> On Feb. 8, 2019, 9:47 p.m., Benjamin Mahler wrote: > > src/tests/hierarchical_allocator_tests.cpp > > Lines 2357-2374 (patched) > > <https://reviews.apache.org/r/69890/diff/2/?file=2124579#file2124579line2357> > > > > Both "empty set" and "NullResourcesAllocatable" don't seem to > > accurately describe what's done? > > > > // Add a framework which specifies minimum allocatable resources > > // with a single, empty resource. > > > > Single empty seems more appropriate? Come to think of it, should we > > validate that all entries are non-empty? There doesn't seem to be any point > > in allowing an empty entry, since any single empty entry renders it > > equivalent to the empty set case? > > Benjamin Bannier wrote: > Fixed the test docstring. > > I think there is some value in allowing this going forward. An empty set > of quantities maps onto an empty set of requirements while a set with a > single, empty quantity maps onto a requirement for anything. WDYT? > > Dropping for now, please feel free to reopen. > > Benjamin Mahler wrote: > Well, what's different about those the two requirements? They sound the > same?
I updated the validation code in https://reviews.apache.org/r/69862/ to reject such empty quantity settings now, and removed the two test cases here. - Benjamin ----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/69890/#review212678 ----------------------------------------------------------- On Feb. 13, 2019, 11:31 p.m., Benjamin Bannier wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > https://reviews.apache.org/r/69890/ > ----------------------------------------------------------- > > (Updated Feb. 13, 2019, 11:31 p.m.) > > > Review request for mesos, Benjamin Mahler and Meng Zhu. > > > Bugs: MESOS-9523 > https://issues.apache.org/jira/browse/MESOS-9523 > > > Repository: mesos > > > Description > ------- > > This patch adds tests for the behavior of per-framework, per-role > minimal allocatable resources. We validate the behavior of framework > minimal allocatable resources below the globally configured limits and > the behavior of empty filters or minimal resource quantities set up for > a framework. > > > Diffs > ----- > > src/tests/hierarchical_allocator_tests.cpp > 1366cc51332484d7ff3b5e5b7ba767efb97172c0 > > > Diff: https://reviews.apache.org/r/69890/diff/4/ > > > Testing > ------- > > `make check` > > > Thanks, > > Benjamin Bannier > >