I was playing with RP tonight specifically memory setting. I noticed that when I disabled expandable reservation
It required me to set a minimum reservation for the vms to even boot up in the rp even though there was no contention on the box.
The reservation that needed to be set seemed to be dependent on how many vm's I put in the pool
and how my vcpu i had in each vm. It did not seem to care how much memory I configured each vm at. (512, 1 GB, 2 GB ect.. ) I am assuming because the more vcpu a vm has the more overhead to run it.
I am wondering what logic it is using, and I did not see this listed as a published number in the resource guide.