> On Jan. 4, 2018, 12:50 a.m., Vinod Kone wrote: > > src/master/master.cpp > > Lines 2013 (patched) > > <https://reviews.apache.org/r/64929/diff/1/?file=1930007#file1930007line2013> > > > > Don't we have the same problem in `SlaveObserver::markUnrechable`? If > > yes, we should probably mention this issue there as well.
I don't think so, we support discarding in the slave observer when another ping comes in while the rate limiting token is outstanding. - Benjamin ----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/64929/#review194730 ----------------------------------------------------------- On Jan. 3, 2018, 10:48 p.m., Benjamin Mahler wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > https://reviews.apache.org/r/64929/ > ----------------------------------------------------------- > > (Updated Jan. 3, 2018, 10:48 p.m.) > > > Review request for mesos, Benno Evers and Vinod Kone. > > > Bugs: MESOS-8386 > https://issues.apache.org/jira/browse/MESOS-8386 > > > Repository: mesos > > > Description > ------- > > In the master failover case of marking agents unreachable, we > do not discard the rate limiting tokens when the agents re-register. > This means that the rate will be slowed down due to the no-op > operations that continue to obtain rate limiting tokens. > > > Diffs > ----- > > src/master/master.cpp 03eb178fa1af7d55ae387e6cb42cdc8d721a2196 > > > Diff: https://reviews.apache.org/r/64929/diff/1/ > > > Testing > ------- > > make check > > > Thanks, > > Benjamin Mahler > >
