> On July 23, 2014, 4:56 p.m., Jay Kreps wrote: > > Is this better than using a condition? I was imagining replacing sleep with > > shutdownCondition.await(backoffMs, TimeUnit.MILLISECONDS) and then in > > shutdown() we call shutdownCondition.signal(). > > Jay Kreps wrote: > Actually a condition isn't quite right as there is a subtle race between > signal and await. Better would be the same approach but using > CountDownLatch.await(backoff, TimeUnit.MS)
Ok..I got your point. I uploaded a new patch which uses CountDownLatch. For this i need to override shutdown method in CleanerThread. - Manikumar Reddy ----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/23858/#review48532 ----------------------------------------------------------- On July 24, 2014, 3:11 p.m., Manikumar Reddy O wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > https://reviews.apache.org/r/23858/ > ----------------------------------------------------------- > > (Updated July 24, 2014, 3:11 p.m.) > > > Review request for kafka. > > > Bugs: KAFKA-1544 > https://issues.apache.org/jira/browse/KAFKA-1544 > > > Repository: kafka > > > Description > ------- > > Used CountDownLatch in CleanerThread to decrease LogCleaner shutdown time > during shutdown > > > Diffs > ----- > > core/src/main/scala/kafka/log/LogCleaner.scala > 2faa196a4dc612bc634d5ff5f5f275d09073f13b > > Diff: https://reviews.apache.org/r/23858/diff/ > > > Testing > ------- > > > Thanks, > > Manikumar Reddy O > >
