what are some good guidelines for calculating memory overcommit, and what is the best way to determine overcommit on a vm by vm basis? (like what metrics in vCenter would you look at on the ESX host level and vm level to make the decision?)
Also, kind of the same question for virtual processors. I always make my VMs with 2, but what is the guideline for when to add 4 or more procs?
The best suggestion here is a question. What do the Virtual Machines need? Don't assign more resources than can actually be used. Use guest based tools to compile usage statistics and then make changes accordingly. Have a look at http://www.vizioncore.com/free/vFoglight-quick-view
downloading foglight quickview now....
I have always found it easier to start small and add when needed. I always do single vCPU to start with it and increase to two if needed. We have had a couple of vendors that stood by their requirement of 4 cpus and after we performed benchmarking the machines were just as efficient with two. I've included a link to a pdf on the cpu scheduler. It's a great read that helps understand how CPU resources are scheduled.
http://www.vmware.com/files/pdf/perf-vsphere-cpu_scheduler.pdf
If you found this or any other post helpful please consider the use of the Helpful/Correct buttons to award points