Open Stack

On Wed, Sep 21, 2016 at 8:59 AM, Rodrigo Duarte <rodrigodsousa at gmail.com>
wrote:

> After some investigation I've found the possible issue: the functional
> tests run in parallel, some of them create and delete roles and others use
> tokens to perform the creation/update/delete of other types of fixtures.
> The problem is that when we delete a role, we also revoke *all* tokens
> from a user that has any assignment containing that role - so? Race
> condition: if we are executing a not related operation and another test
> deletes a role, the user tokens will be revoked resulting in a request
> error.
>
> The strange part is that reverting this commit [1], the tests seem to work
> fine most of the times - what makes think that commit actually *fixes* a
> big issue in our revoke events (since before it, we would not revoke such
> types of tokens).
>
> I can see a couple of options:
> - Create brand new users and role_assignments to be responsible to handle
> operations in the fixtures for each test
>

This makes sense to me. There was a bug in the tests and this corrects it.


> - Change the "framework" of the tests and rely on tempest plugins
>
>
I don't know why this was suggested.

- Brant


> What to think? Makes sense?
>
> On Tue, Sep 20, 2016 at 11:03 AM, Steve Martinelli <s.martinelli at gmail.com
> > wrote:
>
>> Since September 14th the keystoneclient functional test job has been
>> broken. Let's be mindful of infra resources and stop rechecking the patches
>> there. Anyone have time to investigate this?
>>
>> See patches https://review.openstack.org/#/c/369469/ or
>> https://review.openstack.org/#/c/371324/
>>
>> ____________________________________________________________
>> ______________
>> OpenStack Development Mailing List (not for usage questions)
>> Unsubscribe: OpenStack-dev-request at lists.openstack.org?subject:unsubscrib
>> e
>> http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
>>
>>
>
>
> --
> Rodrigo Duarte Sousa
> Senior Quality Engineer @ Red Hat
> MSc in Computer Science
> http://rodrigods.com
>
> _____________
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstack.org/pipermail/openstack-dev/attachments/20160921/a71ad0b8/attachment.html>