> On Aug. 6, 2014, 11:30 a.m., Vinod Kone wrote: > > src/tests/rate_limiting_tests.cpp, lines 973-975 > > <https://reviews.apache.org/r/24343/diff/2/?file=653522#file653522line973> > > > > Isn't this already done by StartMaster()?
Yes. I lost track of the change :) > On Aug. 6, 2014, 11:30 a.m., Vinod Kone wrote: > > src/master/master.hpp, line 792 > > <https://reviews.apache.org/r/24343/diff/2/?file=653520#file653520line792> > > > > seems weird for this to be a CHECK, since the users of this class has > > no idea about the restriction. How about returning an Error instead? > > > > Try<process::Event*> nextPermittedEvent() {} > > > > Also, s/nextPermittedEvent/nextEvent/ ? Removed EventThrottler and made it a dumb struct. Put the logic inside visit() per discussion with Vinod. > On Aug. 6, 2014, 11:30 a.m., Vinod Kone wrote: > > src/master/master.cpp, line 984 > > <https://reviews.apache.org/r/24343/diff/2/?file=653521#file653521line984> > > > > +1. This is not really a continuation i.e., never called asynchronously. In master we already use continuations that are not called asynchronously e.g. __reregisterSlave(), or maybe I don't need to say it's a "continuation". Anyway, __visit() is gone and the behavior the other _visit()s has changed in this new revision, please take a look. > On Aug. 6, 2014, 11:30 a.m., Vinod Kone wrote: > > src/tests/rate_limiting_tests.cpp, line 1048 > > <https://reviews.apache.org/r/24343/diff/2/?file=653522#file653522line1048> > > > > s/reach/processed/ ? I meant to comment on the capacity with "reach". I rephrased it to be: // Send two messages which will be queued up. This will reach but not // exceed the capacity. > On Aug. 6, 2014, 11:30 a.m., Vinod Kone wrote: > > src/tests/rate_limiting_tests.cpp, line 1071 > > <https://reviews.apache.org/r/24343/diff/2/?file=653522#file653522line1071> > > > > how are you making sure the deactivate message is received by the > > master? DeactivateFrameworkMessage is sent by the driver upon receiving the error message so with Clock::settle() we can be sure the master has received it (not necessarily processed). Should there be more comment to explain this? - Jiang Yan ----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/24343/#review49761 ----------------------------------------------------------- On Aug. 6, 2014, 4:13 p.m., Jiang Yan Xu wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > https://reviews.apache.org/r/24343/ > ----------------------------------------------------------- > > (Updated Aug. 6, 2014, 4:13 p.m.) > > > Review request for mesos, Ben Mahler and Vinod Kone. > > > Bugs: MESOS-1578 > https://issues.apache.org/jira/browse/MESOS-1578 > > > Repository: mesos-git > > > Description > ------- > > See summary. > > > Diffs > ----- > > include/mesos/mesos.proto 628cce12d2fae645d2ef55e4809631ca03a56207 > src/examples/load_generator_framework.cpp > 7d94c49cf91bf327ac80f04d9f1a7370996b6ba4 > src/master/master.hpp d8a4d9e04ecff60020b99ea6447055787d187797 > src/master/master.cpp a925a936bc1cd878f1c9ff63a9dcc4ee808ec7a5 > src/tests/cluster.hpp d857fc604f5b47cf9805b37c13051093f486ba31 > src/tests/rate_limiting_tests.cpp fc23a1946ad1a78e699552440df2193ea10dc472 > > Diff: https://reviews.apache.org/r/24343/diff/ > > > Testing > ------- > > make check > > ./bin/mesos-tests.sh --verbose --gtest_filter=*RateLimit* --gtest_repeat=1000 > > > Thanks, > > Jiang Yan Xu > >
