> On June 27, 2019, 10:34 p.m., Greg Mann wrote: > > src/tests/slave_tests.cpp > > Lines 12362-12369 (patched) > > <https://reviews.apache.org/r/70960/diff/2/?file=2152614#file2152614line12362> > > > > I'm not sure why we would receive multiple TASK_LOST updates? I would > > expect just one. > > > > In any case, it seems perhaps easier to read if the test registers an > > expectation of the TASK_LOST update and AWAITs on that, then only once that > > is satisfied, registers the expectation for the TASK_KILLED retry, advances > > the clock, and awaits on that. WDYT?
The test was rewritten and does not rely on resent status updates anymore. We do not handle multiple updates anymore. - Benjamin ----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/70960/#review216203 ----------------------------------------------------------- On June 28, 2019, 1:11 a.m., Benjamin Bannier wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > https://reviews.apache.org/r/70960/ > ----------------------------------------------------------- > > (Updated June 28, 2019, 1:11 a.m.) > > > Review request for mesos, Greg Mann and Joseph Wu. > > > Bugs: MESOS-9860 > https://issues.apache.org/jira/browse/MESOS-9860 > > > Repository: mesos > > > Description > ------- > > This patch adds a test which confirms that the agent leaves a draining > state on its own once all frameworks on the agent have no more pending > tasks and all their executors have neither launched or queued tasks. > > The test uses the fact that the agent rejects task launches while > draining. > > > Diffs > ----- > > src/tests/slave_tests.cpp 7ba2df9fef9966cd09e0af3f4ef3bb260b6167f6 > > > Diff: https://reviews.apache.org/r/70960/diff/3/ > > > Testing > ------- > > `make check` > > > Thanks, > > Benjamin Bannier > >
