[openstack-dev] [nova] [infra] The same SRIOV / NFV CI failures missed a regression, why?
fungi at yuggoth.org
Mon Apr 4 22:56:22 UTC 2016
On 2016-04-04 13:54:34 -0400 (-0400), Jay Pipes wrote:
> On 03/30/2016 11:00 PM, Robert Collins wrote:
> >On 26 March 2016 at 09:08, Jeremy Stanley <fungi at yuggoth.org> wrote:
> >>On 2016-03-25 15:20:00 -0400 (-0400), Jay Pipes wrote:
> >>>3) The upstream Infrastructure team works with the hired system
> >>>administrators to create a single CI system that can spawn
> >>>functional test jobs on the lab hardware and report results back
> >>>to upstream Gerrit
> >>This bit is something the TripleO team has struggled to accomplish
> >>over the past several years (running a custom OpenStack deployment
> >>tied directly into our CI), so at a minimum we'd want to know how
> >>the proposed implementation would succeed in ways that they've so
> >>far found a significant challenge even with a larger sysadmin team
> >>than you estimate being required.
> >I think what Jay is getting at is to have the *exact same approach*
> >third-party CI for NFV and PCI have been using - so whatever
> >$behind-the-abstraction setup they are using, but community accessible
> >and visible, unlike the current behind-corprorate-firewall setups.
> >I'm not saying this is better or worse, but it is different to the
> >tripleo approach of providing a Nova API endpoint for zuul.
> Yes, thank you Rob, that is precisely what I'm getting at.
In that case, I'm not sure a third-party CI system needs close
coordination with "The upstream Infrastructure team" nor "hired
system administrators" employed by the OpenStack Foundation, which
were the parts of the original proposal I was concerned with. Set
up a third-party CI system and start voting on changes (with the
consent of those projects anyway).
More information about the OpenStack-dev