I'm writing a cluster-capacity script to ballpark the number or "standard" VMs worth of capacity left in the environment. We are using vKernel products already, but want to verify some of the capacity numbers ourselves.
I was going to use the following base calculations.....
HA Overhead = FailureTolerance / Number of hosts
Available capacity(CPU) = (Totalmhz - (HA_Overhead * totalmhz) - usagemhz.maximum)
Available capacity(Mem) = (Totalmb - (HA_Overhead * totalmb) - usagemb.maximum)
However, I've hit a couple of snags..... firstly the stats I get through vCenter don't match the ones I get from powercli, 2nd: in vcenter the Usedmhz is greater than the totalmhz...
Any suggestions? I suspect its a result of the way things are aggregated.... but I'd like to confirm that and maybe get some suggestions as to how I might better do this.
Script, screenshot of one of the clusters and dump of the CPU stats attached.
Thanks in advance,
vCenter is 4.0 and all the ESX servers are 4.0U2.
I'll check the powercli version when I get back to the office on monday.
Thanks for the help so far.
VMware vSphere PowerCLI 4.1 build 264274
Really at a loss to explain this behavior... Any help is appreciated.