----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/61573/#review182888 -----------------------------------------------------------
src/launcher/default_executor.cpp Line 233 (original), 233 (patched) <https://reviews.apache.org/r/61573/#comment258825> shouldn't we also pass the `event.kill().kill_policy()`? It allows frameworks/operators to override the kill policy, see: https://github.com/apache/mesos/blob/628d6609b6eeb90767e5799d0177bfe4828d71aa/include/mesos/executor/executor.proto#L96-L106 - Gastón Kleiman On Aug. 14, 2017, 5:19 p.m., Anand Mazumdar wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > https://reviews.apache.org/r/61573/ > ----------------------------------------------------------- > > (Updated Aug. 14, 2017, 5:19 p.m.) > > > Review request for mesos and Vinod Kone. > > > Bugs: MESOS-7879 > https://issues.apache.org/jira/browse/MESOS-7879 > > > Repository: mesos > > > Description > ------- > > This modifies the default executor to perform signal escalation > via the 'KILL_NESTED_CONTAINER' call i.e., send a SIGTERM followed > by a SIGKILL after some grace period. Note that support for kill > policies still needs to be done instead of using a constant grace > period. > > > Diffs > ----- > > src/launcher/default_executor.cpp c25cc941eefc0cca998a99d76497bfdd05babe92 > > > Diff: https://reviews.apache.org/r/61573/diff/2/ > > > Testing > ------- > > make check > > > Thanks, > > Anand Mazumdar > >
