Many people have been asking about the installation process for a clustered SQL Server instance in SQL Server 2008/2008R2. The process has changed from the installation method used in SQL Server 2005. In SQL Server 2005, the installation was executed from the active node and would install to all selected nodes in one setup operation. However, the new installation method requires the installer to be executed separately on each node.

Let’s assume we have 3 computer nodes named NodeA, NodeB, and NodeC with the following IP addresses:

• NodeA – IP Address 10.10.10.10
• NodeB – IP Address 10.10.10.11
• NodeC – IP Address 10.10.10.12

All three nodes are members of the Windows cluster SQLFAILSET01, which has a virtual IP address of 10.10.10.20. We will be installing 2 new clustered instances of SQL Server using unique virtual network names and SQL Server instance names:

• SQLCLUST01 – Virtual IP Address 10.10.10.21
• SQLCLUST02 – Virtual IP Address 10.10.10.22

For the installation, we need to consider virtual network names, virtual IP addresses, shared drives, and NICs. Virtual network names must be unique and follow the usual naming restrictions. Virtual IP addresses are assigned from the client/public network and are used for communication with the SQL Server instances. Shared drives use shared storage and are typically attached over Fibre Channel or iSCSI. NIC stands for Network Interface Card.

Let’s go through the installation process step by step:

1. Launch the SQL Server 2008 installer on NodeA and select the option “New SQL Server failover cluster installation”.
2. Complete the basic checks and provide the license key.
3. Configure the instance by providing the instance name and virtual network name.
4. Assign the cluster resource group and shared disk resources for the new instance.
5. Configure the network using the chosen virtual IP address.
6. Continue through the installer, providing cluster security policy, service configuration, and database engine configuration.
7. On the data directories tab, specify the chosen drive/directory structure using the previously supplied drive letters.
8. Complete the installation process.

Now, failover the cluster disk resources to NodeB and repeat the installation process on NodeB, selecting the option “New SQL Server failover cluster installation”. This time, use different network names, instance names, and logical disk drives.

At this point, we have two new failover instances deployed to nodes A and B. However, we still need to add NodeC as a failover partner for each instance. To do this, launch the SQL Server 2008 installer multiple times on NodeC:

1. Launch the installer on NodeC and select “Add node to a SQL Server failover cluster”.
2. Complete the checks and provide the license key.
3. Select the SQL Server instance you wish to join (SQLCLUST01\INST1).
4. Complete the installation process.
5. Launch the installer on NodeC again, selecting the option “Add node to a SQL Server failover cluster”.
6. Complete the checks and provide the license key.
7. Select the clustered instance name you wish to join (SQLCLUST02\INST2).
8. Complete the installation process.

Now, NodeC has been added as a failover partner for both SQL Server instances.

The installation process for a clustered SQL Server instance has changed significantly, but these changes have made the installation more manageable. It is recommended to make use of the new slipstream features to integrate the Service Pack and/or Cumulative Update, which can decrease the deployment time for your new SQL Server instances.

As always, it is important to test and sanitize this process in an offline environment before implementing it in a production environment. You can easily accomplish this by using VMWare Server 2.0.x or VMWare ESX Server.

If you need a complete tutorial for the VMWare Server\SQL Server 2008 installation, please feel free to request it.