----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/52864/#review153580 -----------------------------------------------------------
Fix it, then Ship it! src/health-check/health_checker.cpp (line 215) <https://reviews.apache.org/r/52864/#comment222984> Seems like you might want a small comment here about how we assume this is a local send? i.e. the health checker library is not used in a binary external to the executor that exits after the promise is failed? If we were to also update send to become satisfying a promise, then we wouldn't need this comment. - Benjamin Mahler On Oct. 14, 2016, 12:35 p.m., Alexander Rukletsov wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > https://reviews.apache.org/r/52864/ > ----------------------------------------------------------- > > (Updated Oct. 14, 2016, 12:35 p.m.) > > > Review request for mesos, Gastón Kleiman and haosdent huang. > > > Repository: mesos > > > Description > ------- > > After we have refactored the HealthChecker into a library from the > binary, there is no more need to wait before failing the promise to > ensure the message has been sent to the executor. HealthChecker > lifetime is managed by an executor, hence it is their responsibility > not to clean the instance until after the message with the `kill_task` > flag is received. > > > Diffs > ----- > > src/health-check/health_checker.cpp > 96ae1a733ff3d211b84d0893b4603873af1c89f0 > > Diff: https://reviews.apache.org/r/52864/diff/ > > > Testing > ------- > > See https://reviews.apache.org/r/52873/. > > > Thanks, > > Alexander Rukletsov > >