On Tue, Aug 4, 2020, 01:50 Michel Alexandre Salim <michel@michel-slm.name> wrote:
Hi Fabio,

On Mon, 2020-08-03 at 19:29 +0200, Fabio Valentini wrote:
> On Mon, Aug 3, 2020 at 6:59 PM Hans de Goede <hdegoede@redhat.com>
> wrote:
> > Hi,
> >
> > On 8/3/20 5:53 PM, Kevin Fenzi wrote:
> > > On Mon, Aug 03, 2020 at 05:21:58PM +0200, Hans de Goede wrote:
> > > > Hi All,
> > > >
> > > > <grumpy mode>
> > > > I just noticed that a lot my packages got a FTBFS because of
> > > > failing to build on s390x. The first set of rebuilds failed
> > > > with:
> > > >
> > > > "BuildrootError: Requested repo (1785390) is DELETED"
> > > >
> > > > The second set of rebuilds failed with:
> > > >
> > > > "rpm.error: error reading package header"
> > > >
> > > > errors.
> > > >
> > > > The last error was also seen quite a bit during the F32 mass
> > > > rebuid ...
> > >
> > > I'm sorry this is happening, and it makes me very grumpy too.
> > >
> > > I have some thoughts on improvements we can make to help try and
> > > make
> > > this better, but I was under the impression it was mostly working
> > > ok for
> > > the second pass.
> > >
> > > We went from 4162 to 2833 failures, so it had to have been
> > > working at
> > > least sometime there?
> >
> > It seems for me the s390x failures on the second build are limited
> > to package names starting with A-Z and "aa*" - "an*" .
> >
> > Any chance we can get a third mass rebuild for package-names
> > starting
> > with A-Z and "a*" ?
> >
> > Or maybe all those where the only failing platform is on s390x ?
> >
> > (no idea how easy it is to script any of this)
>
> I am already resubmitting all builds that failed in koji but that
> currently pass locally in mock with "--enablerepo local".
> So far this has reduced the number of FTBFS packages by almost 100,
> and the script is still running.
> This should take care of all packages that only failed due to infra
> issues.
>

(snip)


Could this be why the queue to get an s390x build seems longer than
usual today, and is there an ETA for when this script will finish?

I have some builds for Eternal Terminal (et) that have all completed on
all platforms (even armv7hl) but are still all queued on s390x

https://koji.fedoraproject.org/koji/packageinfo?packageID=27370

Thanks,

Almost certainly no. Since I'm running a local mock build as sanity check before submitting anything to koji, there's usually no more than 3-4 builds running at one point, and I'm submitting them with both --fail-fast (if they fail anywhere, other arches are cancelled immediately) and --background (lower priority).

Fabio


--
Michel Alexandre Salim
profile: https://keybase.io/michel_slm
chat via email: https://delta.chat/
GPG key: 5DCE 2E7E 9C3B 1CFF D335 C1D7 8B22 9D2F 7CCC 04F2
_______________________________________________
devel mailing list -- devel@lists.fedoraproject.org
To unsubscribe send an email to devel-leave@lists.fedoraproject.org
Fedora Code of Conduct: https://docs.fedoraproject.org/en-US/project/code-of-conduct/
List Guidelines: https://fedoraproject.org/wiki/Mailing_list_guidelines
List Archives: https://lists.fedoraproject.org/archives/list/devel@lists.fedoraproject.org