----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/38251/#review99632 -----------------------------------------------------------
Thanks for fixing this @gyliu ! Just a minor comment about a alternate simple approach that would allow us to not duplicate the same log messages twice. I commented only on the HTTP workflow but the same comment also applies to the PID based workflow. src/master/master.cpp (line 1959) <https://reviews.apache.org/r/38251/#comment156569> I wonder if we can simply do this instead of duplicating the log messages twice. What do you think ? ( Don't copy my suggested comments in the code-snippet though :-) ) ``` if (framework->connected && !subscribe.force()) { FrameworkErrorMessage message; ... blah blah http.close(); return; } // It's now safe to update the framework fields since the request is now gurranteed to be successful. LOG(INFO) << "Updating info for framework " << framework->id(); framework->updateFrameworkInfo(frameworkInfo); allocator->updateFramework(framework->id(), framework->info); framework->reregisteredTime = Clock::now(); if (subscribe.force()) { // already existing logic blah blah. } else { // already existing logic blah blah. } ``` - Anand Mazumdar On Sept. 14, 2015, 2:19 a.m., Guangya Liu wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > https://reviews.apache.org/r/38251/ > ----------------------------------------------------------- > > (Updated Sept. 14, 2015, 2:19 a.m.) > > > Review request for mesos, Joris Van Remoortere and Vinod Kone. > > > Bugs: MESOS-3169 > https://issues.apache.org/jira/browse/MESOS-3169 > > > Repository: mesos > > > Description > ------- > > FrameworkInfo should only be updated if the re-registration is valid > > > Diffs > ----- > > src/master/master.cpp 4b60e637b19e749e27c4a8eb9b0a95c7e9305c5f > > Diff: https://reviews.apache.org/r/38251/diff/ > > > Testing > ------- > > > Thanks, > > Guangya Liu > >
