The Failover Clustering option is a more expensive solution for providing SQL Server availability. It provides very good protection in the case of hardware failure. Failover to a passive node is fairly quick (anywhere between 1-5 minutes depending upon the state of the cluster and database). Failover Clustering provides service availability but does not provide data redundancy like database mirroring and log shipping. Data protection has to be provided at the storage level or by combining with other solutions.
Failover Clustering provides instance-level protection built on Windows Failover Clustering. Nodes of the cluster are typically co-located within the same site or datacenter to provide local availability but can also be deployed regionally. Disaster Recovery (DR) and Business Continuity (BC) are possible using geographically dispersed clustering solutions. These solutions can be difficult to design, implement, and maintain from a cost and operational standpoint.
Implementing SQL Server failover clustering involves a number of steps:
1. Create a virtual machine and install all tested and critical updates and service packs. You should template this virtual machine using sysprep and deploy your virtual machines from templates. This will significantly reduce the time needed to provision your virtual machines.
2. Create a new VMware vSphere virtual switch or port group and add a second network adapter to the cluster’s primary node virtual machine to carry cluster heartbeat communications. If using vSphere, select the vmxnet-3 adapters for your public and heartbeat adapters.
3. After completing the network configuration, you map your storage array LUNs as Raw Device Mappings (RDM) to the primary virtual machine. Keep in mind you need to set up a separate vSCSI controller for clustered disks as clustered disks cannot reside on the same vSCSI controller as the OS boot drive. You also should map your database and log LUNs on separate vSCSI adapters.
Following setup of the SQL Server 2008 primary node virtual machine, you can configure the virtual machine of the secondary SQL Server node.
5. When you are finished with the secondary node, you need to install Microsoft .NET application server roles, configure associated role services, and add the failover cluster feature to both primary and secondary virtual machines in the cluster.
a. On the primary node, open the Disk Manager and bring all newly presented disks online and, using DiskPart, align your disks to prevent split I/Os.
b. On the secondary virtual machine, open the Disk Manager utility to bring disks online to the secondary node virtual machine. Map the disk letters exactly as they are mapped on the primary virtual machine.
c. Use the Microsoft Cluster validation wizard to validate a cluster configuration. When validating a cluster for the first time, you MUST run ALL tests and all tests must pass to be a Microsoft supported failover cluster solution.
6. After completion of the validation, use Microsoft Failover Cluster Management options to create a Windows Server failover cluster to support a clustered SQL Server 2008 instance(s). You will then use the Failover Cluster Management tools to configure the Cluster Quorum or “witness” drive and settings to specify failover conditions for the cluster.
7. With the cluster now fully configured and functional, configure the Microsoft Distributed Transaction Coordinator (MSDTC). SQL Server uses the MSDTC for distributed queries, replication, and two-phase commit transactions. To configure the MSDTC, you open the Microsoft Failover Cluster Management console to configure a service or application. The console provides a High Availability Wizard to configure MSDTC to perform the SQL Server cluster failover.
8. The last step is to actually install SQL Server 2008 on the failover cluster primary node. The procedure to add an additional cluster node to a clustered SQL Server instance is similar to the failover cluster installation of SQL Server. You can perform this task, from either the primary or second node of the Windows Server failover cluster
9. With SQL Server node installation complete, perform instance failover tests.
Availability Guide for Deploying SQL Server on VMware® vSphere” white paper located at: http://www.vmware.com/solutions/business-critical-apps/sql/resources.html