----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/67022/#review208520 -----------------------------------------------------------
Fix it, then Ship it! src/slave/slave.cpp Lines 1607 (patched) <https://reviews.apache.org/r/67022/#comment292502> If I understand the slave code correctly, I think this would checkpoint the new state every time we're losing connection to a master (e.g. when the leader changes). I'd suggest moving it into `Slave::recover()`, i.e. the place where we set `requiredMasterCapabilities.agentUpdate` to true in the first place. - Benno Evers On May 9, 2018, 12:40 a.m., Zhitao Li wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > https://reviews.apache.org/r/67022/ > ----------------------------------------------------------- > > (Updated May 9, 2018, 12:40 a.m.) > > > Review request for mesos, Eric Chung and Jason Lai. > > > Repository: mesos > > > Description > ------- > > Refreshed checkpointed SlaveInfo if `--reconfiguration_policy==any`. > > > Diffs > ----- > > src/slave/flags.cpp a319b5ea633c41fd8a252c5e1617ac52d1480ba5 > src/slave/slave.cpp c6d9152d9de4184f9107bb8242b41d468d76e018 > > > Diff: https://reviews.apache.org/r/67022/diff/2/ > > > Testing > ------- > > > Thanks, > > Zhitao Li > >
