> On Jan. 2, 2015, 10:23 p.m., Ben Mahler wrote: > > What happened previously if a launch takes forever? > > > > What happens now if a launch takes forever?
Ben that's a good point, previously the registration timeout gates the launch not to take forever, however it is too coarse of a timeout as it included the time to launch and also the time it take for the executor to register after launch. Now it will only time out on the latter case, I think we do need a seperate timeout that checks for the launch itself. Ben what do you think? - Timothy ----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/29437/#review66557 ----------------------------------------------------------- On Dec. 31, 2014, 11:57 p.m., Nishant Suneja wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > https://reviews.apache.org/r/29437/ > ----------------------------------------------------------- > > (Updated Dec. 31, 2014, 11:57 p.m.) > > > Review request for mesos and Timothy Chen. > > > Bugs: MESOS-999 > https://issues.apache.org/jira/browse/MESOS-999 > > > Repository: mesos-git > > > Description > ------- > > As part of this bug fix, I have trigerred the executor registration timeout > timer after the container's future object is set, instead of starting the > timer when the container launch is still pending > > > Diffs > ----- > > src/slave/slave.cpp 50b57819b55bdcdb9f49f20648199badc4d3f37b > src/tests/composing_containerizer_tests.cpp > 5ab5a36cadb7f8622bad0c5814e9a5fb338753ad > src/tests/containerizer.hpp 24b014f44d9eec56840e18cf39fbf9100f2c0711 > src/tests/slave_tests.cpp f2896a1fc4521452e29fd261a6f117372345dcfc > > Diff: https://reviews.apache.org/r/29437/diff/ > > > Testing > ------- > > Added the unit test : SlaveTest::ExecutorRegistrationTimeoutTrigger > make check succeeds. > > > Thanks, > > Nishant Suneja > >
