Hi,
I'm just wondering how people capacity manage their ESXi hosts in regards to memory?? In our old 4.1 environment we disable TPS so large pages werent used and we saw around a 20% redunction on our hosts.
We have since upgraded to 5.1 u1a and new hardware (256Gb instead of 148Gb) and left TPS as default settings and the memory usage is around 60-80% used on the hosts.
From my understanding TPS doesnt use small pages until over 80%
So if you have a 11 node cluster with memory (default TPS) at around 80% on each how do you know how what the capacity left is? If you lost two ESXi will the cluster be ok?
Is it worth disabling large page share and take the so called performance hit? This will therefor be easier to manage the memory on the hosts and when to buy additional hardware.