----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/62646/#review186966 -----------------------------------------------------------
Fix it, then Ship it! src/launcher/default_executor.cpp Lines 825-827 (patched) <https://reviews.apache.org/r/62646/#comment263912> One interesting case here is: what should be the taskState if there is a kill from the framework (`killing=true`), and there is a limitation error. Looks like the current code will assume that limitaiton error always take precedence. I think this behavior is fine, but let's add some comment about that semantics. src/launcher/default_executor.cpp Lines 850-853 (patched) <https://reviews.apache.org/r/62646/#comment263914> Looks like we should update `createStatusUpdate`, instead of doing this manually here? - Jie Yu On Sept. 28, 2017, 12:19 a.m., James Peach wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > https://reviews.apache.org/r/62646/ > ----------------------------------------------------------- > > (Updated Sept. 28, 2017, 12:19 a.m.) > > > Review request for mesos, Jie Yu and Qian Zhang. > > > Bugs: MESOS-7963 > https://issues.apache.org/jira/browse/MESOS-7963 > > > Repository: mesos > > > Description > ------- > > In the default executor, propagate newly added `WaitNestedContainer` > response fields to the corresponsing status update. This allows > frameworks to receive more reliably information about nested > containers failures. > > > Diffs > ----- > > src/launcher/default_executor.cpp 136c000917b2c4fdaf68fd460764c3a15fdf3bf0 > > > Diff: https://reviews.apache.org/r/62646/diff/1/ > > > Testing > ------- > > make check (Fedora 26) > > > Thanks, > > James Peach > >
