[pve-devel] cgroups implementation proposal for kvm (cpuunits support for kvm)

[Dietmar Maurer]

> >> I am not even sure if we need to control cpu.shares - users expext 1 CPU if
> they assign 1CPU?
> 
> cpu.share is working like cpuunits, it's not limiting the cpu, only give priorities
> between process when the host cpus are 100%


Seems the old faisched patch still works, for example try:

# kvm -id 300 -cpuunits 1000

Then you can see the created cgroup in:

# ls /proc/vz/fairsched/300/

So all we need to to is to add parameter --cupunits to qemu-server and pass it to kvm.

And we already do that with qemu-server! It is just disabled if you use virtio (because
we got strange behavior earlier)

- Dietmar