So you wanna make em highly available? well, you gotta be careful! Microsoft Fail-Over Clustering (Formerly Microsoft Clustering Services – MSCS) has been greatly optimized for Hyper-V, with the introduction of Cluster Shared Volumes (CSV) you can really see the beauty of a highly available virtual server environment.
In this post I will be generally talking about Live Migration and a few things to keep them in your mind, then show you a problem that a client has faced because of a tiny mistake.
There are many reasons why you would want your virtual environment highly available, primarily to meet SLA agreement, and most definitely prevent down time. The sensitive nature of Fail-Over clusters make them one of the few things that you REALLY should plan out carefully.
1. Check your hardware against the Windows Server Catalog and make sure you only get Certified hardware: www.windowsservercatalog.com
Certified = Passed a lot of tests.. trust me, A LOT.
2. IDENTICAL HARDWARE + IDENTICAL SOFTWARE yes, I know how annoying it is to read capitalized text, but I think I grabbed your attention, by identical I mean same model and same guts, even the specific details (i.e. memory bus speed, CPU clock speed, NIC model and capabilities) I’m not exaggerating! this is how it should be.
As for the software, same OS Edition (you know this by far right?), same Service Pack level, same set of updates, same Server Features or Roles; but does it work without this kind of convergence between the servers? YES IT DOES, should I do it? NO YOU SHOULD’NT.
3. Validate your cluster configuration When you install the Fail-Over Clustering feature, this is the greatest thing MS has added to its HA Clustering, if it passes validation, you’re more likely to have a super healthy cluster (sort of).
4. Read This
This shouldn’t be your ultimate planning guide, but the above points are to keep in your mind, the whole time.