[dpdk-ci] UNH CI failing

[Lincoln Lavoie]

I'm continuing to hunt. I did confirm all of those systems are syncing time
from the same master, using chronyd, and look to be configured correctly.
I'm not sure what would cause the time to jump like that, unless maybe the
"master", which is our IPA server, synced it's time and that caused some
sort of ripple to the downstream other systems.

The patches have been rerun, with most results coming in,
https://lab.dpdk.org/results/dashboard/patchsets/17130/, and are passing
without issue.

The only failures I see in recent patches are on "bugfix for Kunpeng SVE
compile" (https://lab.dpdk.org/results/dashboard/patchsets/17135/), with
unit tests failing on two OSes.

And the intermittent Dynamic Config failure.  We think we have tracked this
down to a patch we sent into DTS, and that was merged into DTS, but was
then reverted a few days ago (
https://git.dpdk.org/tools/dts/commit/?id=90f460df240b3020191916b15705abe208a14694).
I've asked Lijuan why it was reverted.

Cheers,
Lincoln

On Wed, May 19, 2021 at 11:55 AM Aaron Conole <aconole at redhat.com> wrote:

> Lincoln Lavoie <lylavoie at iol.unh.edu> writes:
>
> > As far as I can tell, it looks like one of the clocks fell out of sync
> on the container runners, which caused the
> > builds to fail.
>
> I think that's also been causing some failures with the alarm_test and
> cycles_test unit tests.  If the time source is making adjustments to
> time, we will probably fail these tests as well.
>
> > Also, as far I can tell from an initial look, it impacted the two
> patches Thomas cited.  Two patches that are
> > running now (i.e. they don't have a full set of results yet, look like
> they are running on.  So, it was a transient
> > issue.  Obviously we need to track down its root cause. I suspect
> something happened with NTP, which should
> > be keeping the runners and bare metal systems synced.  I'm looking into
> that now.
> >
> > For the patches with the failed jobs, we will queue those for rerun
> today.
> >
> > Cheers,
> > Lincoln
> >
> > On Wed, May 19, 2021 at 9:05 AM Aaron Conole <aconole at redhat.com> wrote:
> >
> >  Thomas Monjalon <thomas at monjalon.net> writes:
> >
> >  > It seems the IOL CI is failing today:
> >  >
> >  >
> >
> https://patches.dpdk.org/project/dpdk/patch/1621406749-15536-1-git-send-email-changpeng.liu@intel.com/
> >
> >  >
> >
> https://patches.dpdk.org/project/dpdk/patch/20210519032745.707639-1-stevex.yang@intel.com/
> >
> >  >
> >  > That's especially embarassing for closing the release.
> >
> >  I don't see any useful logs in the failures.
> >
> >  What changed?
>
>

-- 
*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)
<https://www.iol.unh.edu>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mails.dpdk.org/archives/ci/attachments/20210519/3a963c8b/attachment-0001.htm>