> On Jan. 27, 2015, 7:55 p.m., Ben Mahler wrote: > > src/tests/master_allocator_tests.cpp, lines 80-81 > > <https://reviews.apache.org/r/29927/diff/3/?file=836337#file836337line80> > > > > Can we be very explicit here about why we need this, and what is > > causing the flakiness? How does this relate to my patches to silence the > > mock warnings? > > Alexander Rukletsov wrote: > The ticket https://issues.apache.org/jira/browse/MESOS-1677 is a bit > misleading: it says the test is flaky but shows just a GMOCK warning in the > log. I tried to reproduce "flakiness" without stopping allocator explicitly, > but without success. Let's keep refactor and this issue separate. I suggest > we keep this patch since it restores status quo, reopen MESOS-1677 and remove > this code afterwards if the issue won't be confirmed. Thoughts? > > Ben Mahler wrote: > > I suggest we keep this patch since it restores status quo > > Hm.. am I correct in understanding that this patch corrects MESOS-1677? > If not, what is the status quo that this restores? Note that my mock warning > fix is committed. > > I'm still a bit confused as to why you want this, given the WARNING fix. > :) > > Alexander Rukletsov wrote: > Currently there exists `MasterAllocatorTest::StopAllocator()` that aims > to fix MESOS-1677. During this refactor I temporary removed it (in /r/29890/) > and restore it here. That is what I mean under the status quo. > > Now given "flaky test" in MESOS-1677 means just GMOCK warning, this code > is safe to remove. I want to separate this change from the allocator refactor.
Ok great! Want to go ahead and discard this patch then? - Ben ----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/29927/#review69870 ----------------------------------------------------------- On Jan. 28, 2015, 5:29 p.m., Alexander Rukletsov wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > https://reviews.apache.org/r/29927/ > ----------------------------------------------------------- > > (Updated Jan. 28, 2015, 5:29 p.m.) > > > Review request for mesos, Bernd Mathiske and Niklas Nielsen. > > > Bugs: MESOS-2213 > https://issues.apache.org/jira/browse/MESOS-2213 > > > Repository: mesos-git > > > Description > ------- > > Several tests have been reported flaky if allocation is not ceased explicitly > (see MESOS-1677). To stop allocation, an allocator instance should be > destroyed. To facilitate this, store a pointer to a MockAllocator instance in > the MasterAllocatorTest fixture and let a test case destroy it. > > > Diffs > ----- > > src/tests/master_allocator_tests.cpp 2430622 > > Diff: https://reviews.apache.org/r/29927/diff/ > > > Testing > ------- > > make check (Ubuntu, OS X) > > > Thanks, > > Alexander Rukletsov > >
