> On April 8, 2016, 11:03 p.m., Ben Mahler wrote: > > src/slave/slave.cpp, lines 4989-4991 > > <https://reviews.apache.org/r/45941/diff/1/?file=1337482#file1337482line4989> > > > > Any reason why this CHECK isn't below the VLOG(1) above? Having it > > above seems to make it more clear that we're validating the input from the > > estimator, no? > > > > Also, for the comment here, we should avoid talking about the > > hierarchical allocator since that just happens to be where the failure > > manifests. > > > > Perhaps something like: > > > > ``` > > // Oversubscrbable resources must be considered revocable. > > // > > // TODO(bmahler): Consider tagging input as revocable > > // rather than rejecting and crashing here. > > ```
I'll move it. - Zhitao ----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/45941/#review127914 ----------------------------------------------------------- On April 8, 2016, 8:02 p.m., Zhitao Li wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > https://reviews.apache.org/r/45941/ > ----------------------------------------------------------- > > (Updated April 8, 2016, 8:02 p.m.) > > > Review request for mesos, Ben Mahler and Joseph Wu. > > > Bugs: MESOS-5131 > https://issues.apache.org/jira/browse/MESOS-5131 > > > Repository: mesos > > > Description > ------- > > Add check in agent for incorrect oversubscribed resource. > > I've decided to let agent crash explicitly here instead of fail more silently. > > > Diffs > ----- > > src/slave/slave.cpp f090c853b8affc4be5eecb4f616ec881fc2b60c3 > > Diff: https://reviews.apache.org/r/45941/diff/ > > > Testing > ------- > > Running existing test, and verify manually that offending resource crashes > the agent. > > (Any suggestion to test `CHECK` is welcomed). > > > Thanks, > > Zhitao Li > >