geode-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Owen Nichols <onich...@pivotal.io>
Subject Re: [DISCUSS] is it time to make Windows tests gating?
Date Fri, 17 May 2019 23:56:49 GMT
Windows tests are now in the main pipeline.  There are still a handful of intermittent failures,
but generally I expect to see them mostly green.  The increased visibility should help drive
the process of shaking out any remaining issues.

The shape of the main pipeline is now very tall, which most people don’t seem to mind, but
if you feel strongly that the font is just too small, a demo of an alternative layout is at
https://github.com/apache/geode/pull/3606

> On May 17, 2019, at 9:53 AM, Dan Smith <dsmith@pivotal.io> wrote:
> 
> You're awesome Jens!
> 
> -Dan
> 
> On Fri, May 17, 2019 at 9:48 AM Jens Deppe <jdeppe@pivotal.io> wrote:
> 
>> I fixed the first two and started ignoring
>> AvailablePortHelperIntegrationTest for Windows (Windows has a default
>> active port range wildly different than Linux which makes this test very
>> flaky, and since it's just testing test code it seemed reasonable to ignore
>> it).
>> 
>> LauncherMemberMXBeanIntegration test still needs some love and I can look
>> at that since I've worked on it in the past.
>> 
>> --Jens
>> 
>> On Fri, May 17, 2019 at 9:33 AM Dan Smith <dsmith@pivotal.io> wrote:
>> 
>>> Looking at the metrics for the windows jobs, it looks like the windows
>>> tests are mostly red due to a few specific tests. The acceptance and gfsh
>>> distributed tests jobs seem to be ok, it's just the unit tests and
>>> integration tests that have problems. It also looks like
>>> ExportConfigCommandTest stopped failing recently. Anyone want to take the
>>> credit?
>>> 
>>> I'd be in favor of putting them in the main pipeline but not the PR
>>> pipeline, since they take so long. Based on the metrics it looks like we
>>> are pretty close to green with these jobs in terms of effort. Anyone want
>>> to volunteer to fix the remaining failures (or see if they have been
>> fixed
>>> recently)?
>>> 
>>> -Dan
>>> 
>>> 
>>> 
>> https://concourse.apachegeode-ci.info/teams/main/pipelines/apache-develop-metrics/jobs/GeodeWindowsUnitTestOpenJDK8Metrics/builds/209
>>> 
>>> CreateDiskStoreCommandTest:  2 failures
>>> ExportConfigCommandTest:  34 failures
>>> 
>>> 
>>> 
>> https://concourse.apachegeode-ci.info/teams/main/pipelines/apache-develop-metrics/jobs/GeodeWindowsIntegrationTestOpenJDK8Metrics/builds/209
>>> 
>>> LauncherMemberMXBeanIntegrationTest:  4 failures
>>> AvailablePortHelperIntegrationTest:  17 failures
>>> 
>>> 
>>> 
>>> 
>>> On Thu, May 16, 2019 at 5:32 PM Owen Nichols <onichols@pivotal.io>
>> wrote:
>>> 
>>>> I’ve created a PR for this: https://github.com/apache/geode/pull/3597
>>>> 
>>>>> On May 16, 2019, at 3:06 PM, Blake Bender <bbender@pivotal.io>
>> wrote:
>>>>> 
>>>>> +1 this needs to happen.  I hope that doesn't cause too much pain for
>>> the
>>>>> dev team, but the native client team has a hard requirement that all
>>> our
>>>>> stuff works properly on Windows at all times, and it can cause
>> trouble
>>> if
>>>>> random builds of the server can break us on Windows.
>>>>> 
>>>>> I would hesitate to run these per-commit if they're taking that long,
>>> but
>>>>> daily is a thing that can easily happen.
>>>>> 
>>>>> On Thu, May 16, 2019 at 2:23 PM Bruce Schuchardt <
>>> bschuchardt@pivotal.io
>>>>> 
>>>>> wrote:
>>>>> 
>>>>>> big +1, as long as artifacts of failed runs can be downloaded
>>>>>> 
>>>>>> On 5/15/19 6:28 PM, Owen Nichols wrote:
>>>>>>> For a very long time we’ve had Windows tests in the main pipeline
>>>>>> (hidden away, not in the default view), but the pipeline proceeds
to
>>>>>> publish regardless of whether Windows tests fail or even run at all.
>>>>>>> 
>>>>>>> Now seems like a good time to review whether to:
>>>>>>> a) treat Windows tests as first-class tests and prevent the
>> pipeline
>>>>>> from proceeding if any test fails on Windows
>>>>>>> b) keep as-is
>>>>>>> c) change Windows tests to trigger only once a week rather than
on
>>>> every
>>>>>> commit, if they are going to remain "informational only"
>>>>>>> 
>>>>>>> One disadvantage to making Windows tests gating is that they
>>> currently
>>>>>> take much longer to run (around 5 hours, vs 2 hours for Linux
>> tests).
>>>>>> 
>>>> 
>>>> 
>>> 
>> 


Mime
View raw message