> On July 30, 2014, 5:12 p.m., Ben Mahler wrote: > > src/tests/zookeeper_tests.cpp, lines 202-205 > > <https://reviews.apache.org/r/24123/diff/1/?file=646315#file646315line202> > > > > What's stopping it from being even lower? > > > > Say, 100 milliseconds?
Ok so with 100 milliseconds I have seen the test fail when run with a large number of iterations due to loss connection to the local ZooKeeperTestServer when the server's up. I haven't found the root cause but 100 milliseconds is way lower than usually zookeeper setup and it may mess with zookeeper's tickTime. And the way we're using the test server is a bit hacky anyway. I think the wait for ZooKeeperTestServer to expire the session is not an essential part of the detector test so I removed it and split the test into two. - Jiang Yan ----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/24123/#review49186 ----------------------------------------------------------- On July 31, 2014, 11:23 p.m., Jiang Yan Xu wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > https://reviews.apache.org/r/24123/ > ----------------------------------------------------------- > > (Updated July 31, 2014, 11:23 p.m.) > > > Review request for mesos and Ben Mahler. > > > Bugs: MESOS-1655 > https://issues.apache.org/jira/browse/MESOS-1655 > > > Repository: mesos-git > > > Description > ------- > > Fixed a flaky test: ZooKeeperTest.LeaderDetectorTimeoutHandling > > - The original ZK session timeout was the same as AWAIT_READY timeout so it's > possible that AWAIT_READY timed out in a race. > - Split the the test into two and removed the part that waits several seconds > for ZooKeeperTestServer to expire the session which is unnecessary to test > the detector and slows down the test. > > > Diffs > ----- > > src/tests/zookeeper_tests.cpp be9fa06818b96e5170c68810fe16cc472f1f8b28 > > Diff: https://reviews.apache.org/r/24123/diff/ > > > Testing > ------- > > Ran the test for 2000 iterations. > > > Thanks, > > Jiang Yan Xu > >
