I have a 2 node (Dell R720's) cluster that was just rebuilt from 2012 to 2012 R2. It's linked to a Dell MD3200 array for the CSV.
I knew multi-pathing had to be installing to see the array properly as one drive but wasn't certain how it should be done. I installed the MPIO feature in Server Manager but I knew that there was software missing that had been there on the nodes when they ran 2012.
I figured out that it was the Dell MDSM so I installed that. Now when running the cluster validation I get errors about a DSM mismatch between nodes. It shows the Microsoft and Dell DSM's installed on both nodes which I now figure is the MPIO and the MDSM. the versions of each manufacturer's DSM do agree between the 2 nodes but of course they don't match each other.
I'm assuming I am not supposed to run both DSM's and need to uninstall one of them. Since the array is a Dell and I'm sure I remember the software being there before I wonder if I should uninstall the MPIO feature via Server Manager. If I do, will this mess up anything? I'd rather try and live with the version mismatch errors than mess up the array or cluster in any way.
I finally got this taken care of. What didn't match was the QFE number (which seems to be part of the DSM number). I found that this is due to not having the same Windows updates installed on both nodes. Once I got the same updates installed on both, the QFE number matched between server and the validation test gave me the green light to join the second node to the cluster. I did that, moved a few non-critical VMs to it for testing and it's running great.
That is correct, the issue is due to repetitive software. When you get the DSM mismatch error it is due to having loaded the MPIO service and then MDSM, which also includes MPIO. I am not certain of the stability of removing one or the other. You would need to completely uninstall the MPIO you loaded AND the MDSM as well. Once they are removed then you can install the MDSM software and it will resolve the issues you are having.
The MD3200 resource guide disk should have the current full version of MDSM on it, you can find it here - http://dell.to/11Dk7sr
Let me know how it goes.
DELL-Chris H
Social Media and Communities Professional
Dell Technologies | Enterprise Support Services
#IWork4Dell
Did I answer your query? Please click on ‘Mark as Accepted Answer’. ‘Thumbs up’ the posts you like!
I don't think there would be an issue migrating them out of the active node we are working with. If there is a concern you may want to wait until you have a maintenance window to perform the steps.
DELL-Chris H
Social Media and Communities Professional
Dell Technologies | Enterprise Support Services
#IWork4Dell
Did I answer your query? Please click on ‘Mark as Accepted Answer’. ‘Thumbs up’ the posts you like!
thanks for your advice, that clears things up on my part about DSM/MPIO. What I'm wondering now is this. With the VMs all failed over to node 2, can I uninstall the DSMs on node 1, reboot, and re-install the Dell DSM without affecting the VMs at all? (and then failover the VMs to node 1 and do the uninstall/reinstall on node 2)
I would think that it would not affect them at all since node 1 was rebuilt with R2 while all the VMs sat on node 2, but I just don't want to adversely affect the operation of the VMs at all.
In my case, the QFE numbers differed because I was doing a cluster upgrade from 2012R2 to 2016 hosts. This is normal. But I got similar problems. They were resolved by upgrading the driver on both hosts for the HBA 12GB card that we use to connect the MD1420
SmallBizAdmin
25 Posts
4646
0
Posted December 18th, 2014 15:00
I finally got this taken care of. What didn't match was the QFE number (which seems to be part of the DSM number). I found that this is due to not having the same Windows updates installed on both nodes. Once I got the same updates installed on both, the QFE number matched between server and the validation test gave me the green light to join the second node to the cluster. I did that, moved a few non-critical VMs to it for testing and it's running great.
Jonathan