canonical-ci-engineering team mailing list archive

[Kevin Gunn <kevin.gunn@xxxxxxxxxxxxx>]

Thanks Francis - I'll leave it up to Stephen how he wants it, advisory or
must-pass

On Thu, Sep 10, 2015 at 9:21 AM, Francis Ginther <
francis.ginther@xxxxxxxxxxxxx> wrote:

> Another way to iterate on these test failures is to manually run the new
> jobs on a bzr branch. Developers should have access to build these two
> jobs, specifying their branch as the "landing_candidate", then watching for
> the results (they will *not* be automatically posted to the MP).
>
> http://s-jenkins.ubuntu-ci:8080/job/mir-wily-i386-ci/
> http://s-jenkins.ubuntu-ci:8080/job/mir-mediumtests-wily-touch/
>
> Only the "landing_candidate" parameter needs to be supplied. All other
> parameters can be left at the default value.
>
> Francis
>
> On Wed, Sep 9, 2015 at 2:28 PM, Francis Ginther <
> francis.ginther@xxxxxxxxxxxxx> wrote:
>
>> Let me clarify that 1.5 hour time increase. That only applies if you want
>> to run these new builds as 'advisory' or not impacting the final result of
>> the build. The reason for the long time in this case is that jenkins has to
>> run the 'must-pass' jobs first and then the 'advisory' builds serially. In
>> my proposal above, the "wily armhf (any real phone hw)" and "wily i386"
>> jobs would be advisory. I only proposed this as a solution to addressing
>> the currently broken tests without blocking unrelated MPs that would fail
>> if the tests were just turned on as 'must-pass'.
>>
>> If all of these jobs are 'must-pass' then they can all be executed
>> concurrently and build times would be the same as they are today assuming
>> nothing else is building and nothing has to wait for resources.
>>
>> Francis
>>
>> On Tue, Sep 8, 2015 at 5:26 PM, Stephen M. Webb <
>> stephen.webb@xxxxxxxxxxxxx> wrote:
>>
>>> On 15-09-08 05:58 PM, Kevin Gunn wrote:
>>> > Whoa! 1.5 hrs longer.
>>> > Per MP that'd probably be too painful as wait times have been an issue
>>> in the past, especially for Mir (not sure what
>>> > we're down to, but i think it was ~30min at best).
>>> >
>>> > Wondering (wanting Cemil/Stephen to chime in) would it make sense to
>>> have all MP's landing on development-branch running
>>> > CI on
>>> > * vivid+overlay armhf (any real phone hw)
>>> > * wily amd64
>>> >
>>> > Then have promotions to release run on all the priority 1
>>> configurations
>>> > * vivid+overlay armhf (any real phone hw)
>>> > * wily amd64
>>> > * wily armhf (any real phone hw)
>>> > * wily i386
>>> >
>>> > in order to avoid sinking so much time into every MP
>>>
>>> I'd really rather see everything thrown at the MP CI so failures just
>>> don't make it as far as the archive.  Promotion to
>>> release is not the time to start finding problems, that's the entire
>>> point of CI at the MP level.
>>>
>>> Why would the builds be slowed down so much?  Is it just that the
>>> hardware is slow or is there a long delay waiting for
>>> resources in general?
>>>
>>> --
>>> Stephen M. Webb  <stephen.webb@xxxxxxxxxxxxx>
>>>
>>
>>
>>
>> --
>> Francis Ginther
>> Canonical - Ubuntu Engineering - Continuous Integration Team
>>
>
>
>
> --
> Francis Ginther
> Canonical - Ubuntu Engineering - Continuous Integration Team
>