I track the metrics on my blog to see which posts are popular, in an effort to gauge my writing and what people are interested in. One post in particular, that I wish wasn’t popular but is, is “Recovering From a Deleted Cluster Resource.”
This post stemmed from an incident at my previous employer, where an offshore DBA had deleted the cluster resource from failover cluster manager “because he used to work at Microsoft, and they did this all the time” (bad idea jeans). In most cases that I’ve seen, people will try to do this when there is some security problem with a cluster. This won’t fix your security problem, however this KB article might.
What will happen however, is that your clustered instance will go down. HARD. Your clustered instance is dependent on that resource–so when the cluster loses contact with the resource, it will stop the SQL Service, and won’t be able to start back up, because of the missing dependency. You actually run the risk of having data corruption if this happens. Fun, fun.
Note—this isn’t the worst thing you can do to a cluster. That would be deleted the active directory cluster computer object—if your AD recycle bin isn’t turned on, there is no recovery except to rebuild the cluster. Also, formatting cluster disks is also a really terrible idea.