[libvirt] Live migration woes

[Thomas Treutner]

On Friday 05 February 2010 21:21:32 Thomas Sjolshagen wrote:
> Quoting Thomas Treutner <thomas at scripty.at>:
> > Hi,
> >
> > is anyone using libvirt-git and qemu-kvm post-0.11 and can live migrate
> > VMs successfully for- and backwards? qemu-kvm-0.11 seems to be the latest
> > release were live migration works, and I'm trying to figure out if its a
> > kvm or libvirt problem, since there were other things (balloon parameter
> > f.e.) that were changed with qemu-kvm-0.12 and caused troubles when using
> > libvirt.
>
> Yes.
>
> I'm using the virt-preview repo (rawhide) and I just noticed, today,
> after re-upgrading to 0.12.2-5 (since -4 has a rather nasty IO problem
> for virtio devices against raw image files) that my live-migrations
> fail with the following error message on the destination system (i.e.
> the system the I'm trying to migrate to):

I now switched back to 0.11 and disabled Enhanced C1 of my PhenomII and 
AthlonII (I think the MSI BIOS calls it just C1), which makes things better, 
but migrations are not completely stable yet. 


>From http://tinyurl.com/yaacdtu:
"In addition, we have C-states, in the case of the Phenom, it is the C0 and C1 
state with the C1 state being enabled if there are no outstanding IPCs and 
resulting in a clock divider of 1/128. If there is no cache activity either 
or outstanding memory requests, then the CPU can go into the enhanced C1 
state, that is the C1e state where the shared cache and memory controllers 
are clock gated back to 1/16 of the original frequency."


So, currently, it seems to be a KVM-bug related to power saving features.


-t