----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/20507/#review47923 -----------------------------------------------------------
src/master/master.cpp <https://reviews.apache.org/r/20507/#comment84160> Instead of persisting completed frameworks in the registry, I believe we would only persist registered (admitted) frameworks. These frameworks could be re-registered (readmitted) so long as they are in the registry. Or, if the registry is non-strict then we would readmit regardless, akin to what is done for slaves. Could you update the TODO accordingly? src/master/master.cpp <https://reviews.apache.org/r/20507/#comment84162> As Adam mentioned, this log message seems misleading because it's not necessarily after a failover timeout. Should we just say: "Framework FID attempted to re-register but is already completed" or "Completed framework FID attempted to re-register" This seems like a WARNING from the master's perspective, unless the importance of the framework is why you're making it an ERROR? src/tests/fault_tolerance_tests.cpp <https://reviews.apache.org/r/20507/#comment84164> Should we have another test for when the scheduler unregisters itself via stop()? - Ben Mahler On July 15, 2014, 12:38 a.m., Vinod Kone wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > https://reviews.apache.org/r/20507/ > ----------------------------------------------------------- > > (Updated July 15, 2014, 12:38 a.m.) > > > Review request for mesos, Benjamin Hindman and Ben Mahler. > > > Bugs: MESOS-1219 > https://issues.apache.org/jira/browse/MESOS-1219 > > > Repository: mesos-git > > > Description > ------- > > This is only a partial fix for MESOS-1219. > > > Diffs > ----- > > src/master/master.cpp 86b147fce153fe3a241dbd841e033f2b7ca07b01 > src/tests/fault_tolerance_tests.cpp > ac65050bec5720b982f53d4dd6797cc3dee285dc > > Diff: https://reviews.apache.org/r/20507/diff/ > > > Testing > ------- > > make check > > > Thanks, > > Vinod Kone > >
