys that are automatically kept in sync across the failover cluster nodes

A SQL Server failover cluster appears on the network as a single SQL Server instance on a single computer. Internally, only one of the nodes owns the cluster resource group at a time, serving all the client requests for that failover cluster instance. In case of a failure (hardware failures, operating system failures, application or service failures), or a planned upgrade, the group ownership is moved to another node in the failover cluster. This process is called failover. By leveraging the Windows Server failover cluster functionality, SQL Server failover cluster provides high availability through redundancy at the instance level.

SQL Server 2008 failover cluster setup uses a new architecture, enabling functionality that was not possible with the previous versions. The following are the key differences from the previous versions of SQL Server with respect to failover cluster functionality:

• There is no remote execution on the cluster nodes.

  • To install, upgrade or maintain a SQL Server failover cluster on multiple nodes, you must run SQL Server Setup separately on each node of the failover cluster.

  • To add a node to an existing SQL Server failover cluster, you must run SQL Server Setup on the node that is to be added and choose the Add Node functionality. Do not run Setup on the node that owns the cluster to add another node.

  Note

  Add Node increases high availability for the existing failover cluster and does not cause any downtime.

  • There is no uninstall SQL Server failover cluster option. To uninstall a SQL Server failover cluster, run Setup with the Remove Node functionality on each of the nodes of the failover cluster to be removed.

  • To remove a node from a SQL Server failover cluster, you must run SQL Server Setup on the node that is to be removed and chose the Remove Node functionality. Do not run Setup on the owning node to remove another node.

  • If Remove Node is run on the node that owns the cluster, and there are other nodes in the failover cluster, it causes a failover to another cluster node before the node is removed.

  • If Remove Node is run on the node that owns the cluster and there is no other node in the failover cluster, it removes the entire SQL Server failover cluster instance.

• There are two installation options:

  • Integrated failover cluster install with Add Node

  • Advanced/Enterprise installation

• Rolling upgrade and patch support with minimal downtime.

• Increased setup reliability.

• Alignment with Windows Server 2008 failover cluster features:

  • Heterogeneous hardware and ISCSI support without a need for certified hardware.

  • The Windows Server 2008 cluster validation tool is used to validate the Windows cluster and block SQL Server Setup when problems are detected.

  • Service SIDs remove the requirement for domain groups on Windows Server 2008.

  • IPV6 and DHCP support.

  • Up to 16-node failover clusters.

  Note

  SQL Server limits you to two nodes for SQL Server Standard.

• Not aligned with OR dependencies.

• Add features to a SQL Server 2008 failover cluster. For example, you cannot add the Database Engine to an existing Analysis Services-only failover cluster or add a Analysis Services to an existing Database Engine-only failover cluster.