Preparing for SQL Server 2000 Clustering
The purpose of this section is to show you how to install SQL Server 2000 clustering onto a Windows 2000 two-node cluster. It already assumes that the cluster server hardware has been installed and configured correctly, and it assumes that Windows 2000 Clustering Services have also been installed and configured correctly.
This document provides an example of how to install an Active/Passive SQL Server 2000 cluster configuration. An Active/Passive SQL Server 2000 cluster installation means that there will only be one instance of SQL Server 2000 running on the cluster at any one time. Generally, the node of the cluster that SQL Server 2000 is running on is called the primary node. The other node of the cluster, which is not running an instance of SQL Server, is called the secondary node. Should the instance of SQL Server 2000 on the primary node fail, then SQL Server 2000 will failover from the primary node to the secondary node in order to run.
It is also possible to install an Active/Active configuration of SQL Server 2000. This option is designed to allow two instances of SQL Server 2000 to run on the cluster, with a single instance of SQL Server 2000 running on each of the two nodes of the cluster. Should one of the two instances of SQL Server 2000 fail, then the instance that failed will failover to the other node of the cluster. At this point, then both instances of SQL Server 2000 will be running on a single node of the cluster. Although most of this document is applicable to installing an Active/Active configuration, the focus is on how to install an Active/Passive configuration, not an Active/Active configuration.
Preparing for the Installation of SQL Server 2000 Clustering Before you begin, you need to be sure you have the following available and ready to go:
- Appropriate cluster hardware that has been certified by Microsoft.
- Windows 2000 Advanced Server or Datacenter Server, and its latest updates, need to have been properly installed on the cluster.
- Windows 2000 Cluster Services must have been properly installed and configured.
- The Windows 2000 Cluster Service needs to have been thoroughly tested to ensure that it is working correctly.
- Ensure that the domain account that you will be using for the SQL Server services has been added to the local administrators groups of both servers.
- Ensure that you (the installer) is a local administrator of both of the nodes of the cluster.
- You have a copy of the SQL Server 2000 Enterprise Edition installation CD.
- You have a copy of the latest Service Pack for SQL Server 2000.
- A name you can assign to the SQL Sever 2000 cluster. This is the virtual name that clients will use to access SQL Server 2000. This name must consist only of letters or number, no special characters.
- An IP address you can assign to the SQL Server 2000 cluster. This is a virtual address that clients will use to access SQL Server 2000.
Copying the Software to the Local Hard Drive While this is not required, I recommend that the SQL Server 2000 Enterprise Edition CD be copied to a local drive of the primary node of the cluster. This will make this installation, and later changes, easier. I also recommend you copy the latest SQL Server 2000 service pack files to the same local hard drive. These files only need to be copied to the node of the cluster that you will consider your primary node, which is the one that you will use to install SQL Server 2000.
Configuring the Microsoft Distributed Transaction Coordinator Unlike non-clustered installations of SQL Server 2000, the Microsoft Distributed Transaction Coordinator (MSDTC) is required when clustering SQL Server 2000. Before SQL Server 2000 can be installed, the MSDTC (which is now a part of SQL Server 2000), must be configured for running in a cluster. Fortunately, this is a simple step. To configure MSDTC on the cluster, you must run the following command on each of the nodes in the cluster. The command should be run from the command prompt. The command is: comclust.exe After running the command on each node, you will receive this feedback from the command prompt, as seen in the following illustration.
To verify that MSDTC has been properly setup, following these steps:
- Start Cluster Administrator. It doesn’t really matter which node of the cluster you do this on, but I always do it from the primary node so I always know which node of the cluster I am working from.
- Next, click on the “Resources” folder. See the following illustration. On the right, you will see that a new cluster resource has been added to the cluster. It is called “MSDTC”. If you see this resource, you can be fairly certain that it was installed correctly, and is now functioning correctly.