This post is more than 5 years old

13 Posts

8814

September 12th, 2010 08:00

Clustering using MD3000 storage

I'm trying to set up a failover cluster using this configuration:

- MD3000 Storage

- Two 6850 PowerEdge servers running windows 2003 server x64 Enterprise edition, each connected to both raid controllers of the MD3000 using two dualport HBAs per server (using one port port of each HBA) - (redundant paths)

Cabling is not the problem as I had a cluster installed using this configuration.

I Installed the Modular disk storage management software on both servers. Crated a host group containing both servers (Topology: Windows 2003 server clustered), created four virtual disks and mapped them to the host group. Restarted the first server (management station) , shut down the second server, started the disk management utility from the first server and created the partitions over the four physical disks (MBR).

In the first disk (RAID 1) created two partitions one 1 GB size for the quorum (Q:) and another for application purposes (L:).

After shutting down the first server went to second server, opened disk management and was able to see all the partitions created in the first server and had write access.

Shut down the second node and on the first node installed the cluster, assigning the Physical disk containing the Q: and L: partitions for the quorum. Brought up the second node and from the first node tried to add the second one to the cluster but I got the message that there's no quorum capable resources accessible from the second node.

Thinking that there's a problem with the communication between second node and the storage removed the cluster and shut down the first node. From the second node installed the cluster ... it ran fine adding this node... but when trying to add the firs node got again the message that there's no quorum capable resource accessible to both nodes...

Removed the cluster again and verified several times that bot servers independently have access to all the disks configured ....

Does any one have any idea about this issue?

Regards,

Felix

 

4 Operator

 • 

9.3K Posts

September 12th, 2010 12:00

When you're in the cluster wizard, select the advanced option and select to skip the disk detection part (don't remember the exact wording).

 

As it sounds like you've make sure both servers can see the disk, it's acceptable to skip this detection, and then the cluster will build just fine and you can test failover of the disks between the servers to make sure it's working.

13 Posts

September 13th, 2010 09:00

Thanks a lot.... that did the trick.

Top