Not sure if this is a common practice. But can't we trigger via travis-ci directly? You can go here <https://travis-ci.org/github/apache/hudi/pull_requests> or here <https://travis-ci.org/github/apache/hudi/builds> and there you can find an option to restart the build (right most column in every row) again if need be. Wouldn't this suffice?
On Wed, May 27, 2020 at 5:50 AM vino yang <[email protected]> wrote: > Hi Lamber-Ken, > > Thanks for opening this discussion. > > +1 to fix this issue. > > About the solution, can we consider to introduce a "CI Bot" just like > Apache Flink has done?[1] > > Just a thought. > > Best, > Vino > > [1]: https://github.com/flink-ci/ci-bot/ > > Lamber Ken <[email protected]> 于2020年5月27日周三 下午2:08写道： > > > Dear community, > > > > Use case: A build fails due to an externality. The source is actually > > correct. It would build OK and pass if simply re-run. Is there some way > to > > nudge Travis-CI to do another build, other than pushing a "dummy" commit? > > > > The way I often used is `git commit --allow-empty -m 'trigger rebuild'`, > > push a dummy commit, the travis will rebuild. Also noticed some apache > > projects have supported this feature. > > > > For example: > > 1. Carbondata use "retest this please" > > https://github.com/apache/carbondata/pull/3387 > > > > 2. Bookkeeper use "run pr validation" > > https://github.com/apache/bookkeeper/pull/2158 > > > > But, I can't find a effective solution from Github and Travis's > > documentation[1], any thoughts or opinions? > > > > Best, > > Lamber-Ken > > > > [1] https://docs.travis-ci.com https://support.github.com > > > -- Regards, -Sivabalan
