Triage and migration report - Friday

[Christian Ehrhardt]

On Fri, Mar 6, 2020 at 2:21 PM Christian Ehrhardt <
christian.ehrhardt at canonical.com> wrote:

> ## Triage ##
>
> I got 21 bugs
>
> I was happy to find that most of them were bumps to fix released or fixes
> that were started.
> One can see that the Feature Freeze has passed and bug work is increasing
> again.
>
> The TLS v1.3 bugs were mentioned by others yesterday, no further updates
> to those yet since yesterday.
>
> Three bugs needed a proper initial triage and some minor testing.
>
> One of these that is worth mentioning is:
>   https://bugs.launchpad.net/ubuntu/+source/clamav/+bug/1865818
> I think we should try to take a look at picking the patch and modifying
> the postinst beofre 20.04.
> Subscribed the team and tagged as server-next.
>
>
> ## Proposed migration ##
>
> Most things in our queue are valid but depending on bigger migrations.
> The obvious ruby/php candidates are being worked on.
>
> The one thing worth to look at (again) is python-cffi which breaks and
> holds back python-defaults a while now.
> Paride mentioned that test on 2020-02-27:
> It seems the autopkgtest run got a sigkill:
>
> [...]
> ..............................s...s.............ss.Killed
>
> It was even mentioned before on 2020-01-16 missing dep: libffi but that is
> good now.
>
>
> Nowadays it is important to know that there is a new version of it
>  python-cffi | 1.13.2-1build3 | focal/universe          | all
>  python-cffi | 1.14.0-1build1 | focal-proposed          | source
>  python-cffi | 1.14.0-1build1 | focal-proposed/universe | all
>
> Recent tests had:
> - the new python3-defaults against the old version of python-cffi
> - the new python-cffi against python3-defaults
>
> One outlier also had "ModuleNotFoundError". But that was before a rebuild
> to drop python 3.7 and is resolved.
> The other fails of the new version still fail with the "Killed" message.
>
> Lets try to run that with new python + new python-cffi and do so in a
> local VM for debugging right away.
>
> The test has a python2 and a python3 test component.
> In the past both failed with the kill-symptom.
>
> I ran the new cffi against three sets:
> - old python3-defaults
> - new python3-defaults
> - all-proposed
>
> And of course, they ALL passed the tests :-/
> So much for reproducibility.
>
> ====== 1912 passed, 88 skipped, 4 xfailed, 206 warnings in 447.83 seconds
> ======
> ====== 1912 passed, 88 skipped, 4 xfailed, 231 warnings in 492.50 seconds
> ======
>
> I was running the case at lower memory then, in case the issue would be an
> OOM (to see if it breaks at the same place). And that indeed is it, hitting
> the same spot now.
>
> So it is quite likely that if we mark it as huge test it will work.
> After Andreas was so kind to refresh my memory on how this is done
> (Thanks!) I was opening an MP for that.
> =>
> https://code.launchpad.net/~paelzer/autopkgtest-cloud/+git/autopkgtest-cloud/+merge/380362
>

Merged and re-tests worked, python3-cffi should now drop out of our
proposed issues.


> --
> Christian Ehrhardt
> Staff Engineer, Ubuntu Server
> Canonical Ltd
>


-- 
Christian Ehrhardt
Staff Engineer, Ubuntu Server
Canonical Ltd
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.ubuntu.com/archives/ubuntu-server/attachments/20200306/e0683557/attachment.html>