Virtual Environment: Reactivation Required?

NoelDP said:

As you've probably noticed, I'm at the limit of my knowledge when you start talking about VM's and servers in the same breath :)
One or the other I can cope with, but put the two together and I start to think

I have a suspicion that your scenario of reactivations isn't far off the mark - but that so long as the node is in a stable state, once re-activated, it shouldn't get upset until you change the Virtual hardware again, either by switching virtual motherboards or real ones. I don't know how much of the real board a VMWare VM actually sees, and the triggers for re-activation aren't the same as they were in XP, where changing a network card and a graphics card at the same time meant that if you changed the HD as well within 120 days, you could pretty much bank on needing reactivation.
The theory is that only motherboard changes will cause reactivation in Win 7/8 - but how the motherboard is defined in the activation/SPP software isn't published AFAIK, and anyhow I've at least heard of cases where updating the chipset drivers can cause a re-activation request, and I know that broken Inter Rapid Storage drivers can lead to systems failing the WGA testing that goes on, resulting in non-genuine notifications.

I'm not sure either. The proper solution is to use Enterprise and volume licensing that checks in with a local server (like any number of the 2k8r2 servers we have) and then that 2k8r2 server can sync with MSFT. This is the "correct" way to handle it in my eyes and how I'll be setting up the W8 stuff in about 18 months.

You can only run a VM on a single host meaning the W7 instance isn't being driven by cores/threads/memory, etc. across 20 different hosts/nodes/computers, etc. it physically resides on a single motherboard with a single (or dual) CPU, etc. This limits things. You can overbook resources. There are ratios and best practices so this is an extreme example but if you have 36GB of RAM on a Host you can build 10 W7x64 VMs and give them each 6GB of RAM.

VMware has these cool features in it's lineup that allow you to cluster servers (or use bigger multi-node servers). The VMware will monitor each host and auto-detect when a Host is starting to reach capacity or has a hardware failure. In the example above, if all 10 W7 VM's turn on and only are using ~2GB ram each there will be no issue but if they all start to approach ~3GB each (or some more, some less but the host's resources are almost at a threshold) the VMware can automatically move and migrate a VM -- with no interruption or change to the end-user experience -- to another Host that isn't at capacity. The same is true in the event of a hardware failure, etc.

So currently we are overbooking resources but not using vMotion to migrate in the event of a failure or capacity issue - things just get throttled but the VM's stay on the same physical Host. I want to change this because its better in general but also because we aren't using our hardware to it's full capacity and we need to add a few more VMs. In doing that I have no idea where a given VM will be moved to or what Host it will be run on - so there could be re-activation issues all day every day which would totally interrupt the end-user experience (as well as make me look bad when "this is not genuine" notices start popping up).

Hopefully that better explains the situation. I have asked MSFT to work with me on a volume license solution but it may not be worth it because we'd have to re-do or rebuild many of the VMs. I'm looking for an 18-month band-aid rather than an actual solution... I think...