Mark Minasi's Tech Forum
Register Calendar Latest Topics Chat
 
 
 


Reply
  Author   Comment  
Pieter

Avatar / Picture

Senior Member
Registered:
Posts: 131
Reply with quote  #1 
I know I can over commit CPU's on Hyper-V.
Does that imply that I have to look at the host for measuring the load ?
Are there any rules of thumb ?

Example:
Host has 16 cores, I can have 16 VM's with some of them (or even each) more than 1 vCPU. To measure the load I have to look at the physical CPU's on the host.


Sorry for asking something I surrely can Google (or Bing or...), but I'm just hoping to get a quick answer, I' m a bit running out of time...

__________________
Pieter Demeulemeester
0
cj_berlin

Avatar / Picture

Senior Member
Registered:
Posts: 177
Reply with quote  #2 
Hi,

overcommitting CPU is known to cause all kinds of strange behaviour, the most frequent one being (Host load = low, VM load = low, VM speed = horrible).

Like with VMware, your friends here are the somewhat more obscure PerfMon counters, my favourite being Hyper-V Hypervisor Virtual Processor\CPU Wait Time Per Dispatch which, per official definition, will give you “The average time (in nanoseconds) spent waiting for a virtual processor to be dispatched onto a logical processor.” Lower is better ;-) Everything above 2,5 ms (2,500 ns) is a cause for concer, above 5 ms is bad.

This example is certainly open for improvement (i.e. put in more cores ;-) ):
cpu.PNG 


__________________
Evgenij Smirnov

My personal blog (German): http://www.it-pro-berlin.de/
My stuff on PSGallery: https://www.powershellgallery.com/profiles/it-pro-berlin.de/
0
wobble_wobble

Avatar / Picture

Associate Troublemaker Apprentice
Registered:
Posts: 741
Reply with quote  #3 
Veeam One - small VM and its free after 30 days, but limited historical storage.

Great for VMware and Hyper-V

__________________
Have you tried turning it off and walking away? The next person can fix it!

New to the forum? Read this
0
Previous Topic | Next Topic
Print
Reply

Quick Navigation: