I have a 2008R2 file server cluster that is extremely critical. It is actually running 2008R2 Storage server with SIS and Microsoft recommends re-inflating the volumes before moving to a new cluster. The current cluster has one node that is so corrupted that we have disabled the cluster service on it, and the shared storage also has enough latency that it causes the other node to occasionally have problems when we try the sisadmin commands or sometimes just fails a datastore as offline. We are thinking that because of these cluster issues with one bad node+bad performing storage that we are probably better to simplify everything and destroy the cluster and go back down to a single node and hopefully it will handle the storage concerns better. Then we can focus on getting the data re-inflated and migrated to a new 2012R2 cluster. That sounds a lot simpler than hoping the cluster becomes more stable when messing with cluster timeouts, etc.
I have never done this, but I am thinking that I just destroy the cluster, delete the cluster computer account, rename the current working node to the cluster name and use the cluster vip, and recreate the shares that it should work fine. Most of our ownership on the datastores are set to the default 500 account and I am thinking that data should come across with all the same ACLS, etc. Does anyone see any concerns with this plan or have any other advice?
Thanks,
Dave