> On July 3, 2016, 11:07 p.m., Michael Park wrote: > > src/master/allocator/mesos/hierarchical.cpp, lines 631-642 > > <https://reviews.apache.org/r/49377/diff/2/?file=1434836#file1434836line631> > > > > Any particular reason why this moved?
It seemed cleaner to me to update all of the allocator state first, then update the state of the sorters. The previous coding updated some allocator state, then the sorters, then the rest of the allocator state. > On July 3, 2016, 11:07 p.m., Michael Park wrote: > > src/master/allocator/mesos/hierarchical.cpp, lines 658-662 > > <https://reviews.apache.org/r/49377/diff/2/?file=1434836#file1434836line658> > > > > I don't think I quite follow this. A framework can only be in 1 role > > `R`, and can only operate on resources to make them transition between `*` > > and `R`. I don't see what we would be missing here. Hmmm, fair enough -- comment removed. - Neil ----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/49377/#review140577 ----------------------------------------------------------- On July 1, 2016, 9:47 a.m., Neil Conway wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > https://reviews.apache.org/r/49377/ > ----------------------------------------------------------- > > (Updated July 1, 2016, 9:47 a.m.) > > > Review request for mesos, Alexander Rukletsov, Joris Van Remoortere, and > Michael Park. > > > Bugs: MESOS-5698 > https://issues.apache.org/jira/browse/MESOS-5698 > > > Repository: mesos > > > Description > ------- > > Each DRFSorter tracks the total resources in the cluster. This means > that each sorter must be updated when the resources in the cluster have > changed, e.g., due to the creation of a dynamic reservation or a > persistent volume. In the previous implementation, the quota role sorter > was not updated for non-quota roles when a reservation or persistent > volume was created by a framework. This resulted in inconsistency > between the total resources in the allocator and the quota role sorter. > > This could cause several problems. First, removing a slave from the > cluster would leak resources in the quota role sorter. Second, certain > interleavings of slave removals and reserve/unreserve operations by > frameworks and via HTTP endpoints could lead to CHECK failures. > > > Diffs > ----- > > src/master/allocator/mesos/hierarchical.cpp > 38381237fa6ceb3f21fd0d4b07d7c3787f0129df > src/master/allocator/sorter/drf/sorter.cpp > 967290d4d1100208900b4b724422c3218abc23cb > src/tests/persistent_volume_endpoints_tests.cpp > 971a7b38d26b26c8543815e467f30cbb83ee412c > > Diff: https://reviews.apache.org/r/49377/diff/ > > > Testing > ------- > > make check on OSX and recent Arch Linux. > > > Thanks, > > Neil Conway > >
