On Wed, Aug 20, 2014 at 12:07 PM, Karanbir Singh mail-lists@karan.org wrote:
On 08/11/2014 06:58 PM, George Dunlap wrote:
On Mon, Aug 11, 2014 at 12:01 PM, Russell Bryant rbryant@redhat.com wrote:
That's right. We want to test OpenStack with bleeding edge versions of libvirt and qemu, but want the underlying OS to be something supported for a longer period of time than a given Fedora release. CentOS + a testing repo with that software included would be perfect.
But on the whole, it sounds like your goals and the goals of the Virt SIG are at odds. The Virt SIG wants to provide a stable base; the "(1)" group of people mentioned in the Fedora virt-preview wiki page. As it happens, we plan on updating our libvirt package fairly frequently at first; but that's just to get some important Xen functionality in as soon as possible. Once the Xen functionality for libvirt has stabilized, we'll probably stop. Our plan for qemu was to re-build the exact RHEL package, but with snapshotting enabled.
What you're describing would essentially be a completely separate project: designed for people (like yourselves) who want bleeding-edge versions.
could we do this as a part of a -testing or -next repo, but still be a part of the VirtSIG ? I think it would be great to see some of the upstream devel stuff being built and tested, specially if it can be automated. Needing to do this manually, and curate it locally would be quite hard.
Yeah, I can see the usefulness of that -- particularly with me "upstream" hat on. It's just a matter of effort and priorities. :-)
And it would be different from Fedora's virt-next, because it would be focusing on virtualization stuff coming down the pipeline from individual projects upstream, rather than virtualization stuff likely to end up in the next CentOS.
-George