F40 CI Failures on Previously Passing Code - 23.11 branch

[Lincoln Lavoie]

So, the plot thickens a little bit.

On F44, it seems like 23.11 may have a compile issue with the Clang version
that comes with that OS release. What is the expectation / desire for
testing older LTS versions with updated OS releases (that technically came
out far after the original LTS was cut).

The only other issue with F43 and F44, that we are solving is the meson
(and by extension, the python versions), that need to be running on the
systems. But we have that pretty sorted, just needs to get rolled out into
the lab. Part of the "trick" there is, the LTS and main versions have
different min / max meson versions required or used.

Cheers,
Lincoln

On Wed, May 6, 2026 at 10:14 AM Aaron Conole <aconole at redhat.com> wrote:

> Patrick Robb <patrickrobb1997 at gmail.com> writes:
>
> > Just spitballing but if the team built a new F40 DPDK image from the
> > template engine, there might be a difference because a newer base
> > image of F40 or newer versions of dependencies were baked in (for any
> > which are not version tagged).
> >
> > Another thought - if F40 is EOL it might make sense to disable it in
> > CI and for Shani and the community to disregard this failing test. It
> > does make sense to enable the newer versions of fedora though.
>
> +1 to just disabling older fedora releases - F40 is EOL
>
> > I had a PR for enabling new fedora versions up when I was at UNH but
> > it was blocked because DPDK did not build on the newest fedora (this
> > is a pretty common occurrence for DPDK when bleeding edge fedora
> > versions come out) but most likely this is no longer an issue. If the
> > new fedora still wont build DPDK, that can be ticketed on Bugzilla and
> > someone in DPDK will figure it out. Sorry for leaving this as an open
> > item on my way out of UNH - I probably could have touched base with
> > the team better about this PR.
>
> +1 to this plan.
>
> >> On May 4, 2026, at 8:34 AM, Lincoln Lavoie <lylavoie at iol.unh.edu>
> wrote:
> >>
> >> Hi Shani,
> >>
> >> I've asked the team to look into it.  I know they had just done some
> >> work with the Fedora systems, to get the latest releases enabled and
> >> drop the versions that are EOL (that also includes Fedora 40, that
> >> went end of life in May of 2025).
> >>
> >> Cheers,
> >> Lincoln
> >>
> >>> On Mon, May 4, 2026 at 4:33 AM Shani Peretz <shperetz at nvidia.com>
> wrote:
> >>>
> >>> Hey,
> >>>
> >>> I’m seeing on F40 CI failures on code that previously passed.
> >>>
> >>> This reproduces on both the 23.11 and 23.11-staging branches. For
> >>> example, on 23.11, tarball 37012 passed while 37022 failed, even
> >>> though they contain the same code.
> >>>
> >>> Re-testing older commits that previously passed now fails with the
> >>> same errors.
> >>>
> >>> Were there any recent updates on the F40 CI side?
> >>>
> >>>
> >>>
> >>> Thanks a lot for your help,
> >>> Shani
> >>>
> >>>
> >>>
> >>>
> https://lab.dpdk.org/results/dashboard/periodic_testing/?branch_id=20&page=0&rerun_status=None
> >>
> >>
> >>
> >> --
> >> Lincoln Lavoie
> >> Principal Engineer, Broadband Technologies
> >> 21 Madbury Rd., Ste. 100, Durham, NH 03824
> >> lylavoie at iol.unh.edu
> >> https://www.iol.unh.edu
> >> +1-603-674-2755 (m)
>
>

-- 
*Lincoln Lavoie*
Principal Engineer, Broadband Technologies
21 Madbury Rd., Ste. 100, Durham, NH 03824
lylavoie at iol.unh.edu
https://www.iol.unh.edu
+1-603-674-2755 (m)
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mails.dpdk.org/archives/ci/attachments/20260506/2c63c0da/attachment.htm>