----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/71029/#review216422 -----------------------------------------------------------
src/master/master.hpp Lines 2728 (patched) <https://reviews.apache.org/r/71029/#comment303649> If we want this to accurately capture the hierarchical consumption, this needs to loop over all frameworks (i.e. `master->frameworks`) instead. - Benjamin Mahler On July 8, 2019, 6:07 p.m., Benjamin Mahler wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > https://reviews.apache.org/r/71029/ > ----------------------------------------------------------- > > (Updated July 8, 2019, 6:07 p.m.) > > > Review request for mesos, Andrei Sekretenko and Meng Zhu. > > > Bugs: MESOS-9871 > https://issues.apache.org/jira/browse/MESOS-9871 > > > Repository: mesos > > > Description > ------- > > Quota consumption is computed as follows: > > Allocation + Unallocated Reservation == > Reservations + Unreserved Allocation > > That is, reservations count towards quota regardless of whether > they're allocated. Allocation counts towards quota. Offered resources > *do not* count towards quota, this is to (1) provide stability of the > quota consumption metrics in the face of offers flowing in and out, > and (2) to ensure that we treat offers as rescindable and therefore > not yet "counting" towards quota. Also, if in the future schedulers > are offered more than their quota to improve their choices, counting > offered resources as quota consumption will be problematic. > > > Diffs > ----- > > src/master/framework.cpp c35de2574ef683a9f397a957422fdbf8fa751313 > src/master/master.hpp 23dafe746b6f9b3d70ad7220f54c4d49068b8af8 > > > Diff: https://reviews.apache.org/r/71029/diff/1/ > > > Testing > ------- > > Added a test in a subsequent patch. > > > Thanks, > > Benjamin Mahler > >
