Windows Server File Cluster help needed!

I hope you’ll all forgive my unusual approach here but I’ve been having major problems with a file cluster we built a few weeks ago. Naturally I’ve been trawling around the net for a solution but so far I’ve not found anyone else having the same problem.

The problem I am having is with a network name (or Alias) that I’ve assigned to my file cluster so if you are a cluster expert perhaps you could read on and offer some advice?

To everyone else, I’m sorry for the interruption to our normal technology coverage!

Ok, lets me start with a quick summary of the problem before I go on to explain what I’ve already looked at.

Several weeks ago we built a new cluster comprising of two nodes. After testing the cluster for a while we were happy that it was working properly, it would fail over properly from one node to the other and the resources and network name that we gave the cluster were visible and working.

Three weeks ago one of the servers that we were due to decommission anyway began to fail and developed a problem with AD. Our only option was to copy the data from the faulty server to the cluster, remove the server from the network and seize all of its roles on a new server. This in itself worked just fine so we then added the old faulty server name as a network name on the cluster so that existing shortcuts and shares would still work.

After testing this out of hours on the move day everything seemed to work just fine and for two weeks the people in the organisation (about 100 clients) were blissfully unaware that anything had changed!

However, last week a few clients began to experience problems. This comes in the form of a connect to server window asking for a username and password. Even if you enter the correct credentials here you are unable to connect to the resource, windows will simple report that the resource is unavailable.

The problem doesn’t affect all clients (just a handful at the moment) and the problem is intermittent. If you leave the client for a while you can generally reconnect to the resource later without seeing the logon screen.

The other strange thing is that even when the client begins to prompt for password via the network name ‘NTSERVER’  the other cluster network name ‘CLUSTER’ still works and does NOT prompt for password. Both network names are on the same cluster and same cluster group and have been set up in the same way.

We are pretty sure that this isn’t a DNS related issue as pinging the network name resolves the correct IP address and gets a response, NSLOOKUP also returns the correct results.

If you do not make any changes to a client it will go between working and not working throughout the day. Changing the DNS servers and IP address from DHCP to static has no effect. When the problem occurs nothing seems to resolve it, likewise we’re not able to cause the problem.

Nothing is being logged in the event log on either server.

I’ve tried removing the network names and recreating them, other network names are fine but the one ‘NTSERVER’ has a problem, I’m assuming because it’s the same network name as a previous physical server but cant find where this would cause a problem.

If you can offer any ideas please, please do email me, the problem is driving me nuts! [email protected]

Thanks!!