IGT Part 4: Failover frustration

This article is part of a series on Incoming Google Traffic (IGT).

Want to know what happens if your virtualization platform uses native OS clustering technologies to provide highly-available virtual machines?  You become an unwitting cluster administrator.  I’m referring to MSCS-style clustering, now known as Failover Clustering.  You know, one VM per LUN stuff.

Take a look at this Google query:

how+to+shutdown+vm+when+ha+hyper-v

If you have ever administered a clustered service, like MS SQL Server, you learned quickly not to manage it with the native service control tools — either by reading the documentation or the hard way.  The same goes for virtual machines that are cluster resources:  Shut them down the wrong way and they will be dutifully restarted by the cluster service.  This is because the cluster sees such a shutdown as a failure.

For folks visiting VCritical in search of such a solution, here is your tip: do not shut down your guest OS via the Start menu.  Oh, and don’t use Hyper-V Manager, either.  Try the other single panes of glass.

Related posts:

1. What would things be like without VMFS?
2. Choose any two: Hyper-V, HA, Linux
3. Microsoft Private Cloud: No CSV Allowed!
4. IGT Part 1: Mojave Only Need Apply
5. IGT Part 3: One-VM-per-LUN doubters