----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/53378/#review154710 -----------------------------------------------------------
Fix it, then Ship it! src/launcher/executor.cpp (lines 305 - 306) <https://reviews.apache.org/r/53378/#comment224377> Not a native speaker, but I'd say passive voice would have been more appropriate here: the task has *been* transitioned. Also, though not documented anywhere, we try to format short comments in a way that the right edge is less jagged. Some people find it easier to read, while others don't care, hence a nice change : ) ``` // This check prevents us from sending `TASK_RUNNING` update // after the task has been transitioned to `TASK_KILLING`. ``` - Alexander Rukletsov On Nov. 2, 2016, 9:51 p.m., Gastón Kleiman wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > https://reviews.apache.org/r/53378/ > ----------------------------------------------------------- > > (Updated Nov. 2, 2016, 9:51 p.m.) > > > Review request for mesos, Alexander Rukletsov and Benjamin Mahler. > > > Bugs: MESOS-6457 > https://issues.apache.org/jira/browse/MESOS-6457 > > > Repository: mesos > > > Description > ------- > > The command executor would send health updates even after having > transitioned a task to `TASK_KILLING`. This plus the fact that health > updates always have the `TASK_RUNNING` state leads to `TASK_KILLING -> > TASK_RUNNING` transitions. > > This patch makes the executor skip health updates after starting to kill > a task. > > > Diffs > ----- > > src/launcher/executor.cpp 0544121e679db503fe4eaf23a24e315eb188a520 > > Diff: https://reviews.apache.org/r/53378/diff/ > > > Testing > ------- > > Test in the next patch in the chain. > > > Thanks, > > Gastón Kleiman > >
