Working on speccing out some shared storage for running VMs, but one thing I am working on figuring out is how many machines a single storage node could support. I figure, throughput isn't going to be much of an issue. Don't usually see VMs pulling a lot of throughput, mostly IOps.
Curious what others factor in per VM when speccing out machines. Guess there are two tiers, one for your entry level machines and one for performance.
I've been figuring 60-70 for higher end nodes, since that equates to about one 7200rpm SATA drive per VM.