Wondering if there is a good detailed explanation of the settings for CPU and memory when set directly on a VM and when set on the resource pool a VM belongs to. What does each really mean and why would certain settings/changes do different things.

Example: We have a R900 (4) socket quad code w/128 GB hosting about 15 SAP Dev and QA VM's. I have created 3 separate Resource Pools depending on utilization, etc... Also when a SAP install is occurring, I will move that machine into the 'High' Pool.

Several times now we have seen the CPU within the VM Windows OS spiked at 100%, even though the VM has not ballooned and the performance for that VM on the CPU is maybe 60% when seen from the VI. Similar experience with the memory settings.

The only fix I found (and I tried uping the CPU reservation to 20480 mhz for this one machine) was to shutdown the machine and change the CPU setting under 'Edit Settings' to 4 (it was set at 2). This made all the difference.

Why would that make such a difference when the other changes and Resource Pool membership basically did nothing? I have also seen the same thing under the memory settings. Why would these changes be needed when the Resource Pool says the greater amount that I then change the VM to manually?

Anyone else see these type of issues? Am I just not configuring from the beginning? Am I expecting too much of the dynamic changes that can be made in ESX?

Thanks!

1  out of 1 found this valuable. Do you find this valuable? Do you?