> On Feb. 15, 2017, 10 p.m., Vinod Kone wrote: > > src/launcher/default_executor.cpp > > Lines 869 (patched) > > <https://reviews.apache.org/r/56449/diff/2/?file=1627710#file1627710line875> > > > > What's the guarantee that this task is still present in `containers`? > > AFAICT this is a deferred callback invoked by the health checker.
You're right: we erase the taskId right after we reap the task container, while a check can be in flight and report its status afterwards. - Alexander ----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/56449/#review165770 ----------------------------------------------------------- On Feb. 28, 2017, 2:35 p.m., Alexander Rukletsov wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > https://reviews.apache.org/r/56449/ > ----------------------------------------------------------- > > (Updated Feb. 28, 2017, 2:35 p.m.) > > > Review request for mesos, Anand Mazumdar, Gastón Kleiman, and Vinod Kone. > > > Repository: mesos > > > Description > ------- > > With the recent introduction of the `Container` struct in the default > executor, tasks' health checkers should be moved to this struct. Also, > health checking is stopped on per-task basis and not on shutdown. > > > Diffs > ----- > > src/launcher/default_executor.cpp e63cf153831088851863d0956455a024e9bc172a > > > Diff: https://reviews.apache.org/r/56449/diff/2/ > > > Testing > ------- > > See https://reviews.apache.org/r/56218/ > > > Thanks, > > Alexander Rukletsov > >
