----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/56268/#review164178 -----------------------------------------------------------
src/launcher/default_executor.cpp (lines 674 - 677) <https://reviews.apache.org/r/56268/#comment235795> not new in this patch chain, but: do we care about non-zero exit code if the task was killed by the framework? the `taskState == TASK_KILLED` suggests that we don't (which is reasonable) but the code above in #637 ignores kill intent if exit status is zero ? src/launcher/default_executor.cpp (lines 720 - 729) <https://reviews.apache.org/r/56268/#comment235800> I would rename s/killing/kill/ and s/killed/killing/ see below. src/launcher/default_executor.cpp (line 820) <https://reviews.apache.org/r/56268/#comment235798> s/it has already been killed/it is in the process of getting killed/ src/launcher/default_executor.cpp (line 1022) <https://reviews.apache.org/r/56268/#comment235797> s/killed/killing/ sounds more appropriate. - Vinod Kone On Feb. 3, 2017, 4:45 a.m., Anand Mazumdar wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > https://reviews.apache.org/r/56268/ > ----------------------------------------------------------- > > (Updated Feb. 3, 2017, 4:45 a.m.) > > > Review request for mesos and Vinod Kone. > > > Bugs: MESOS-6296 > https://issues.apache.org/jira/browse/MESOS-6296 > > > Repository: mesos > > > Description > ------- > > This ensures that when we add support for the default executor > launch multiple task groups, kill for a particular task should > _only_ result in the killing of a task group and not the entire > executor as was the case currently. > > > Diffs > ----- > > src/launcher/default_executor.cpp 97eee05cac8cb1f62d43e2aecc08a8e54e49eac3 > src/tests/default_executor_tests.cpp > e8e0aa2f0d2508de6db03685c70d1ed9c1da3659 > > Diff: https://reviews.apache.org/r/56268/diff/ > > > Testing > ------- > > make check > > > Thanks, > > Anand Mazumdar > >
