> On Aug. 26, 2015, 11:13 p.m., Joseph Wu wrote: > > We probably want to keep the check. It's good defensive coding. > > Yong Qiao Wang wrote: > Thanks Joseph Wu for your comments. > > But Mesos now always hardcode the * role in roleSorter, so do we real > need to have this obviously useless check?
I think it is worthwhile having this check because the fact that the 'Master' actor adds "*" to roles is unknown to the "Allocator" actor. In other words, it needs non-local reasoning. - Vinod ----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/37784/#review96610 ----------------------------------------------------------- On Aug. 31, 2015, 7:58 a.m., Yong Qiao Wang wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > https://reviews.apache.org/r/37784/ > ----------------------------------------------------------- > > (Updated Aug. 31, 2015, 7:58 a.m.) > > > Review request for mesos, Benjamin Hindman, Jie Yu, Joseph Wu, and Vinod Kone. > > > Bugs: MESOS-3301 > https://issues.apache.org/jira/browse/MESOS-3301 > > > Repository: mesos > > > Description > ------- > > Remove the redundant check in HierarchicalDRFAlocator. > > > Diffs > ----- > > src/master/allocator/mesos/hierarchical.hpp > 38f8fd2c84314bb3731684d0e9795cb4f50a227e > > Diff: https://reviews.apache.org/r/37784/diff/ > > > Testing > ------- > > > Thanks, > > Yong Qiao Wang > >