Hi,

On Fri, Apr 26, 2013 at 11:27 AM, Christoph Pleger <
[email protected]> wrote:

> Hello,
>
> > I'm not that much familiar with cgroups, but the way I understand it the
> first VM does not get 100%, it will get 1024 shares out of 1536 (2/3 of
> cpu
> > time).
>
> Yes, that's how cgroup works, and I wonder why this is not a conflict with
> what CPU= in VM Templates means
>

Could you please elaborate a bit more about this conflict?
The only "problem" I can think of is that VMs will get more cpu time that
requested if the Host isn't full; but we still ensure that the VM gets at
least as much as requested.


> and why the documentation even says that
> the cgroup "feature can be used to enforce the amount of CPU assigned to a
> VM, as defined in its template". What about enforcement without cgroup?
> Isn't there any?
>

We have the "logical" OpenNebula level enforcement, and the hypervisor
level.

First one: the OpenNebula scheduler will not allocate more cpu/memory than
the Host reports as available.

Hypervisor level reservation: cgroups for kvm, credit scheduler for xen,
and the esx cpu scheduler for vmware


>  I am also wondering why, in sunstone, one cloud node with one VM running
> on it shows a CPU usage of 6400. I thought that this is because the user
> set CPU=8, VCPU=8 and 8*8=64, but if CPU= is related to one physical CPU,
> that cannot be true ...
>

Can you please paste the output of 'onehost show <id>' ?


>
> Regards
>   Christoph
>
>
Regards

--
Carlos Martín, MSc
Project Engineer
OpenNebula - The Open-source Solution for Data Center Virtualization
www.OpenNebula.org <http://www.opennebula.org/> | [email protected] |
@OpenNebula <http://twitter.com/opennebula> <[email protected]>



>
>
> _______________________________________________
> Users mailing list
> [email protected]
> http://lists.opennebula.org/listinfo.cgi/users-opennebula.org
>
_______________________________________________
Users mailing list
[email protected]
http://lists.opennebula.org/listinfo.cgi/users-opennebula.org

Reply via email to