Open Stack

On 9/26/2016 5:15 PM, Dave McCowan (dmccowan) wrote:
> I don't know what triggered the update.  Our gates started breaking on
> September 23, but I can't find a commit around that time that would have
> caused this to happen.
>
> From: Clay Gerrard <clay.gerrard at gmail.com <mailto:clay.gerrard at gmail.com>>
> Reply-To: "OpenStack Development Mailing List (not for usage questions)"
> <openstack-dev at lists.openstack.org
> <mailto:openstack-dev at lists.openstack.org>>
> Date: Monday, September 26, 2016 at 6:03 PM
> To: "OpenStack Development Mailing List (not for usage questions)"
> <openstack-dev at lists.openstack.org
> <mailto:openstack-dev at lists.openstack.org>>
> Subject: Re: [openstack-dev] Pecan Version 1.2
>
> I'm interested to hear how this works out.
>
> I thought upper-constraints was somehow supposed to work to prevent
> this?  Like maybe don't install a brand new shiny upstream version on
> the gate infrastructure test jobs until it passes all our tests?
> Prevent a fire drill?  That bug was active back in July - but I guess
> 1.2 was released pretty recently?  .... maybe I don't understand the
> timeline.
>
> -Clay
>
> On Mon, Sep 26, 2016 at 2:21 PM, Dave McCowan (dmccowan)
> <dmccowan at cisco.com <mailto:dmccowan at cisco.com>> wrote:
>
>
>     The Barbican project uses Pecan as our web framework.
>
>     At some point recently, OpenStack started picking up their new
>     version 1.2.  This version [1] changed one of their APIs such that
>     certain calls that used to return 200 now return 204.  This has
>     caused immediate problems for Barbican (our gates for /master,
>     stable/newton, and stable/mitaka all fail) and a potential larger
>     impact (changing the return code of REST calls is not acceptable for
>     a stable API).
>
>     Before I start hacking three releases of Barbican to work around
>     Pecan's change, I'd like to ask:  are any other projects having
>     trouble with
>     Pecan Version 1.2?  Would it be possible/appropriate to block this
>     version as not working for OpenStack?
>
>     Thanks,
>     Dave McCowan
>
>
>     [1]
>     http://pecan.readthedocs.io/en/latest/changes.html
>     <http://pecan.readthedocs.io/en/latest/changes.html>
>     https://github.com/pecan/pecan/issues/72
>     <https://github.com/pecan/pecan/issues/72>
>
>
>     __________________________________________________________________________
>     OpenStack Development Mailing List (not for usage questions)
>     Unsubscribe:
>     OpenStack-dev-request at lists.openstack.org?subject:unsubscribe
>     <http://OpenStack-dev-request@lists.openstack.org?subject:unsubscribe>
>     http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
>     <http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev>
>
>
>
>
> __________________________________________________________________________
> OpenStack Development Mailing List (not for usage questions)
> Unsubscribe: OpenStack-dev-request at lists.openstack.org?subject:unsubscribe
> http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
>

There is a bot that updates upper-constraints, so it was updated here:

https://github.com/openstack/requirements/commit/21015dfb3c3e93222265721f589d11910a366f83

Reviews on these are basically, if they pass CI they get merged, unless 
we're in an release candidate mode, which for master we aren't anymore 
(since master is now ocata).

As fungi pointed out, there are some representative jobs run on these 
changes but it's not an exhaustive list, it's mostly the integrated-gate 
jobs, which barbican is not a part of which is how it slipped through.

By the way, you're broken on stable/mitaka because barbican isn't using 
upper-constraints in barbican. Note the version of pecan in 
stable/mitaka is 1.0.4. Same story for stable/newton, pecan is 1.1.2 in 
stable/newton and is frozen.

So a large part of the fix here is for barbican to use upper-constraints 
in it's unit test jobs. Looks like you can thank tonyb for doing this 
for you:

https://review.openstack.org/#/c/358404/

Which says it's also in stable/newton, so I don't know how you're busted 
in stable/newton.

-- 

Thanks,

Matt Riedemann