----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/13446/#review24942 -----------------------------------------------------------
Ship it! It does feel slightly odd that the statusUpdate() acts as both a message handler for StatusUpdateMessage, and as a helper for sending status updates generated by the master. The code appears correct but it seems odd that we simulate an incoming message in order to send a message from the master itself. So I then thought we should set the pid to the Master's pid, but there are also some implicit semantics regarding the StatusUpdateMessage.pid, where if it is not present, the status update does not require an acknowledgement. Perhaps we should document that inside messages.proto? - Ben Mahler On Aug. 9, 2013, 6:45 p.m., Vinod Kone wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > https://reviews.apache.org/r/13446/ > ----------------------------------------------------------- > > (Updated Aug. 9, 2013, 6:45 p.m.) > > > Review request for mesos, Benjamin Hindman and Ben Mahler. > > > Bugs: MESOS-635 > https://issues.apache.org/jira/browse/MESOS-635 > > > Repository: mesos-git > > > Description > ------- > > Generated TASK_LOST updates for valid tasks are now sent via statusUpdate() > instead of directly sending them. This ensures we don't miss critical stuff > that needs to be done when handling terminal updates. > > > Diffs > ----- > > src/master/master.cpp a2f8929987f95e796aba71063223aea7003bf3df > > Diff: https://reviews.apache.org/r/13446/diff/ > > > Testing > ------- > > make check > > > Thanks, > > Vinod Kone > >
