Clearly I should have given more detail. This is for a thin client system running in the local school. Currently we have two servers each serving part of the school, but sometimes run into problems with one of the servers getting overloaded (clearly we can’t afford to just throw more processing at it). Have been looking into alternative VM software which provides clustering, but none of them seem to provide the functionality of our current system (well the paid for ones might, but we need something free), hence thinking of doing it manually using our current server software.
Therefore I’m only looking for the number of VM processes currently running (top wouldn’t help much as I’m not interested in the current load, but the potential load from the number of concurrent VMs – typically new VM sessions get started when other sessions are idle). Can’t really just alternate stuff on the servers, as each client only ever fires up a single process and I don’t know which ones will get turned on – though as a first step to improve the current system we’re going to alternate them physically which should work fairly well – as suggested we’re also after improved availability.
Hopefully I’m not quite as daft as it would seem from not remembering I could do what I wanted using ps.