----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/19006/#review37439 -----------------------------------------------------------
src/master/master.cpp <https://reviews.apache.org/r/19006/#comment68971> Lets kill this TODO. It's better to explicitly remove it in exited() rather than doing it here. As an aside, we should really kill the slave checkpointing flag at this point i.e., slave should always checkpoint. src/master/master.cpp <https://reviews.apache.org/r/19006/#comment68973> I don't see any "depends on" in this review. So this should not be here? src/master/master.cpp <https://reviews.apache.org/r/19006/#comment68975> How about renaming this to shutdownSlave to avoid confusion? - Vinod Kone On March 11, 2014, 2:18 a.m., Adam B wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > https://reviews.apache.org/r/19006/ > ----------------------------------------------------------- > > (Updated March 11, 2014, 2:18 a.m.) > > > Review request for mesos and Vinod Kone. > > > Bugs: MESOS-804 > https://issues.apache.org/jira/browse/MESOS-804 > > > Repository: mesos-git > > > Description > ------- > > Moved slave deactivation code out of Master::exited() into new > deactivate(Slave). > > Still need to reconcile this code with deactivateSlave(SlaveId) which right > now just calls removeSlave(Slave). > From what I can tell, checkpointing slaves can just be deactivated, but > non-checkpointing slaves get removed altogether. Correct? > I welcome thoughts/suggestions on how to clean this up. > > > Diffs > ----- > > src/master/master.hpp 49a3e15 > src/master/master.cpp 2a40333 > > Diff: https://reviews.apache.org/r/19006/diff/ > > > Testing > ------- > > make check > > > Thanks, > > Adam B > >
