Hyper-V and VMM Memory stats…complexity abounds

Someone requests you to build a new environment.  Do we have enough resource?  In VMM go to “Fabric” workspace, find the Cluster, and hit Overview at the top.  Hey no problem!  I have tons of space here.  Well that 256GB is only 128GB when you have a 2 node cluster with 1 reserve.

If you mouse over the memory graph it says 41GB allocated to VMs.  This is calculated from the Startup parameter so it’s not the real truth!  This is enough to meet the demand to START the vms.  They will and do consume more RAM as the apps start up.  So what do we do?  You can open up VMM and look at “Memory Assigned” column and get out your calculator…OR you can run this script from chfvwp-scrdm01 or any hyper-v host.  Specify the physical hosts you want to aggregate in the get-vm section.

get-vm -ComputerName hv01,hv02 | Measure-Object -Sum ‘MemoryAssigned’ | select Count, @{n=’UsageGB’;e={[int]($_.Sum/1GB)}}

SOOOO our 14 VMs are actually using 51gb of ram in production at this moment.  We can have a total of about 120GB with 1 node failover so we should be in the clear.

Want to really go deep?  Check out how VMM calculates over-commit in 2012 R2.

Leave a Reply

Your email address will not be published. Required fields are marked *