----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/50103/#review142547 -----------------------------------------------------------
Ship it! This is the downside of leveraging task status updates for the health check information: we should keep the last sent status update in order not to accidentally omit fields. Let's think about it when we work on health checks. - Alexander Rukletsov On July 16, 2016, 3:31 a.m., haosdent huang wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > https://reviews.apache.org/r/50103/ > ----------------------------------------------------------- > > (Updated July 16, 2016, 3:31 a.m.) > > > Review request for mesos, Alexander Rukletsov and Benjamin Mahler. > > > Bugs: MESOS-5294 > https://issues.apache.org/jira/browse/MESOS-5294 > > > Repository: mesos > > > Description > ------- > > When we enable health check for the tasks run in docker container, > their `TaskStatus` messages generated by docker executor may miss > `NetworkInfo` field and agent would fill the host ip as default value. > In this changes, we cache the `NetworkInfo` of the task and reuse it > when generate `TaskStatus` messages which state is `TASK_RUNNING`. > > > Diffs > ----- > > src/docker/executor.cpp 69511044e39bc05d7d6240264ec70b6e6f44edba > > Diff: https://reviews.apache.org/r/50103/diff/ > > > Testing > ------- > > > Thanks, > > haosdent huang > >
