I ran into an interesting problem with one of our Windows 2003 clusters the other day. I presented a new 2 TB LUN from our EVA8400 to one of the cluster nodes and for some reason the physical disk signature never replicated to the secondary cluster node. The result being, when I failed the cluster over to the secondary node, the disk immediately failed to come online, as did the share folder which was dependent upon the physical disk. My solution was quite simple, although I still don’t know what caused the problem.
Within the registry of both nodes, I drilled down to HKLM\System\CurrentControlSet\Serivces\ClusDisk\Parameters\Signatures and compared values. I noticed that on the server which could bring up the disk, there was a Key value of 0D637FE6, but on the node that couldn’t bring up the disk, there was no entry for this signature. So I created a new Key value with the same signature (0D637FE6) and left it blank. I rebooted the secondary node and once it was back on-line, I failed the cluster over. Sure enough, the disk came online as did the share folder.
As mentioned before, I don’t know why the signature never replicated over. When I looked through the event log, all I found was event id 1069 in which it stated it couldn’t bring the share folder online. It never indicated any problem with not being able to bring the disk online.
Anyhow … simple fix for a rather bizarre problem.