----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/64589/#review193769 -----------------------------------------------------------
Fix it, then Ship it! src/slave/slave.cpp Line 3847 (original), 3847 (patched) <https://reviews.apache.org/r/64589/#comment272411> I feel like there are too many helper functions in different places. Is there any uses of `stripAllocationInfo` other than applying resource conversions? If not then it seems to make more sense to me that we have one strip or unallocate function and a get RP ID method for `ResourceConversion`, then keep the original prototype for `Slave::apply`. Otherwise LGTM to this patch. - Chun-Hung Hsiao On Dec. 14, 2017, 1:38 a.m., Jie Yu wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > https://reviews.apache.org/r/64589/ > ----------------------------------------------------------- > > (Updated Dec. 14, 2017, 1:38 a.m.) > > > Review request for mesos, Benjamin Bannier, Greg Mann, and Jan Schlicht. > > > Repository: mesos > > > Description > ------- > > The bug is introduced in this patch: > https://reviews.apache.org/r/64477/ > > Given that we also keep track of each resource provider's total > resources in addition to the total resources of the agent, we need to > make sure we update those totals after applying an operation. > > The bug may manifest as a CHECK failure in the agent that checks if > `totalResources` of the agent is a super set of all the resource > provider resources. > > > Diffs > ----- > > src/slave/slave.hpp de2b2e2e81ed860e6a33ce1b93d859d816ba1021 > src/slave/slave.cpp e8f7591dc0d57ca8a0eb72f6c1c008d4005a524d > > > Diff: https://reviews.apache.org/r/64589/diff/1/ > > > Testing > ------- > > make check > > > Thanks, > > Jie Yu > >
