> On Dec. 17, 2013, 1:25 a.m., Bill Farner wrote: > > src/main/java/com/twitter/aurora/scheduler/PulseMonitor.java, line 120 > > <https://reviews.apache.org/r/16247/diff/2/?file=398286#file398286line120> > > > > Without a remove(), this now stands out as a memory leak. We either > > need to remove() at some point or revert to Cache (preferably not relying > > on the map view of it).
Removing an item from the map would erase its history trail and it would be treated as a new one with a completely new initial delay. That will mess up our GC interval predictability per host. I'd rather not do that. - Maxim ----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/16247/#review30511 ----------------------------------------------------------- On Dec. 16, 2013, 6:06 p.m., Maxim Khutornenko wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > https://reviews.apache.org/r/16247/ > ----------------------------------------------------------- > > (Updated Dec. 16, 2013, 6:06 p.m.) > > > Review request for Aurora, Kevin Sweeney and Bill Farner. > > > Repository: aurora > > > Description > ------- > > Modified PulseMonitor to support random initial jitter delay. > > > Diffs > ----- > > src/main/java/com/twitter/aurora/scheduler/PulseMonitor.java > e31ce3d71c895f43ff73ad8e904124443ebdd72a > src/main/java/com/twitter/aurora/scheduler/SchedulerModule.java > bd7929d631cf45b4c2c7f39177bbafbd8f659071 > src/main/java/com/twitter/aurora/scheduler/periodic/GcExecutorLauncher.java > 5a8e13f0d719052286a7136707d7ed1aca87bab2 > src/test/java/com/twitter/aurora/scheduler/PulseMonitorImplTest.java > PRE-CREATION > > src/test/java/com/twitter/aurora/scheduler/periodic/GcExecutorLauncherTest.java > ac5e3102925e2f26a12e18e95817403d18e25cb3 > > Diff: https://reviews.apache.org/r/16247/diff/ > > > Testing > ------- > > gradle build > > > Thanks, > > Maxim Khutornenko > >
