[openstack-dev] [release][all] What is upper-constraints.txt?

Amrith Kumar amrith at tesora.com
Wed Mar 30 14:30:00 UTC 2016


> -----Original Message-----
> From: Andreas Jaeger [mailto:aj at suse.com]
> Sent: Wednesday, March 30, 2016 9:03 AM
> To: OpenStack Development Mailing List (not for usage questions)
> <openstack-dev at lists.openstack.org>
> Subject: Re: [openstack-dev] [release][all] What is upper-constraints.txt?
> 
> On 2016-03-30 14:33, Davanum Srinivas wrote:
> > Folks,
> >
> > Quick primer/refresh because of some gate/CI issues we saw last few
> > days with Routes===2.3
> >
> > upper-constraints.txt is the current set of all the global libraries
> > that should be used by all the CI jobs.
> 
> We're not ready yet for such a general recommendation, see below for some
> details.
> 
> > This file is in the openstack/requirements repo:
> > http://git.openstack.org/cgit/openstack/requirements/tree/upper-constr
> > aints.txt
> > http://git.openstack.org/cgit/openstack/requirements/tree/upper-constr
> > aints.txt?h=stable/mitaka
> >
> > Anyone working on a project, please ensure that all CI jobs respect
> > constraints, example from trove below. If jobs don't respect
> > constraints then they are more likely to break:
> > https://review.openstack.org/#/c/298850/
> >
> > Anyone deploying openstack, please consult this file as it's the one
> > *sane* set of libraries that we test with.
> >
> > Yes, global-requirements.txt has the ranges that end up in project
> > requirements files. However, upper-constraints.txt is what we test for
> > sure in OpenStack CI.
> 
> 
> Note that upper-constraints is not ready for full usage in the project.
> It only works in check and gate jobs but especially does not work in post
> jobs. If you implement it improperly, your jobs will fail - and post jobs
> will fail silently (see also [1] for an infra discussion).
> Before infra can support this everywhere, our tools need to be fixed to
> handle constraints in all queues.
> 
> Right now, I consider upper-constraints experimental since it does not
> work for all queues and is not fool-proof. So, if you go down this road,
> triple check that *all* your jobs do the right thing,

[amrith] I'm really thrilled that dims sent this to the ML, and that we got your review comments Andreas. I'll try my best to fix them and ping you for a re-review.

> 
> Andreas
> 
> [1]
> http://eavesdrop.openstack.org/meetings/infra/2016/infra.2016-03-15-
> 19.03.html
> 
> 
> 
> --
>  Andreas Jaeger aj@{suse.com,opensuse.org} Twitter: jaegerandi
>   SUSE LINUX GmbH, Maxfeldstr. 5, 90409 Nürnberg, Germany
>    GF: Felix Imendörffer, Jane Smithard, Graham Norton,
>        HRB 21284 (AG Nürnberg)
>     GPG fingerprint = 93A3 365E CE47 B889 DF7F FED1 389A 563C C272 A126
> 
> 
> __________________________________________________________________________
> 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



More information about the OpenStack-dev mailing list