> On 十一月 23, 2015, 12:58 a.m., Guangya Liu wrote: > > src/slave/slave.cpp, line 4419 > > <https://reviews.apache.org/r/40379/diff/2/?file=1135192#file1135192line4419> > > > > I think that the oversubscribable resources should always > > has_revocable()? For which case does the oversubscribable resource do not > > has_revocable()? Caused by a bad resource estimitator? For such case, can > > we let slave throw exception? > > Klaus Ma wrote: > Should we keep the compatitility?
Yes, we should keep compatibility. But the problem is that all of the revocable resources no matter usage_slack or allocation_slack should has_revocable()? Comments? - Guangya ----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/40379/#review107527 ----------------------------------------------------------- On 十一月 24, 2015, 11:02 a.m., Klaus Ma wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > https://reviews.apache.org/r/40379/ > ----------------------------------------------------------- > > (Updated 十一月 24, 2015, 11:02 a.m.) > > > Review request for mesos and Guangya Liu. > > > Bugs: MESOS-3930 > https://issues.apache.org/jira/browse/MESOS-3930 > > > Repository: mesos > > > Description > ------- > > In Optimistic Offer Phase 1, it introduce `RevocableInfo::type`: USAGE_SLACK > for Oversubscription and ALLOCATION_SLACK for Optimistic Offer. Slave helps > to update `RevocableInfo::type` for Oversubscription. > > > Diffs > ----- > > src/slave/slave.cpp 9055f2a > > Diff: https://reviews.apache.org/r/40379/diff/ > > > Testing > ------- > > make (make check is on-going) > > > Thanks, > > Klaus Ma > >