DPDK CI discussions
 help / color / mirror / Atom feed
From: "Juraj Linkeš" <juraj.linkes@pantheon.tech>
To: Aaron Conole <aconole@redhat.com>
Cc: "ci@dpdk.org" <ci@dpdk.org>
Subject: Re: [dpdk-ci] OBS pull request in pw-ci
Date: Fri, 17 Jul 2020 07:05:51 +0000	[thread overview]
Message-ID: <73bd99e112d74c1fa5e60872a27ce62e@pantheon.tech> (raw)
In-Reply-To: <f7t5zansjrp.fsf@dhcp-25.97.bos.redhat.com>

Hi Aaron,

By community maintained I meant that it's not solely on one person's shoulders - that is, if need be, someone else (from the community) would also be able to do the maintenance work.

Juraj

> -----Original Message-----
> From: Aaron Conole <aconole@redhat.com>
> Sent: Thursday, July 16, 2020 2:54 PM
> To: Juraj Linkeš <juraj.linkes@pantheon.tech>
> Cc: ci@dpdk.org
> Subject: Re: OBS pull request in pw-ci
> 
> Juraj Linkeš <juraj.linkes@pantheon.tech> writes:
> 
> > Hi Aaron,
> >
> > The dpdk community decided that the best place to maintaing the .spec
> > and .control files would be in an OBS repository. Luca is currently
> > doing that so that's where it comes from. As to how to update it - I
> > included the exact string "home:bluca:dpdk" so that it's clear we want
> > to use this repo, but it should be handled the same way as
> > UPLOAD_GIT_REPO. I didn't see how this is supplied from the code in
> > pw-ci, so I assumed it's something that's changed locally. I followed
> > the same logic for OBS_SOURCE_PACKAGE and OBS_GIT_URL - the values
> > that are there are what we should be using for our use case at this
> > point in time.
> >
> > I have a minor question about the pw-ci repo - is this your personal
> > repo (i.e. not maintained by dpdk community)?
> 
> I maintain it - but it's publicly accessible / available.  I don't know what else it
> means.
> 
> > One last thing - we'll talk today about setting up a community OBS
> > project (to mirror https://travis-ci.com/github/ovsrobot/dpdk - is
> > this actually managed by the community?) in the CI call.
> 
> Again, I maintain it - but it's for the community.  Actually the service that runs
> the robot is a Red Hat server, but we choose to push it to a public github and use
> a public CI so that it is all part of the community.
> 
> I don't know what it means to be community maintained ;)
> 
> > Thanks,
> > Juraj
> >
> >> -----Original Message-----
> >> From: Aaron Conole <aconole@redhat.com>
> >> Sent: Wednesday, July 15, 2020 9:14 PM
> >> To: Juraj Linkeš <juraj.linkes@pantheon.tech>
> >> Cc: ci@dpdk.org
> >> Subject: Re: OBS pull request in pw-ci
> >>
> >> Aaron Conole <aconole@redhat.com> writes:
> >>
> >> > Juraj Linkeš <juraj.linkes@pantheon.tech> writes:
> >> >
> >> >> Hi Aaron,
> >> >>
> >> >> Have you had a chance to look at the pull request?
> >> >
> >> > Only a bit.  I will do something more thorough on Wednesday.
> >>
> >> I looked through it and it seems okay.  I'll set it up to do a test
> >> run tomorrow, and merge it once that goes well.
> >>
> >> If the obs source project changes (right now, it's set to
> >> home:bluca:dpdk) how should we update it?  Meaning, where does that
> >> variable come from (and if the robot needs to provide one at some
> >> point ever, how can we do that)?
> >>
> >> >> Thanks,
> >> >> Juraj
> >> >>
> >> >>> -----Original Message-----
> >> >>> From: Aaron Conole <aconole@redhat.com>
> >> >>> Sent: Monday, July 6, 2020 7:32 PM
> >> >>> To: Juraj Linkeš <juraj.linkes@pantheon.tech>
> >> >>> Cc: ci@dpdk.org
> >> >>> Subject: Re: OBS pull request in pw-ci
> >> >>>
> >> >>> Juraj Linkeš <juraj.linkes@pantheon.tech> writes:
> >> >>>
> >> >>> > Hi folks,
> >> >>> >
> >> >>> >
> >> >>> >
> >> >>> > I've submitted a PR with OBS changes to pw-ci:
> >> >>> > https://github.com/orgcandman/pw-ci/pull/12
> >> >>>
> >> >>> Cool!  I'll take a look.  I was on PTO, so didn't get a chance yet.
> >> >>>
> >> >>> >
> >> >>> >
> >> >>> > I have a couple of points/questions:
> >> >>> >
> >> >>> > ·        The scripts are using
> >> >>> https://build.opensuse.org/package/show/home:bluca:dpdk/dpdk to
> >> >>> fork the debian and
> >> >>> > redhat spec and control files.
> >> >>>
> >> >>> Okay.  A quick look shows that it's configurable, though.  Should
> >> >>> be okay.  I refer
> >> >>> to:
> >> >>>
> >> >>> https://github.com/orgcandman/pw-
> >> >>>
> ci/pull/12/commits/93f6baa497c85d7aca36cda6661e90f7a69d4709#diff-
> >> >>> e2ab32f32889186bdeb0ab4f0ef55282
> >> >>>
> >> >>> > ·        We should create a community OBS project (that the community
> can
> >> >>> manage) where the forked packages will be
> >> >>> > created and where the build will run (OBS_TARGET_PROJECT in
> >> >>> > 3rd-party/dpdk/jenkins-rc)
> >> >>>
> >> >>> Okay.
> >> >>>
> >> >>> > o   The project should be configured according to what we want to
> build -
> >> >>> distros/architectures. Can be done via OBS GUI.
> >> >>> >
> >> >>> > ·        Only after all patches have been submitted to a series branch does
> >> the
> >> >>> DPDK Jenkins job upload and configure an OBS
> >> >>> > package. This means that we'll have a build just for the whole
> >> >>> > series, not for
> >> >>> each patch. Not sure if this is a problem.
> >> >>>
> >> >>> The flip side is the amount of time to build.  We'd prefer to
> >> >>> build just one at a time.
> >> >>>
> >> >>> > ·        One thing that doesn't make sense to me is that the DPDK Jenkins
> job
> >> >>> builds dpdk after each patch - is this needed,
> >> >>> > since Travis is going to do the same thing anyway?
> >> >>>
> >> >>> Travis doesn't.  The ovsrobot travis build is configured to
> >> >>> cancel a build on a branch if a new one comes.  That way we only
> >> >>> build the latest.  Builds on travis take 40+ min to complete, so
> >> >>> if a 20-patch series came in, we wouldn't be done building it for
> >> >>> more than half a day.  Then if a v2+ is posted in rapid
> >> >>> succession, we lost all this time.  So for now, we only build the
> >> >>> absolute last patch in the series on
> >> Travis.
> >> >>>
> >> >>> Also, the build server we run the robot on doesn't do the builds
> >> >>> anymore because it also got overloaded - so meson and ninja are
> >> >>> not the actual meson/ninja but just dummy commands.  :-/  Maybe I
> >> >>> should re-enable and see if things improved.
> >> >>>
> >> >>> > ·        I've added some more comments to the PR itself.
> >> >>> >
> >> >>> >
> >> >>> >
> >> >>> > Regards,
> >> >>> >
> >> >>> > Juraj
> >> >>>
> >>
> 


  reply	other threads:[~2020-07-17  7:05 UTC|newest]

Thread overview: 9+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-07-03 13:08 Juraj Linkeš
2020-07-06 17:31 ` Aaron Conole
2020-07-13 14:20   ` Juraj Linkeš
2020-07-13 20:17     ` Aaron Conole
2020-07-15 19:13       ` Aaron Conole
2020-07-16 11:58         ` Juraj Linkeš
2020-07-16 12:54           ` Aaron Conole
2020-07-17  7:05             ` Juraj Linkeš [this message]
2020-07-17 13:24               ` Aaron Conole

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=73bd99e112d74c1fa5e60872a27ce62e@pantheon.tech \
    --to=juraj.linkes@pantheon.tech \
    --cc=aconole@redhat.com \
    --cc=ci@dpdk.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).