----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/65026/#review195104 -----------------------------------------------------------
Ship it! The description is helpful but probably needs some re-wording for a reader with less context? e.g. the old code used the FD based version, etc. - Benjamin Mahler On Jan. 8, 2018, 6:14 p.m., Michael Park wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > https://reviews.apache.org/r/65026/ > ----------------------------------------------------------- > > (Updated Jan. 8, 2018, 6:14 p.m.) > > > Review request for mesos and Benjamin Mahler. > > > Bugs: MESOS-8375 > https://issues.apache.org/jira/browse/MESOS-8375 > > > Repository: mesos > > > Description > ------- > > The existing `ResourcesState::recoverResources` had complicated logic > for reading back the checkpointed resources. The logic is typically > used not merely for a sequence of messages, but rather for dynamically > appended sequence of messages (e.g. status updates). The `Resources` > however is not a dynamically appended sequence of messages, but rather > the full state and does not require truncation of a partial message. > > > Diffs > ----- > > src/slave/state.hpp 01abb5095cf09dd36d7f154d007336892bd8df46 > src/slave/state.cpp 5428b341b061c8209a5cfe63e17b71467d9f4c48 > src/tests/slave_recovery_tests.cpp e305d7430b64409fdac84a33ee072707471d2c49 > > > Diff: https://reviews.apache.org/r/65026/diff/1/ > > > Testing > ------- > > > Thanks, > > Michael Park > >
