Hi Sheng, I agree with doubling-down on the efforts to fix the flaky tests but do not agree with compromising the stability of the test automation.
As a compromise, we could probably run the flaky tests as part of the nightly test automation -- would that work? I like your suggestion of using this: https://pypi.python.org/pypi/flaky in another email thread. May be we could have a higher rerun count as part of the nightly test to have better test automation stability. Bhavin Thaker. On Sun, Jan 14, 2018 at 12:21 PM, Sheng Zha <[email protected]> wrote: > Hi Bhavin, > > Thanks for sharing your thoughts. Regarding the usage of 'flaky' plugin > for retrying flaky tests, it's proposed as a compromise, given that it will > take time to properly fix the tests and we still need coverage in the > meantime. > > I'm not sure if releasing before these tests are re-enabled should be the > way, as it's not a good practice to release features that are not covered > by tests. Having done it before doesn't make it right. In that sense, > release efforts shouldn't be a blocker for re-enabling tests. Rather, it > should be the other way around, and release should happen only after we > recover the lost test coverage. > > I hope that we would do the right thing for our users. Thanks. > > -sz > > On 2018-01-14 11:00, Bhavin Thaker <[email protected]> wrote: > > Hi Sheng, > > > > Thank you for your efforts and this proposal to improve the tests. Here > are > > my thoughts. > > > > Shouldn’t the focus be to _engineer_ each test to be reliable instead of > > compromising and discussing the relative tradeoffs in re-enabling flaky > > tests? Is the test failure probability really 10%? > > > > As you correctly mention, the experiences in making the tests reliable > will > > then serve as the standard for adding new tests rather than continuing to > > chase the elusive goal of reliable tests. > > > > Hence, my non-binding vote is: > > -1 for proposal #1 for renabling flaky tests. > > +1 for proposal #2 for setting the standard for adding reliable tests. > > > > I suggest to NOT compromise on the quality and reliability of the tests, > > similar to the high bar maintained for the MXNet source code. > > > > If the final vote is to re-enable flaky tests, then I propose that we > > enable them immediately AFTER the next MXNet release instead of doing it > > during the upcoming release. > > > > Bhavin Thaker. > > > > On Sat, Jan 13, 2018 at 2:20 PM, Marco de Abreu < > > [email protected]> wrote: > > > > > Hello Sheng, > > > > > > thanks a lot for leading this task! > > > > > > +1 for both points. Additionally, I'd propose to add the requirement to > > > specify a reason if a new test takes more than X seconds (say 10) or > adds > > > an external dependency. > > > > > > Looking forward to getting these tests fixed :) > > > > > > Best regards, > > > Marco > > > > > > On Sat, Jan 13, 2018 at 11:14 PM, Sheng Zha <[email protected]> > wrote: > > > > > > > Hi MXNet community, > > > > > > > > Thanks to the efforts of several community members, we identified > many > > > > flaky tests. These tests are currently disabled to ensure the smooth > > > > execution of continuous integration (CI). As a result, we lost > coverage > > > on > > > > those features. They need fixing and to be re-enabled to ensure the > > > quality > > > > of our releases. I'd like to propose the following: > > > > > > > > 1, Re-enable flaky python tests with retries if feasible > > > > Although the tests are unstable, they would still be able to catch > > > breaking > > > > changes. For example, suppose a test fails randomly with 10% > probability, > > > > the probability of three failed retries become 0.1%. On the other > hand, a > > > > breaking change would result in 100% failure. Although this could > > > increase > > > > the testing time, it's a compromise that can help avoid bigger > problem. > > > > > > > > 2, Set standard for new tests > > > > I think having criteria that new tests should follow can help > improve the > > > > quality of tests, but also the quality of code. I propose the > following > > > > standard for tests. > > > > - Reliably passing with good coverage > > > > - Avoid randomness unless necessary > > > > - Avoid external dependency unless necessary (e.g. due to license) > > > > - Not resource-intensive unless necessary (e.g. scaling tests) > > > > > > > > In addition, I'd like to call for volunteers on helping with the fix > of > > > > tests. New members are especially welcome, as it's a good > opportunity to > > > > familiarize with MXNet. Also, I'd like to request that members who > wrote > > > > the feature/test could help either by fixing, or by helping others > > > > understand the issues. > > > > > > > > The effort on fixing the tests is tracked at: > > > > https://github.com/apache/incubator-mxnet/issues/9412 > > > > > > > > Best regards, > > > > Sheng > > > > > > > > > >
