[Openstack-operators] A Hypervisor supporting containers
matt at nycresistor.com
Fri May 2 04:32:27 UTC 2014
you know, i've never worked anywhere that didn't roll their own distro.
that being said, people ( manager types ) look for check marks on support
boxes... they start asking questions and getting all weary of the world
when you bring things like stackforge into the mix.
I am all for enforcing CI. But, my understanding of the workflow is code
doesn't go in without unit tests. Frankly you guys removing sections of
code for not having proper unit testing is downright terrifying. Doubly so
when it's major feature sets.
On Fri, May 2, 2014 at 12:26 AM, Michael Still <mikal at stillhq.com> wrote:
> lxc via libvirt is still supported. Docker is still supported via
> stackforge. Very little has been lost, especially if your distro is
> packaging them all together for you.
> This is all about making sure we do enough testing of these things, so
> that we have the highest quality drivers we can. Surely operators want
> that? The drama in the mean time is just the process of sorting it all
> On Fri, May 2, 2014 at 2:16 PM, matt <matt at nycresistor.com> wrote:
> > not going to get into the drama of the inner working of the dev teams...
> > but, losing container support from a release is a major feature hit
> > openstack. the people will lament.
> > On Fri, May 2, 2014 at 12:06 AM, Michael Still <mikal at stillhq.com>
> >> On Fri, May 2, 2014 at 1:39 AM, Eric Windisch <eric at windisch.us> wrote:
> >> > On Wed, Apr 30, 2014 at 7:59 PM, Michael Still <mikal at stillhq.com>
> >> > wrote:
> >> >>
> >> >> Yes, the matrix is trying to day that lxc under libvirt is not well
> >> >> tested. I don't think there are any plans to remove it... Its much
> >> >> more likely we'll fix the testing than remove that code.
> >> >>
> >> >> The biggest problem here is that no one is really actively working on
> >> >> container support.
> >> >
> >> > That's not true, is it? Much of the effort just happens to live in
> >> > stackforge right now. There are arguments that that work, by virtue of
> >> > being
> >> > in Stackforge, isn't "valid" or subsequently, "doesn't exist", I've
> >> > heard
> >> > such arguments in the past... but it doesn't invalidate the fact that
> >> > there
> >> > is legitimate work and effort toward containers (and Docker) in
> >> > OpenStack.
> >> >
> >> > Also, it isn't just me. The Docker driver (now in Stackforge) seems to
> >> > have
> >> > 6-10 active contributors. Missing features such as Neutron support
> >> > been
> >> > added. A multi-contributor effort to gate with Tempest is now in
> >> I think it was clear during Icehouse that you were the only person
> >> working on CI for docker, which was why it didn't make it before the
> >> deadline. I think you're doing good work and don't mean to diminish
> >> that. However, its good to hear you're getting some more help now.
> >> > I agree the dedicated containers API / service has had limited
> >> > The
> >> > proposal to discuss this during the summit was rejected.
> >> > the
> >> > unconference track was removed, but several of the interested parties
> >> > have
> >> > discussed organizing a less formal meeting co-located with the
> >> > conference.
> >> This is more what I was referring to. People have been working on a
> >> container service for an entire release and we've heard nothing. The
> >> summit session was an API brainstorming session -- that's what's meant
> >> to have been happening for the last six months. I think the nova team
> >> is waiting to see a concrete proposal, not yet more discussion.
> >> Michael
> >> --
> >> Rackspace Australia
> Rackspace Australia
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the OpenStack-operators