Despite the plethora of discussion threads and documents I'm still trying to figure out something about memory usage reporting.
Forgive the bullets. It's how I organize my thoughts:
- If you click on a host in the cluster you see "Memory Usage" and "Capacity". Capacity is clearly the total amount of memory installed on the system. Is "usage" the total of all committed across all VMs or the total of all active memory on all VMs?
- If you go to a VM's summary tab you see Consumed Host Memory and Active Guest Memory. See picture below. Consumed Host Memory appears to be the amount allocated plus a little overhead. The Active Guest Memory doesn't match the amount cached (in use). In a simple way, why is this? Is it always less due to deduplication across VMs?
- We bought a certain amount of memory based on the expected needs of the total number of VMs we planned on deploying. Not it appears that due to some efficiency of VMware we bought WAY more than we needed since the total actively being used by each VM is far less than the committed amount almost always. It seems as if we can build and deploy far more VMs than our memory committment adds up to because of this. Am I wrong in this conclusion?
