Re: make check failures in run-tox-qa

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On Fri, Apr 22, 2022 at 11:34 PM kefu chai <tchaikov@xxxxxxxxx> wrote:
>
> hi Casey,
>
> On Sat, Apr 23, 2022 at 2:04 AM Casey Bodley <cbodley@xxxxxxxxxx> wrote:
> >
> > i'm seeing a lot of these failures on PR checks lately:
> >
> > >   from nose.tools import eq_ as eq
> > E   ModuleNotFoundError: No module named 'nose'
> >
> > tasks/rgw_multi/tests.py:18: ModuleNotFoundError
> > ----------------------------- Captured stdout call -----------------------------
> >
> > (example from https://jenkins.ceph.com/job/ceph-pull-requests/94129/consoleFull#1181206387c19247c4-fcb7-4c61-9a5d-7e2b9731c678)
> >
> > these errors are coming from rgw multisite test cases, which
> > 'run-tox-qa' appears to be linting? we haven't made any interesting
> > changes to the tests, so i'm not sure whether something changed in the
> > lab, or if it's just the regular kind of python breakage. does anyone
> > know what's up with the missing 'nose'?
> >
>
> my guess is that this regression is introduced by
> https://github.com/ceph/teuthology/pull/1743/files#diff-4d7c51b1efe9043e44439a949dfd92e5827321b34082903477fd04876edb7552L146.
> before this change, the teuthology python package had been preparing
> the "nose" for the tests. after the overhaul of python packaging, nose
> was somehow dropped in
> https://github.com/ceph/teuthology/pull/1743/commits/21372fd9b2b87bb6417929f10097fba7db9cdca9.
> i failed to find the explanation in the commit message though. if we
> should follow the protocol of "hey, teuthology! please prepare the
> test dependencies for us, as we are imported by you, and are relying
> on you to get us settled before we start doing our business", we need
> to add "nose" to the options.extras_require section in
> teuthology/setup.cfg. IMHO, we should follow this protocol unless we
> develop another way to prepare the deps for these innocent tests.
>
> just created https://github.com/ceph/teuthology/pull/1749 to address this issue.

thanks!

>
> > separately, it seems a bit silly for 'make check' to be doing anything
> > with our crusty old python tests. i don't think anyone's concerned
> > about code quality there, only that the tests run and pass in
> > teuthology. is there a way to disable these checks under
> > qa/tasks/rgw_multi/?
>
> this change was introduced back when we were migrating the tests from
> python2 to python3. Thomas kindly helped to create
> https://github.com/ceph/ceph/pull/32129 in hope to ensure that we are
> python3 compatible and don't break the compatibility by adding this
> guard as part of "make check" run. i agree that we don't need to keep
> the same standard of code quality there as the one we are trying to
> keep under src/, but it probably still makes sense to have a minimal
> check for qa/ by running the python interpreter against them?
>
> >
> > _______________________________________________
> > Dev mailing list -- dev@xxxxxxx
> > To unsubscribe send an email to dev-leave@xxxxxxx
>
>
>
> --
> Regards
> Kefu Chai
>

_______________________________________________
Dev mailing list -- dev@xxxxxxx
To unsubscribe send an email to dev-leave@xxxxxxx



[Index of Archives]     [CEPH Users]     [Ceph Devel]     [Ceph Large]     [Information on CEPH]     [Linux BTRFS]     [Linux USB Devel]     [Video for Linux]     [Linux Audio Users]     [Yosemite News]     [Linux Kernel]     [Linux SCSI]

  Powered by Linux