[automated-testing] Supporting GD2 & GD1 in glusto-tests

[Jonathan Holloway]

Sounds good. My comments are inline.

On Wed, Aug 8, 2018 at 1:49 AM Nigel Babu <nigelb at redhat.com> wrote:

> I had a chat with Vijay and Akarsha yesterday. Here's a summary of what we
> decided.
>
> * We want to have two versions of the glustolibs-gluster library.
> * We will create a new repo called glusto-libs. Nigel will create this
> repo. To start off, we'll copy the content from glustolibs-gluster to this
> repo.
>

Both the GCS and non-GCS namespace packages will be located in the new
glusto-libs library, right?
Essentially going back to what we discussed way back about maintaining the
two separate repos for tests and libs.


> * Akarsha will land changes to this library to support GD2. Vijay and
> Akarsha will take responsibility for keeping them in sync in terms of
> features that apply in both places.
> * Given that moving the current version which supports 3.x to this repo
> will break a bunch of automation both upstream and downstream, we will
> defer it to the future. However, this is the end goal.
>

I think this answers my above question, but wanted to be sure.


> * Akarsha will look at marking tests that are relevant for GD2 in the
> existing tests so we can have a glusterfs + GD2 sanity test (before going
> to the container layer).
>

Will that be done via pytest markers and run via the py.test -m
<marker_name>?

Cheers,
Jonathan


>
> On Tue, Aug 7, 2018 at 8:22 PM Jonathan Holloway <jholloway at redhat.com>
> wrote:
>
>> Vijay,
>>
>> Is there a specific way you had in mind to implement the GCS version of
>> the glustolibs namespace package?
>> Do we maintain a separate branch for now and then when we have everything
>> in the branch working for both GCS and non-GCS swap it to master?
>> What will that mixed architecture look like? Config setting to enable GCS
>> or non-GCS and the logic in the libraries? Class-based?
>> Or are we maintaining both separately for the long-term?
>>
>> Cheers,
>> Jonathan
>>
>>
>> On Tue, Aug 7, 2018 at 1:51 AM Nigel Babu <nigelb at redhat.com> wrote:
>>
>>> Hello,
>>>
>>> Is there a POC that you can show yet? This is the second blocker to test
>>> GCS. The other one is a glusto[1] bug.
>>>
>>> [1]: https://github.com/loadtheaccumulator/glusto/issues/17
>>>
>>> On Thu, Jul 26, 2018 at 7:38 PM Nigel Babu <nigelb at redhat.com> wrote:
>>>
>>>> My primary concern here is that this may not work like you think it
>>>> will. We may want to test this idea for it's limitations and problems
>>>> before we decide.
>>>>
>>>> On Thu, Jul 26, 2018 at 7:35 PM Nigel Babu <nigelb at redhat.com> wrote:
>>>>
>>>>> Do you have a PoC that we can play with some time next week?
>>>>>
>>>>
>>>>
>>>
>>> --
>>> nigelb
>>>
>>
>
> --
> nigelb
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.gluster.org/pipermail/automated-testing/attachments/20180808/f5f08c77/attachment.html>