----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/57364/#review168131 -----------------------------------------------------------
Offhand, I would think the right logic should be: ``` // Set up expectation for `registered` event driver.start(); Clock::advance(authentication_backoff_factor); Clock::advance(registration_backoff_factor); AWAIT_READY(registered); process::Time registerTime = Clock::now(); ``` i.e., settling clock should be unnecessary if we wait for `registered` before calling `Clock::now()`, and we still want to advance the clock before waiting for `registered`. Does that make sense to you? src/tests/fault_tolerance_tests.cpp Lines 826 (patched) <https://reviews.apache.org/r/57364/#comment240268> With the clock paused, I believe we need to wait for `authentication_backoff_factor` and `registration_backoff_factor` to ensure that registration will occur. - Neil Conway On March 7, 2017, 6:48 a.m., Jay Guo wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > https://reviews.apache.org/r/57364/ > ----------------------------------------------------------- > > (Updated March 7, 2017, 6:48 a.m.) > > > Review request for mesos and Neil Conway. > > > Bugs: MESOS-7029 > https://issues.apache.org/jira/browse/MESOS-7029 > > > Repository: mesos > > > Description > ------- > > Fixed flaky test FaultToleranceTest.FrameworkReregister. > > > Diffs > ----- > > src/tests/fault_tolerance_tests.cpp > b13a7e2527189931b733fb4f188b1463fe1f919a > > > Diff: https://reviews.apache.org/r/57364/diff/1/ > > > Testing > ------- > > ./bin/mesos-tests.sh --gtest_filter="FaultToleranceTest.FrameworkReregister" > --gtest_break_on_failure --gtest_repeat=-1 > > > Thanks, > > Jay Guo > >
