VMware Environment:
- 9 Hosts
- 43 Guests
- vSphere 5.5
- vCenter
- NO vCOPs
- NO Centralized Storage
We are planning to replace our VMware hosts (HP blades with local datastores) with 2 hosts that will be configured with enough local storage and memory to support all of our current guests and any additional guests that might be needed in the next year. Although there are a number of things to consider I am focused right now on determining how much memory we need to put in each of these hosts. The company management doesn't want to spend the money on 32GB sticks of RAM which, by my calculations, are the minimum size we need for our setup. Our setup will be to split 50 guests between both hosts leaving enough resources for each host to support ALL of the guests should the other host fail and, of course, to add guests as needed.
Which brings me to the guest memory configuration. We have a significant number of guests that have been configured with 16GB of memory per the software developer's spec. The majority of these guests are using between 8 and 16GB of memory according to the Consumed Host Memory statistic BUT the Active Guest Memory spec is substantially lower. I read a comment from someone in another forum that advised the OP to give the server ONLY what it needs to run (referncing only the Host Memory Consumed stat and the Guest Memory Active stat) and at this point I'm a little confused. So I'm hoping someone can help me figure out how to squeeze as many VMs on a host as possible by limiting the amount of memory that is available to a guest to just what it needs. I've attached a screenshot of a typical guest. Is there a way to reconfigure the memory settings to allow this VM to continue to run well while freeing up memory on the host?
Thanks for your help.
MJ
