Hi everyone,

I'm looking for clarification on how cloud capacities in VMM work and wondered if anyone here could be so good as to assist.

Point 10 of the "To create a private cloud from host groups" section of this page states that the Virtual CPUs quota type is, "Applied against running virtual machines.".

Likewise, for the Memory quota type it states; "Applied against running virtual machines only."

This suggests to me that I should be able create a cloud with a vCPU capacity of 4 and a memory capacity of 8GB. Then, within that cloud, I could have any number of VMs configured to use 2 CPUs and 4GB of RAM but I would only be able to have two of those running at any one time.

When I try this configuration however, I find I'm unable to deploy a 3rd VM to the cloud I question, even if the first two are shutdown. The error I'm presented with is, "The operation results in a violation of the processor count capacity of cloud <cloud name>."

Am I misunderstanding something? Perhaps the definition of 'running' in the article is that the VM is deployed to the cloud but not necessarily running?

I'm running SCVMM 2012 R2 with update roll-up 6.

Any assistance or clarification would be greatly appreciated!

Best regards,

Barney.

A Cloud is a representation of capacity that a customer or group of customers can consume.  It is simply an abstraction above Host Groups.

"Applied against running VMs only' means that when the VM is powered off it is not consuming memory nor CPU, just storage.  Therefore there is now way to enforce a memory quota when a powered off VM is not using any.

You have all of that correct.

There are also quota points which apply to deployed VMs that may or may not be running.

But I understand your confusion.

I wonder if your first two VMs did not equal your quota, but were just short, if the error would be the same or different.

Hi Barney,

I would like to check if you need further assistance .

Best Regards,

Elton Ji

Thanks Brian, for your response. I'm glad the issue isn't down to a mis-configuration on my part!

I can see that if cloud capacity checks are done at VM provisioning time then modifying it to run on VM start up / shut down would be a not insignificant change. I wonder if the wording on the TechNet page could be updated to avoid confusion in the future?

It's not quite as convenient, but I think my use case will just have to make liberal use of 'storing' VMs to a library. In effect, that'll achieve what I want.

Thanks again for your help.

Best regards,

Barney.