----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/36814/#review93995 -----------------------------------------------------------
Ship it! LGTM. I like the idea of not setting the executorId=taskId in the actual TaskInfo/Task, since that could confuse other logic downstream that expects the executor/executorId to be empty for command executors. However, since this is exposed in state.json, there might be external tools/scripts that expect the same. Now they can just check if executorId==taskId, assuming there are no custom executors using the same naming scheme. Could you please add a note to docs/upgrades.md mentioning the API change, in case anybody does depend on the old behavior? Does this actually solve the problem of: "The webui is broken for command executors because the master does not know the executor ID for the tasks using a command executor."? @bmahler do you have any further thoughts/questions on this simple change? - Adam B On Aug. 2, 2015, 12:57 a.m., haosdent huang wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > https://reviews.apache.org/r/36814/ > ----------------------------------------------------------- > > (Updated Aug. 2, 2015, 12:57 a.m.) > > > Review request for mesos, Adam B and Ben Mahler. > > > Bugs: MESOS-527 > https://issues.apache.org/jira/browse/MESOS-527 > > > Repository: mesos > > > Description > ------- > > Fill executor_id in state.json when task is run in CommandExecutor. > > > Diffs > ----- > > src/common/http.cpp a74c51d9392d0b0a67d51a0143eb314cfca11245 > src/tests/common/http_tests.cpp 38d062b2b4062e0a2fc912bad0cc2d73339eb66e > > Diff: https://reviews.apache.org/r/36814/diff/ > > > Testing > ------- > > > Thanks, > > haosdent huang > >