Hi all,

I have a strange problem with a DAG cluster in a test environment.

The test environment has a two-member DAG which had been working fine for some time.  Some time ago (I dont access the test environment often so I dont know how long ago) the DAG has failed.  The databases are dismounted and will not mount, with an error indicating that Exchange is unable to read from the cluster database.  Even trying to access the properties of the DAG generates the same error.

When I go into the Failover Cluster Manager I can see that the failover cluster is gone literally! The Failover Cluster Manager shows no clusters at all.  I know for a fact it used to be there (and surely the DAG could not have been created in the first place without an associated Failover Cluster). The DAG computer object still exists in Active Directory.

Any idea how the failover cluster can just disappear?

And Im wondering, if I manually recreate the failover cluster will it fix the DAG, or is it not that simple?

Check whether the Cluster Service is still there and its status is Running.

Thanks Li.  The Cluster Service is there and is running on both member servers, so that doesn't seem to be the cause of the problem.

When I go into the Failover Cluster Manager I can see that the failover cluster is gone literally! The Failover Cluster Manager shows no clusters at all. 

Did you try this?

Right click Failover Cluster Manager -> Connect to Cluster...

Hi Gloneill,

How about the result if you run the following command?

cluster node 

You can try to run the following command to force cleanup the server from the cluster through an elevated command prompt.

Cluster Node ServerName /forcecleanup 

And then, once the cleanup is successful, open Failover Cluster Manager and add the failed server manually back into the cluster.

Best regards,

I was able to connect to the cluster on one of the nodes with 'Right click Failover Cluster Manager -> Connect to Cluster...' but the cluster showed both nodes as down.  Restarting the Cluster Service would not bring them back up.  On the other node I could not connect to the cluster (I can't recall the precice message but I recall it was suggesting the cluser name might have been misspelled - thus perhaps inferring a DNS issue, but the server could resolve the cluster hostname when attempting to ping).

I then tried removing the nodes with the Cluster Node [ServerName]/forcecleanup command, but now I can't re-add them as the Cluster Service will not start on either server (the event log states

An error occurred while trying to write to the cluster database. Error: OpenCluster failed with error 1753.)

Any other thoughts welcomed (although I'm about to go on vacation for two weeks so it will be at least two weeks before I look at and actions any suggestions posted).

Right click Failover Cluster Manager -> Connect to Cluster...