> On July 4, 2016, 11:29 p.m., Guangya Liu wrote: > > src/tests/hierarchical_allocator_tests.cpp, lines 3208-3218 > > <https://reviews.apache.org/r/49610/diff/1/?file=1436511#file1436511line3208> > > > > Does there are any special reason that you are using `dynamic > > reservation` here? Seems using `static reservation` may be more simple as > > your test is against allocation but not dynamic reservation.
I think using dynamic reservations is better, because we're partly trying to test the behavior of `Resources::createStrippedScalarQuantity` in the allocator: stripping dynamic reservation info is okay, but stripping static role information is not. - Neil ----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/49610/#review140704 ----------------------------------------------------------- On July 4, 2016, 8:39 p.m., Neil Conway wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > https://reviews.apache.org/r/49610/ > ----------------------------------------------------------- > > (Updated July 4, 2016, 8:39 p.m.) > > > Review request for mesos and Alexander Rukletsov. > > > Repository: mesos > > > Description > ------- > > This test checks that when setting aside unallocated resources to ensure > that a quota guarantee can be met, we don't use resources that have been > reserved for a different role. > > > Diffs > ----- > > src/tests/hierarchical_allocator_tests.cpp > 0498cd5e54b0e4b87a767585a77699653aa52179 > > Diff: https://reviews.apache.org/r/49610/diff/ > > > Testing > ------- > > `./src/mesos-tests > --gtest_filter="HierarchicalAllocatorTest.QuotaSetAsideReservedResources" > --gtest_repeat=100 --gtest_break_on_failure` passes. When you apply a patch > that breaks quota allocation behavior > (https://gist.github.com/neilconway/97eed34e9b0b312b8511f0f4f57fe061), > previously `make check` passes, but it fails with this new test added. > > > Thanks, > > Neil Conway > >
