> On June 11, 2019, 7:36 p.m., Benjamin Mahler wrote: > > src/tests/master/update_framework_tests.cpp > > Lines 780 (patched) > > <https://reviews.apache.org/r/70755/diff/2/?file=2147660#file2147660line780> > > > > do you need to resume here for the test to pass?
Yes, it turns out we do. We need some allocation interval to start after the FrameworkInfo update is applied by the allocator. The V1 test acheives this by advancing clock after the update is confirmed by the master. ... wait, now I begin to suspect that the V1 test without resume() has the same issue (albeit less prominent) - there the test can receive the update confirmation BEFORE the allocator process handles updateFramework. I'll need to address that separately. - Andrei ----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/70755/#review215810 ----------------------------------------------------------- On June 12, 2019, 4:44 p.m., Andrei Sekretenko wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > https://reviews.apache.org/r/70755/ > ----------------------------------------------------------- > > (Updated June 12, 2019, 4:44 p.m.) > > > Review request for mesos, Benjamin Mahler and Meng Zhu. > > > Bugs: MESOS-9793 > https://issues.apache.org/jira/browse/MESOS-9793 > > > Repository: mesos > > > Description > ------- > > Added tests for the V0 UPDATE_FRAMEWORK scheduler call. > > > Diffs > ----- > > src/tests/master/update_framework_tests.cpp > 78304c211c5eb9f50ea629121a6dddfe47942e9f > > > Diff: https://reviews.apache.org/r/70755/diff/5/ > > > Testing > ------- > > > Thanks, > > Andrei Sekretenko > >
