Resource Pool Reservation

[joshp]

I am slightly confused on the implementation details of resource pools. Lets say I create a Resource Pool and have assigned it a CPU reservation of 10000MHz. If I do not allow it to be expandable; does this mean that any virtual machines running under this resource pool are limited to a maximum of 10000MHz combined? If yes, by deselecting expandable reservations, is this similar to setting the Limit at 10000MHz?

Thanks for the help.

[fejf · Accepted solution]

Reserveration = what you reserve for this pool. That means if there is congestion than this resource pool (and so the VMs inside) will always get a minimum of 10000Mhz (they may get more but that's the minimum which is guaranteed)

Limit = limit for this resource pool. That means if there is congestion than this resource pool (and so the VMs inside) will get a maximum of x Mhz.

Expendable Reservation means that this resource pool can use the reservation of the parent pool.

All the reservation values don't matter if there isn't congestion (aka there are enough resources for all VMs of all resource pools).

--

There are 10 types of people. Those who understand binary and the rest. And those who understand gray-code.

[joshp]

Thank you; the explanation is clear.

[weinstein5]

To add to what has been said - the reservation of a resource pool is the resources set aside for the reservations of the vm's and child resource pools in the resource pool - so to use you example if the reservation of your resource pool is set to 100000 MHz - and you have the practice o build VM's with a CPU reservation of 10000 MHz and remembering a VM will not start if its reservation can not be satisfied - that you can only have 10 running VMs in this resource pool (10 x 10000 = 100000) with out expandable reservations enabled - those 10 VM's than can utilize up to the limit of the resource pool and if it is left unlimited it will utilize all the resouces -

Also with CPU you have to remember tha a virtual CPU will only run on one core at a time - for example a single vcpu vm running on an esx host with dual 3 GHz quad core CPU will only ever get a 3 GHz cpu - dual vcpu VM will get 2 3 GHz cores -