----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/71435/#review217570 -----------------------------------------------------------
Ship it! Ship It! - Meng Zhu On Sept. 4, 2019, 10:53 a.m., Andrei Sekretenko wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > https://reviews.apache.org/r/71435/ > ----------------------------------------------------------- > > (Updated Sept. 4, 2019, 10:53 a.m.) > > > Review request for mesos, Benjamin Mahler and Meng Zhu. > > > Bugs: MESOS-1452 and MESOS-9949 > https://issues.apache.org/jira/browse/MESOS-1452 > https://issues.apache.org/jira/browse/MESOS-9949 > > > Repository: mesos > > > Description > ------- > > Currently, when accepting offers for a slave already removed from > the master, Master::_accept() dispatches `recoverResources()` > unconditionally (with empty resources if, due to framework/slave removal > the offers were already rescinded and their resources recovered). > > The depending patch removes this redundant recovery of empty resources, > thus exposing a race in some of `MasterAuthorizationTest.*` tests > between `FUTURE_DISPATCH(_, &MesosAllocatorProcess::recoverResources)` > and the dispatch which performs the actual resource recovery. > > This patch is fixing these tests to wait for the first dispatch of > `recoverResources()`. > > > Diffs > ----- > > src/tests/master_authorization_tests.cpp > 29325642d17fa00067eacfa7415686bb42b7b137 > > > Diff: https://reviews.apache.org/r/71435/diff/1/ > > > Testing > ------- > > `./bin/mesos-tests.sh --gtest_filter="*MasterAuthorizationTest*" > --gtest_break_on_failure --gtest_repeat=1000` > > > Thanks, > > Andrei Sekretenko > >
