> On Feb. 5, 2015, 7:13 p.m., Vinod Kone wrote: > > src/Makefile.am, lines 466-469 > > <https://reviews.apache.org/r/29925/diff/6/?file=848172#file848172line466> > > > > How about this structure for consistency: > > > > master/allocator/allocator.hpp > > master/allocator/mesos/hierarchical.hpp > > master/allocator/sorter/sorter.hpp > > master/allocator/sorter/drf/sorter.hpp > > Alexander Rukletsov wrote: > > master/allocator/* > > I would argue `allocation/` is a more general term, that unifies not only > allocators, but also sorters and whatever comes in the future related to > resource allocation. For example, we have `authentication/`. > > > master/allocator/sorter/drf/sorter.hpp > > Could you please elaborate on what consistency do you mean? I took the > `containerizer/isolators/` folder as an example: > > ``` > ??? containerizer > ? ??? isolator.hpp > ? ??? isolators > ? ? ??? cgroups > ? ? ??? filesystem > ? ? ??? posix > ? ? ? ??? disk.cpp > ? ? ? ??? disk.hpp > ? ? ??? posix.hpp > ? ??? mesos > ? ??? containerizer.cpp > ? ??? containerizer.hpp > ``` > > Vinod Kone wrote: > we used 'authentication' because there are two top level components > there, 'authenticator' and 'authenticatee'. a 'sorter' on the other hand > belongs to (or resides in or used by) an 'allocator'. most of our directories > inside src/ are named after the components (e.g, authorizer, watcher, > zookeeper, launcher, etc). > > regarding consistency, i meant an interface is at the top level (e.g, > slave/containerizer/containerizer.hpp) and a specfic implementation should be > in a sub-directory named after the implementation > (slave/containerizer/mesos/containerizer.hpp). > > also, chatting with benm, we might even want to put the drf sorter under > "mesos" because it's only used by mesos allocator for now, but i'll leave > that upto you. > > does that make sense?
Agreed, makes sense. Let's keep `sorter` folder on the same level with `mesos` since theoretically we can combine different sorters with different allocators. - Alexander ----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/29925/#review71273 ----------------------------------------------------------- On Feb. 4, 2015, 4:49 p.m., Alexander Rukletsov wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > https://reviews.apache.org/r/29925/ > ----------------------------------------------------------- > > (Updated Feb. 4, 2015, 4:49 p.m.) > > > Review request for mesos, Bernd Mathiske, Niklas Nielsen, and Vinod Kone. > > > Bugs: MESOS-2213 > https://issues.apache.org/jira/browse/MESOS-2213 > > > Repository: mesos > > > Description > ------- > > See summary. > > > Diffs > ----- > > src/Makefile.am 07bea1f > src/local/local.cpp 76e73a4 > src/master/allocation/allocator.hpp PRE-CREATION > src/master/allocation/mesos/hierarchical.hpp PRE-CREATION > src/master/allocation/sorter.hpp PRE-CREATION > src/master/allocation/sorters/drf.hpp PRE-CREATION > src/master/allocation/sorters/drf.cpp PRE-CREATION > src/master/allocator.hpp 318a756 > src/master/drf_sorter.hpp 5a66313 > src/master/drf_sorter.cpp 584e26c > src/master/hierarchical_allocator_process.hpp 6b44892 > src/master/main.cpp e5e76ce > src/master/master.cpp 1005686 > src/master/sorter.hpp 8915c61 > src/tests/cluster.hpp 90fda52 > src/tests/fault_tolerance_tests.cpp f927d4a > src/tests/hierarchical_allocator_tests.cpp f44d9e9 > src/tests/master_allocator_tests.cpp 018a6cc > src/tests/master_authorization_tests.cpp 20adaa9 > src/tests/master_slave_reconciliation_tests.cpp 04806ed > src/tests/master_tests.cpp 678d27f > src/tests/mesos.hpp 17c2d8f > src/tests/rate_limiting_tests.cpp 7f5ca25 > src/tests/resource_offers_tests.cpp 24a7eab > src/tests/sorter_tests.cpp 520a42e > > Diff: https://reviews.apache.org/r/29925/diff/ > > > Testing > ------- > > make check (Ubuntu, OS X) > > > Thanks, > > Alexander Rukletsov > >