[virt-tools-list] Fedora 13 beta - virtualization testing - comments/question

[Daniel P. Berrange]

On Tue, May 04, 2010 at 09:01:29PM -0500, Mike Hinz wrote:
> 
> On Mon, 2010-05-03 at 20:12 +0100, Richard W.M. Jones wrote:
> > On Sat, May 01, 2010 at 12:45:44PM -0500, Mike Hinz wrote:
> > > - The cpu usage for a single Win XP seem high at about 22% CPU looking
> > > at qemu in TOP.  This is with VM minimized with absolutely nothing
> > > running on it.
> > 
> > Also try virt-top.
> > 
> > > - USB hotplug is a bit of a mixed page:
> > > [...]
> > 
> > Many of these things will be issues with the hypervisor (ie. KVM)
> > itself.  You might try looking in /var/log/libvirt/qemu/... and file
> > bugs as appropriate.  On the other hand, USB hardware is so varied
> > that it's hard for the developers to reproduce issues.
> > 
> > > I still need to try storage pools on NFS and test other functionality,
> > > but I thought it might be helpful to put this report out now as the
> > > release date is getting closer.
> > 
> > Don't forget to test the libguestfs tools!
> > 
> > Rich.
> > 
> Rich,
> 
> Interestingly, after doing the latest updates as of 4 May evening, the
> cpu usage is considerably more reasonable.  Now with an idle Win XP VM,
> TOP reports qemu-kvm using slightly < 8% CPU.  Virt-top reports always
> about 1/2 of what TOP reports, in this case slightly < 4%.  Is this
> expected behavior?   Uname now reports as follows:

It depends what you consider 100% to be.  100% can refer to either the
cumulative capacity of all CPUs, or 100% can refer to capacity of a single
CPU. Thus if you have 2 CPUs in a machine, 8% vs 4% can both be correct
if they're using different frames of reference.


Daniel
-- 
|: Red Hat, Engineering, London    -o-   http://people.redhat.com/berrange/ :|
|: http://libvirt.org -o- http://virt-manager.org -o- http://deltacloud.org :|
|: http://autobuild.org        -o-         http://search.cpan.org/~danberr/ :|
|: GnuPG: 7D3B9505  -o-   F3C9 553F A1DA 4AC2 5648 23C1 B3DF F742 7D3B 9505 :|