[qa][stable] tempest.api.volume.test_versions.VersionsTest.test_show_version fails on stable/pike
Ghanshyam Mann
gmann at ghanshyammann.com
Thu Sep 26 00:16:31 UTC 2019
---- On Wed, 25 Sep 2019 13:20:23 -0500 Ghanshyam Mann <gmann at ghanshyammann.com> wrote ----
> ---- On Wed, 25 Sep 2019 09:51:15 -0500 Matt Riedemann <mriedemos at gmail.com> wrote ----
> > A new test was recently added to tempest [1] which uses the volume API
> > v3 client to list and show versions and compare those against the
> > response schema. It's failing against stable/pike changes [2] presumably
> > because the volume v3 API does not show the (deprecated) v1 API, though
> > I'm a bit confused because the test is doing a GET / and then for each
> > version in the response it's doing a GET /{version}/ so I'm not sure
> > what's up with cinder's pike API version interface where it would list a
> > version but not be able to show it.
>
> Yeah, it is not so clear from failure and logs that which version is missing.
> let me add more clear logs in tests.
>
> Seems like cinder v1 version got removed from the list as well as from show API at the same time
> - I03bf2db5bd7e2fdfb4f6032758ccaf2b348a82ba
Logs were missing for show_version because it directly call raw_request() which does not do logging.
Adding the logging for rest_client raw_request() also so that we can check the failed version info
- https://review.opendev.org/#/c/684858/
testing here -
-https://review.opendev.org/#/c/684859/
-gmann
>
> -gmann
>
> >
> > Anyway, it sounds like this is another case where we're going to have to
> > pin tempest to a tag in devstack on stable/pike to continue running
> > tempest jobs against stable/pike changes, similar to what recently
> > happened with stable/ocata [3].
> >
> > [1] https://review.opendev.org/#/c/680790/
> > [2]
> > https://145b0d5be77f0cbe05ff-d67a6226f6eb6d34ab4efd951744a16f.ssl.cf5.rackcdn.com/684658/1/check/tempest-full/9e1de42/testr_results.html.gz
> > [3] https://review.opendev.org/#/c/681950/
> >
> > --
> >
> > Thanks,
> >
> > Matt
> >
> >
>
More information about the openstack-discuss
mailing list