[automated-testing] gluster-block: Discussion regarding glusto tests

Pranith Kumar Karampuri pkarampu at redhat.com
Mon Jul 2 12:51:58 UTC 2018


On Mon, Jul 2, 2018 at 5:50 PM, Nigel Babu <nigelb at redhat.com> wrote:

> I've already tried to redirect this thread once. Please use the upstream
> automated-testing@ list rather than CCing half the world. This is a
> discussion that needs to happen and be decided in the open.
>
> Pranith/Prasanna: At the moment your problem is that you want to write
> tests in conjunction with your pull request and splitting into multiple
> repos do not give you this ability. Is that the compelling reason why you
> want this to happen in the gluster-block repo directly rather than
> glusto-tests?
>

Yes. Goal is to be able to test HEAD with tests at HEAD, write tests once
and be done. It would be nice if we can come up with a solution that
doesn't lead to duplication of efforts from contributors to these two
different projects like it is happening with GlusterFS now. How can we make
that possible?


>
>
> On Mon, Jul 2, 2018 at 5:40 PM Pranith Kumar Karampuri <
> pkarampu at redhat.com> wrote:
>
>>
>>
>> On Mon, 2 Jul 2018, 4:02 pm Rahul Hinduja, <rhinduja at redhat.com> wrote:
>>
>>>
>>>
>>> On Mon, Jul 2, 2018 at 3:54 PM, Prasanna Kalever <pkalever at redhat.com>
>>> wrote:
>>>
>>>> On Mon, Jul 2, 2018 at 3:41 PM, Vijay Bhaskar Reddy Avuthu
>>>> <vavuthu at redhat.com> wrote:
>>>> > + Rahul
>>>> >
>>>> > Currently all the components of Gluster are under "gluster-tests"
>>>> repo, I
>>>> > would recommend the same for block. Another reason is its easy for QE
>>>> to
>>>> > track and trigger runs whenever is needed.
>>>>
>>>> +Vijay Bellur, +Amar
>>>>
>>>> The block dev is keeping the continuous effort here to catch the
>>>> regression that might be introduced at each PR level, it will be easy
>>>> for the contributes to run the tests locally, if they are part of the
>>>> same repository. Downloading a different repository just to test the
>>>> code that goes in guster-block repository doesn't make any sense to
>>>> me. I'm not aware of any opensource project which would like their
>>>> tests to maintained as part of a separate repository. At-least I
>>>> wouldn't encourage this with gluster-block.
>>>>
>>>
>>> Thats not true, if we go with this path than whats blocking us from
>>> having multiple repositories like "ganesha", "geo-rep", "smb", "afr" etc.
>>>
>> Even though currently gluster dev worked on gluster block libs and cases
>>> but we worked as a team in sprint-2 to achieve this. QE has contributed
>>> maximum across glusto and will continue to invest time and resources in
>>> glusto project [1] including block.
>>>
>>
>> Our work of integration with glusto which started even before stop-gap
>> work was to enable development team to write test cases along with the
>> feature/bug-fix and send them in same PR. This way, the behaviour of the
>> project at that HEAD will be tested correctly at that HEAD. If
>> functionality changes then tests will change automatically.
>> Asking contributors to send two separate patches to two different repos
>> to get a change in the long run leads either to stale tests or stale code.
>> It is a head ache better to avoid for both the teams.
>>
>>
>>> All the cases irrespective of features must go to a common glusto-tests
>>> repository. This would also help in central-CI and polarion reporting.
>>>
>>
>> This is definitely something we should get right. Otherwise it will again
>> become the same story as glusterfs where ongoing development and QE will
>> have to duplicate efforts in automating the test cases.
>>
>> If we get this right by the time QE cycle starts there would be good
>> amount of work already done by development team for automation and only the
>> missing parts need to be automated by QE.
>>
>> Is there a way to get central-CI/polarion reporting to pull in tests from
>> another repo and run them?
>>
>>
>>> [1]: https://review.gluster.org/#/q/project:glusto-tests
>>>
>>
>>
>>
>>
>>>
>>>
>>>>
>>>> Thanks,
>>>> --
>>>> Prasanna
>>>>
>>>>
>>>> >
>>>> > Regards,
>>>> > Vijay A
>>>> >
>>>> > On Mon, Jul 2, 2018 at 2:05 PM, Prasanna Kalever <pkalever at redhat.com
>>>> >
>>>> > wrote:
>>>> >>
>>>> >> This was the plan long back, which Shwetha and myself had discussed,
>>>> >> before even start work on the glusto tests for gluster-block, let me
>>>> >> try to put the summary here:
>>>> >>
>>>> >> Plan 1:
>>>> >> Create a 'glusto/libs/' folder under gluster-block repo and start
>>>> >> adding libraries there and then keep the tests at 'glusto/tests/'
>>>> >> folder under gluster-block repo.
>>>> >>
>>>> >> Plan 2:
>>>> >> Push the libraries to glusto-tests repo and maintain the tests under
>>>> >> gluster-block repo.
>>>> >>
>>>> >> I'm okay with any of the above approaches.
>>>> >>
>>>> >> Thanks!
>>>> >> --
>>>> >> Prasanna
>>>> >>
>>>> >>
>>>> >> On Mon, Jul 2, 2018 at 11:46 AM, Bhumika Goyal <bgoyal at redhat.com>
>>>> wrote:
>>>> >> > Hi all,
>>>> >> >
>>>> >> > During stop-gap, we wrote some libraries(patch on gerrit) and test
>>>> >> > cases(PR on gluster-block repo currently) for gluster-block using
>>>> glusto.
>>>> >> > Some of us think that having the glusto based test cases in the
>>>> >> > gluster-block repo instead of glusto-tests repo would be easier for
>>>> >> > developers to test their changes locally. In this way, we don't
>>>> have to deal
>>>> >> > with two repos for testing the changes every time.
>>>> >> > What are your suggestions and is it okay to go with this approach?
>>>> >> >
>>>> >> > Thanks,
>>>> >> > Bhumika
>>>> >
>>>> >
>>>>
>>>
>>>
>
> --
> nigelb
>



-- 
Pranith
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.gluster.org/pipermail/automated-testing/attachments/20180702/368896e8/attachment-0001.html>


More information about the automated-testing mailing list