Unreadable disk in MSCS

[wkwicker]

Hello,

I am attempting to create a two node cluster on 2 VM's on different hosts. I have read the guide but have and issue with the quorum (and other) shared disk resource. The first node to boot up will see the disk fine. The second node will see the disk as Unknown and unreadable. I can shutdown the nodes, power on node 2 first and it will see the disk but node 1 will not. It is always the first server to power on that can see the disk.

I have an EMC CX-500 as the storage device.

Both nodes have the shared disk as an RDM in physical mode on a 2nd LSI scsi card.

The 2nd LSI is configured for physical SCSI Bus Sharing

Hosts are ESX 3.0.1

Guests are MS Windows Server 2003 Enterpirse Edition SP2

Even though RDMs are used and the SCSI card is set the physical SCSI bus sharing it does not appear that the sharing is actually working. Any thoughts on this?

[formulator · Accepted solution]

That's the way MSCS works, the active node locks the LUNs so that there is no chance of filesystem corruption. To test you'll want to do "move group" on the cluster admin to see if all resources failover properly.

[wkwicker]

Thanks. I guess you can see this is my first MS cluster also. I figured there would be some type of locking mechanism but didn't realize it would mark it unreadable. I didn't see that mentioned in any documents I read.

Moving the group activates the disk on the other node so it looks like everything is working fine so far.

[Yevgen_Morgensh]

I have same problem. But I can't add second node to cluster because shared disk is locked by first node.

How i can fix this?

[Yevgen_Morgensh]

I have found workaround to resolve this problem.

I use MNS Resource (Majority Node Set) as a Quorum and everything work properly.