The other night I was onsite at a client site to build a second Named Instance of SQL Server 2008 on a Windows Server 2008 Stretch Cluster. You may also know it as a Geo-Cluster or Multi-Site Cluster. The Storage had been presented to the applicable nodes of the cluster and the disks just needed to be added to the Windows Failover Cluster. Your thinking yeah this is a simple task of adding disk from the Cluster Manager GUI. This my friends is where it went wrong and we received this disk error:
“No disks suitable for cluster disks were found”
A little bemused with this as we checked the nodes and in Disk Management we could see the shared storage had been presented and could be seen online and offline respectively. This should work right? Obviously not. As I had not come across this issue before a little bit of investigation was needed. I came across the following article by John Toner where he outlines exactly the issue I was facing:
Following the steps outline in John’s article I was able to add the disk to the cluster using the cluster.exe command line. Thanks John. There are some steps I would like to inform you of that I had to do to get the disk to fail back and forth successfully.
On Node1 you have Added the disk with a drive letter of H:. This drive is visible on Node1 in windows explorer. You have added the disk to a “testfailover” Resource group in the cluster.
Failover the “testfailover” resource to Node2
You will receive a disk failure in Cluster Manager. This failure is indicating that the disk resource has failed to come online.
On Node2 go into Disk Manager, find the appropriate disk (it will be online) and you will have to manually assign it the drive letter H: so that it matches between nodes in the cluster.
Bring the disk online
Fail the disk back and forth between nodes now to test your newly attached cluster disk
Being presented with the opportunity I decided to have a bit of a play with the adding of the disk to see if there are any issues with adding the disk into different Storage Groups.
As per John’s article it mentions to add the new disk to the Available Storage group. This group is by default part of the cluster and holds the available storage that you can choose from to allocate to the appropriate Cluster Resources. Adding the disk to this Group and then creating a TestFailOver resource I was able to fail the disk over successfully between the nodes in the cluster. I was also able to move the disk back to the Available Storage group.
However if you Create the disk into the TestFailOver group you can still fail the resource around the nodes however you are not able to assign it to the Available Storage group. So remember that you will have to put it to the Available Storage group or you will have further issues.
I hope you have found this helpful. Enjoy….