----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/20080/#review39749 -----------------------------------------------------------
src/slave/containerizer/mesos_containerizer.cpp <https://reviews.apache.org/r/20080/#comment72362> For posterity I'd like you to add a comment here that explains why it's OK to return a Failure. In particular, we should be asking ourselves the question: how does the slave know that it should be sending TASK_FAILED or TASK_LOST for the tasks that would have been launched? In this case any subsequent 'wait' will return a failed future as well which the slave knows to treat as a "terminated" container. This makes me think that 'wait' should have really returned an Option<Termination> to distinguish the case when it doesn't know about a container ID! But that's for another review. - Benjamin Hindman On April 7, 2014, 1:19 p.m., Till Toenshoff wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > https://reviews.apache.org/r/20080/ > ----------------------------------------------------------- > > (Updated April 7, 2014, 1:19 p.m.) > > > Review request for mesos, Benjamin Hindman, Ian Downes, Niklas Nielsen, and > Vinod Kone. > > > Repository: mesos-git > > > Description > ------- > > Introduces the ContainerInfo protobuf as part of CommandInfo. > Right now, if present, the mesos containerizer fails the task launch to point > out that we do not support it on that containerizer. > > This will be needed for the ExternalContainerizer and possibly other > containerizers as well. > > > Diffs > ----- > > include/mesos/mesos.proto 37f8a7f > src/slave/containerizer/mesos_containerizer.cpp c819c97 > > Diff: https://reviews.apache.org/r/20080/diff/ > > > Testing > ------- > > make check > > > Thanks, > > Till Toenshoff > >
