On Fri, Oct 23, 2015 at 06:21:23AM +0000, Tim Bell wrote:
> 
> On 22/10/15 20:01, "Marc Heckmann" <marc.heckm...@ubisoft.com> wrote:
> 
> >Hi,
> >
> >On Thu, 2015-10-22 at 08:17 -0700, Abel Lopez wrote:
> >> I've actually looked for this for our RBD backed ephemeral instances,
> >> but found the options lacking. I last looked in Juno. 
> >> 
> >> On Thursday, October 22, 2015, Tim Bell <tim.b...@cern.ch> wrote:
> >>          
> >>         
> >>         Has anyone had experience with setting up Nova with KVM so it
> >>         has raw ephemeral disks but qcow2 images for the VMs ? We’ve
> >>         got very large ephemeral disks and could benefit from the
> >>         performance of raw volumes for this.
> >
> >We looked into this for the very same reasons and it doesn't seem to be
> >supported.
> >
> >That being said, I'm fearful of the boot time performance impact of
> >using RAW for ephemeral. 
> >
> >I suggest you check out the following presentation about qcow2
> >performance if you haven't already done so:
> >
> >http://events.linuxfoundation.org/sites/events/files/slides/p0.pp_.pdf

I was about to recommend this (/me was present in this session in
person).

Also, slightly off-topic: I'd recommend the "Incremental Backups"
session:

    
http://events.linuxfoundation.org/sites/events/files/slides/kvm2015_rh_light_44_vfinal.pdf

> >I think it would be worthwhile for Openstack (and libvirt if required)
> >to support the "l2-cache-size" option for qcow2.
>
> +1 for l2-cache-size. This presentation was the basis under which I
> was looking for a temporary approach until we get the latest qcow2
> support.

Would you or Marc like to file a bug report to track this?

The 'l2-cache-size' runtime pption to '-drive' command-line argument is
a recent addition (AUG-2015):

    [...]-drive file=hd.qcow2,l2-cache-size=2097152[...]

-- 
/kashyap

_______________________________________________
OpenStack-operators mailing list
OpenStack-operators@lists.openstack.org
http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-operators

Reply via email to