I have read a fair amount of information about resource pools but I am struggling with this example. Can anyone give me a quick walkthrough on how to accomplish this.
At a remote site we have a single ESX cluster of three host that have 40 logical procs and about 30 VMs
3 of the 30 Vms have been giving 40 procs each. (these are indexing type machines that use all 40 procs 100% of the time)
What we would like to do is:
To give each of these 3 VMs all available proce on the host (40 procs) and allow those vms to completely saturate all available host CPU all the time when there is no contention or other VMs are not in use. However, the second any other VM needs even a sliver of CPU the 3 VM hogs need to take a back seat and give it up.
Basically I want these 3 VMs to run wild until someone else needs CPU. Does that mean every VM in the cluster needs to go into one of two pools? 1 pool that consist of the 3 VM hogs and a 2nd pool of all the others? Then we define CPU shares on each pool?
In short. 3 ESX host with 30 vms. We want one vm on each host that can take 100% of the CPU on the entire host. However the second another vm wants cpu it lets it go.
Im a reading this correclty?