Last time we looked at getting yourself ready for Hyper-V and SCVMM 2008.  A little best practice will go a very long way.

But if you don’t, well you can still survive.  But here’s where…

“InConsistency will put you on the throne”(and not the good one)

Now you can still use SCVMM 2008 without an expensive SAN and individual servers through network wiring.  Hyper-V doesn’t care, nor does the system.  It will work with it.   But when you get into some mismatching (different versions of Hyper-V), different speeds of drives on hosts, mixed brands of network cards.  (And Please I don’t recommend ANY of this!  That’s a problem waiting to happen).   But being aware that as you veer away from “Datacentre Nirvana” you will have problems, I though it would be good to pass on how to DEAL with them.

Because problems will happen.  As always.  Plan for it and you’ll be fine.  Don’t and well… Data recovery companies will always take your money and your job.

Here’s a few I’ve encountered.  They’re not very common, but knowing how to deal with it can save your bacon.   And this does NOT involve rebooting the physical server.  I like to avoid that if I can WHENVER possible.

“I look at my server and can’t view any machines!  It’s not manageable but I’m the server!”

This one is the easiest of all.  There are three services in use by “Hyper-V”.  One is the Hyper-V manager service.  Probably needs to be restarted.   What trips this off?  Well honestly this is where you are probably pushing the available ram to it’s limit.  At least that’s where I’ve encountered it.   On a machine with 32gigs of ram, I was pushing 85% of the available ram, watching the CPU’s max out.   In that scenario, I saw the Hyper-V Management service shut down a lot.   Take that as a hint to look into another server or more ram / CPU power on the existing server.

“I have machines running.   There’s no errors in event viewer.  I shut one down and it won’t start up!”

Most likely (failing a corrupt VHD or config file which isn’t common) it’s going to be the certificate for Hyper-V.  It’s good for a year, and if your has expired the process to fix it is very easily.  Restart your three Hyper-V services and it will generate a new one.  No reboot of the server is necessary.   Other than a temporary interruption in the network, all the Child Partitions should still run.   Remember.  You’re not running that Child operating system on the parent.  You’re running it on the Hypervisor which is an O/S BELOW Server 2008.   Your Child partitions shouldn’t care about the Services on the parent, only what the parent provides (some resources like the Virtual Networks)

“I migrated a machine and started it up before properly assigning it to a Virtual Network.  Now it’s GONE!”

This can happen, and again very easy to fix.   On the machine in question, you’ll most likely find the files are where they should be and everything is ok.  Don’t bother trying to re-create it.  Why?  It’s not gone.   Just go to www.sysinternals.com and download Process Explorer.  Do a search for the VHD files from “Find” double click on the list and it will carry you over to the process using it.  Under the process accessing your VHD (or VHD’s), Right click on the particular VHD and choose “Close Handle”.  You’ll get a warning about this being a “bad idea” but that’s ok.  Do it.   Do it for all the VHD’s from that PARTICULAR Child partition.

What’s interesting is it doesn’t seem very effective.   What’s more interesting is you will go back to Hyper-V Manager and probably (after a refresh) find your missing Child partition back with it’s configuration completely intact.

“It won’t let me migrate and there’s nothing wrong with either server!”

Not a problem.  Read the error message.  We’re back to “Best Practice”.   You may find it’s simply because you’re running two different versions of Hyper-V or different patch levels to be correct.   The solution is to do a full shutdown of the Child Partition and then migrate.  Then (after reassigning the Virtual Network on the destination Hyper-V box) starting it back up.

“SCVMM 2008 is showing that The destination server has my computer, it’s existing on the Source, there’s two entries and I can’t fix either one!”

Simple but a little scary to fix.   What is actually true is the Destination computer has it but the computer controlling it (including the location of the source files) is on the source computer.  To fix this scenario you need to shut down the Child Partition O/S.   Then go to whatever folders have the VHD files for the the Child, and backup the files, the configuration to an alternate location (maybe even a different drive to be safe).  Document all the configuration settings (Virtual Lan, IP addresses, attached hardware) from Hyper-V Manager for that partition.  

Then in SCVMM 2008, under the Virtual Machines, delete them.   Yes you heard me right.  Delete them.

Once they are successfully removed, recreate the settings on the original server by copying the files back and putting in your settings.   A little rough but you’ll be good to go.

Of course the key thing here is always follow best practice, have Backup in place.   But in the words of Douglas Adams, “Don’t Panic”

Sean
The Energized Tech