Hi, I was hoping for some help in understanding CPU load with multiple CPU's in ESXi, I will lay out the issue below and would really appreciate some advice
We have just virtualised a 2003 Std Terminal server, that has around 12 users that log onto it, before it was running on hardware with two Dual core Xeon 2.8ghz procs, and 4GM memory. The hardware was getting dated, and you get much more bang for buck out of hardware these days so VM it is.
Anyway, the new hardware is 2x X5650 procs, so total of 24 logical procs with hyper heading enabled.
I have allocated 4GB to the above converted VM, and 4x vCPU's so the allocated hardware is the same as it was in the past.
at the moment there are no other VM's running on that hardware, the plan is to migrate more onto it down the track though
Since the convert performance is terrible, the CPU's max out at 100% every time a couple of people browse the net, and a range of other light weight apps seem to put the CPU usage up way higher than they used to on the old hardware
other odd behaviour in the environment that wasn't happening before standard apps that users use every day will be fine 70% of the time, and then will stop loading and throw out strange messages, log off and on again usually fixes the issue, but this never occurred before the change over
There are numerous posts about this already in this forum so I do apologies for asking the same questions again, but I have read through the docs on the CPU scheduler and various other docs, and most of them seem to point towards CPU load sharing between VM's. as there are no other VM's on this host yet, I can't see why I would be getting such terrible performance
Can anyone share some expertise / recommendations for me here? I would like to keep the 4x CPU's available due to the amount of users so the resources are there if they are needed.
Thanks in advance for you help