After you have installed a Microsoft® SQL Server™ 2000 failover cluster, you can change or repair your existing setup. For example, you can add additional nodes to a virtual server in a failover cluster, run a clustered instance as a stand-alone instance, remove a node from a clustered instance, or recover from failover cluster failure.
During SQL Server Setup, you are given the option of maintaining an existing virtual server. If you choose this option, you can add other nodes to your failover cluster configuration at a later time. You can add up to three additional nodes to an existing virtual server configured to run on one node.
To add a node to an existing virtual server
You can remove a node from a virtual server (for example, if a node is damaged). Each node in a virtual SQL Server is considered a peer, and you can remove any node.
A removed node can be added back to a failover cluster at any time. For example, a removed node can be rebuilt after a failure and added back to the failover cluster. Alternately, if a node is temporarily unavailable and later comes back online and an instance of SQL Server 2000 from the affected virtual server is still in place, the Setup program removes this instance from the computer before installing the binaries on the node again.
Note A damaged node does not have to be available to be removed, but the removal process will not uninstall any of the binaries from the unavailable node.
To remove a node from an existing failover cluster
Usually, you run a clustered instance of SQL Server under the control of Microsoft Cluster Service (MSCS). However, it may be necessary to run a clustered instance of SQL Server as a stand-alone instance (for example, when you want to perform administrative operations like running an instance of SQL Server in single-user mode). To connect to a clustered instance of SQL Server 2000 in stand-alone mode using sockets, both the IP address and network name resources must be online for the virtual server on which the instance was installed.
If these resources cannot be online, connect using Named Pipes. However, you must create an alias on the client side to talk to the pipe name on which the instance of SQL Server is listening. Use SQL Server Network Utility to find out the pipe name. For more information, see Failover Cluster Troubleshooting.
Usually, there are two scenarios that cause failover cluster failure:
To recover from this failure, first remove the failover cluster using the Setup program.
To remove a failover clustered instance
To recover from failover cluster failure in Scenario 1
To recover from this failure, first remove the failover cluster using the Setup program.
To remove a failover clustered instance
To recover from failover cluster failure in Scenario 2
You should not change the passwords for any of the SQL Server service accounts when a failover cluster node is down or offline. If you have to do this, you will need to reset the password again using Enterprise Manager when all nodes are back online.
If the service account for SQL Server is not an administrator in a cluster, the administrative shares cannot be deleted on any nodes of the cluster. The administrative shares must be available in a cluster for SQL Server to function.