----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/36620/#review92457 -----------------------------------------------------------
src/linux/cgroups.cpp (line 1504) <https://reviews.apache.org/r/36620/#comment146628> This is actually the Freezer right? And we should make this VLOG(1) src/linux/cgroups.cpp (line 1627) <https://reviews.apache.org/r/36620/#comment146629> To me this doesn't killer doesn't "atomically" kill all of them, as that implies either all or none. I think the TaskKiller actually best effort try to kill by signal and reintrospect all tasks in a cgroup within the timeout src/linux/cgroups.cpp (line 1650) <https://reviews.apache.org/r/36620/#comment146630> 100 ms seems way too fast? And why not use the default timeout that's already there? src/linux/cgroups.cpp (line 1668) <https://reviews.apache.org/r/36620/#comment146632> This should go to line 1674? src/linux/cgroups.cpp (line 1676) <https://reviews.apache.org/r/36620/#comment146631> Move to previous line. src/linux/cgroups.cpp (line 1684) <https://reviews.apache.org/r/36620/#comment146633> This fits on to previous line? - Timothy Chen On July 21, 2015, 5:18 p.m., Joerg Schad wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > https://reviews.apache.org/r/36620/ > ----------------------------------------------------------- > > (Updated July 21, 2015, 5:18 p.m.) > > > Review request for mesos and Timothy Chen. > > > Bugs: MESOS-3086 > https://issues.apache.org/jira/browse/MESOS-3086 > > > Repository: mesos > > > Description > ------- > > WIP Added Non-Freezeer Task Killer. > > > Diffs > ----- > > src/linux/cgroups.cpp b7d11ac4eb8b78664c8c7bd7a5b0d566b5bf2f0e > > Diff: https://reviews.apache.org/r/36620/diff/ > > > Testing > ------- > > sudo make check > > > Thanks, > > Joerg Schad > >