Does anyone have experience installing SME onto a Microsoft Exchange 3 node active/active/passive cluster? The documentation only describes a 2 node active/passive. I want to be sure about caveats.
Is there anything in particular that you're looking for? I've installed 3 node (a/a/p) in my test lab before without issues. As far as SDW and SME/SMSQL is concerned, it shouldn't matter if it's a 2, 3, 4 node cluster.
If there's something else that you're looking for info on, please let us know and we'll get you the info...
Thanks for the reply. I'm most interested in if there are any caveats I should be aware of. Is it just a normal install with just another active node? In your test lab I assume they were net new installs of the Exchange cluster. I'll be working on a production Exchange cluster. I've run into problems before when moving the quorum resource from one disk to another - like what I'd be doing in this situation with SME. It complains about disk signatures not being what it expected for the quorum resource and breaks the cluster. Do you have any reccommendations for successfully moving the cluster quorum and DTC resources from direct attached storage over to the Netapp?
Moving a quorum disk resource to a new disk is pretty easy.
Use SnapDrive to create a new LUN (500M - 1GB is fine) in the Cluster Group, give it any free drive letter.
Right-click the cluster name at the top of the tree, choose Properties.
Click on the Quorum tab, and select the new disk as the Quorum. Apply.
You may also have to swap out the disk dependancies..
As for the MSDTC, you'll have to destroy the original one and create a new one with the new quorum disk as a dependancy.
It's only used for installation of Exchange, unless you're also running SQL Server in the same cluster.
Unfortunately, it's not quite that simple I've done just that before and had the cluster be broken. What happens is that the disk signature of the quorum gets changed when moved. The cluster gets confused and complains that the disk signature that was expected could not be found. Is there a way to preserve the disk signature of the quorum resource?
I just found this article on Cluster disk signature mismatch:
I think I'll give it a try and see if it works.
You shouldn't have the cluster service complain about disk signatures when you're actually moving the quorum to another location. Sounds like there might be a MS bug that you could be hitting.
Did you try the steps in the link that you referenced? Playing with disk signatures is never fun...and you really shouldn't have to do that. If the quorum move is failing because of it, I'd follow up with MS.
Unfortunately, moving the quorum to another disk does indeed change the disk signature. It's a well known issue. (see Microsoft Q305793). I followed up with Microsoft and they said the problem was with Netapp.
I apologize but I have not found anything out about this yet. I'll ring the proper bells again and see what shakes down. The response time right now might be a tad slower than normal...with the holidays and vacation and such.
I'll report back with anything I find out.
I got a reply back from our dev folks and tdhey don't see anything with this kind of configuration that would cause the disk signatures to change (during the quorum switch) and cause the cluster service to fail. We're not exactly sure why the Microsoft rep would say that this is a NetApp issue. NetApp (SDW) does nothing special to shared LUNs that would cause the disk signature to change after, say, a reboot or something. Once the quorum is moved (a Microsoft procedure), and the cluster comes back online, you shouldn't have to worry about disk signatures changing.
LEt me know if there's something I haven't addressed with this explanation. Or if you have other questions. I'll be online as much as possible over the next day or two...then I'll be on vacation till the New Year.
NetApp Wins One Silver and One Bronze Stevie® Award in 2022 Stevie Awards for Sales and Customer Service