> On Nov. 7, 2017, 5:09 p.m., Benjamin Mahler wrote: > > src/master/allocator/mesos/hierarchical.hpp > > Lines 319 (patched) > > <https://reviews.apache.org/r/63175/diff/1/?file=1864532#file1864532line319> > > > > I think this should be `connected`? > > > > Right now there's some confusion around `active` and `connected`. > > `active` in the past meant that the framework was receiving offers, but we > > never added the ability for frameworks to change this state. When we added > > suppression, that was equivalent to deactivation (note that activation has > > become a per-role thing). > > > > Mostly `connected` == `active`. But I think the point of this boolean > > is to track whether we can talk to the framework? Thoughts?
`active` is already used in the class and this review doesn't change its meaning so I think we can defer this for later? For activeness vs. connectedness, I feel it's a larger effort to address it consistently. I'll follow up on this. - Jiang Yan ----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/63175/#review190412 ----------------------------------------------------------- On Nov. 7, 2017, 5:02 p.m., Jiang Yan Xu wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > https://reviews.apache.org/r/63175/ > ----------------------------------------------------------- > > (Updated Nov. 7, 2017, 5:02 p.m.) > > > Review request for mesos, Benjamin Mahler, James Peach, and Joseph Wu. > > > Bugs: MESOS-8085 > https://issues.apache.org/jira/browse/MESOS-8085 > > > Repository: mesos > > > Description > ------- > > Do not generate UnavailableResources for inactive frameworks. > > > Diffs > ----- > > src/master/allocator/mesos/hierarchical.hpp > c2346054b2c98516f15ab8ce2dc798224ff4def4 > src/master/allocator/mesos/hierarchical.cpp > 5b6efe5faa3c3b10f1f714f582a155b368f8ccaf > > > Diff: https://reviews.apache.org/r/63175/diff/1/ > > > Testing > ------- > > make check. > > I didn' write a new test as the externally observable behavior doesn't change. > > > Thanks, > > Jiang Yan Xu > >
