----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/35935/#review89583 -----------------------------------------------------------
src/tests/reconciliation_tests.cpp (lines 905 - 908) <https://reviews.apache.org/r/35935/#comment142218> I don't follow why you couldn't have just used the master in this test? What's the hard part? AFAICT, you can reconcile a bunch of unlaunched tasks which would stress test the createStatusUpdate() part? - Vinod Kone On June 26, 2015, 9:33 p.m., Ben Mahler wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > https://reviews.apache.org/r/35935/ > ----------------------------------------------------------- > > (Updated June 26, 2015, 9:33 p.m.) > > > Review request for mesos, Jie Yu and Vinod Kone. > > > Bugs: MESOS-2941 > https://issues.apache.org/jira/browse/MESOS-2941 > > > Repository: mesos > > > Description > ------- > > We currently don't have any nice testing abstractions for scaling up a master > in-memory, so this simulates the reconciliation logic instead. Not ideal, but > was very valuable for improving the performance. > > > Diffs > ----- > > src/tests/reconciliation_tests.cpp 6042d8c02d86f486e0c4d82d5a70666d7ac9019b > > Diff: https://reviews.apache.org/r/35935/diff/ > > > Testing > ------- > > make check + ran the benchmark (per MESOS-2940) > > > Thanks, > > Ben Mahler > >
