How do I deal with a SQL Server cluster that is not acting properly?

Question

We are having a problem with our SQL Server cluster. Everything was originally set up and configured properly, then for some reason the SQL Server Fulltext service failed on the primary node, and will not run on this node any longer. I changed the configuration of Fulltext so that this failure would not prevent the group from failing over, and this works. 

However, I'm trying to get to the bottom of why Fulltext failed, and in the event log for the node I see the following message:

'An error occurred during the online operation for instance : 80040154 - Class not registered'

So at this point I am thinking its time to reinstall the SQL cluster. The problem is that the SQL Setup process no longer recognizes the Virtual Server! How should I go about uninstalling at this point?

Answer

While I have not seen your specific problem, I have seen similar problems, especially the problem of SQL Server setup not being able to see the virtual server. In these cases, even though you will not like what I have to say, the best option is to rebuild your cluster from scratch.

The purpose of a cluster is to ensure high reliability, and you can't guarantee this if a cluster begins acting up like you describe. Whenever I see the slightest hint of a problem, I start over from scratch, reinstalling the operating system SQL Server, and service packs as appropriate.

Only when I get a clean install, with no odd behavior, do I believe the cluster is ready for production.