----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/14910/#review27479 -----------------------------------------------------------
Ship it! We should likely clarify that this is for recovering the "latest run" for completed executors? src/slave/slave.cpp <https://reviews.apache.org/r/14910/#comment53397> We should only attach directories for executors that exist in the slave's structs, otherwise the webui will not show the executor and hence doesn't need to have the path attached. So not sure if this is the right TODO or if the TODO should be to recover previous runs? - Ben Mahler On Oct. 24, 2013, 6:01 p.m., Vinod Kone wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > https://reviews.apache.org/r/14910/ > ----------------------------------------------------------- > > (Updated Oct. 24, 2013, 6:01 p.m.) > > > Review request for mesos, Benjamin Hindman and Ben Mahler. > > > Bugs: MESOS-737 > https://issues.apache.org/jira/browse/MESOS-737 > > > Repository: mesos-git > > > Description > ------- > > Now the slave recovers completed executors instead of skipping them. This > allows the webui to show them even after a restart. > > Also simplified the semantics of removeExecutor(). It no longer implicitly > does removeFramework(). > > > Diffs > ----- > > src/slave/slave.cpp debb2f4ce05fbfec450197e68bc8a0c78f1d0adf > > Diff: https://reviews.apache.org/r/14910/diff/ > > > Testing > ------- > > make check > > GLOG_v=1 ./bin/mesos-tests.sh --gtest_filter="*RecoverCompletedExecutor*" > --verbose > > > Thanks, > > Vinod Kone > >
