> On Sept. 9, 2013, 11:44 p.m., Neha Narkhede wrote: > > core/src/main/scala/kafka/client/ClientUtils.scala, line 100 > > <https://reviews.apache.org/r/14041/diff/2/?file=349971#file349971line100> > > > > The concern I have with this change is the potential performance hit to > > the consumer rebalance process. Since it affects the end-to-end message > > latency, I would encourage you to run some performance tests for both a > > MirrorMaker that consumes ~500 topics as well as a consumer that consumes 5 > > topics, to quantify the performance hit. > > > > Also, what do you think about the alternate solution that I described > > on the JIRA?
Yeah I will do the perf test. About the alternative approach, that would require the handling api to recognize if the current broker is controller or not; and if yes, hand over the request to the controller module. I think that would complicate the logic. What do you think? - Guozhang ----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/14041/#review26005 ----------------------------------------------------------- On Sept. 9, 2013, 9:31 p.m., Guozhang Wang wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > https://reviews.apache.org/r/14041/ > ----------------------------------------------------------- > > (Updated Sept. 9, 2013, 9:31 p.m.) > > > Review request for kafka. > > > Bugs: KAFKA-1030 > https://issues.apache.org/jira/browse/KAFKA-1030 > > > Repository: kafka > > > Description > ------- > > Using the approach of reading directly from ZK. > > > Diffs > ----- > > core/src/main/scala/kafka/client/ClientUtils.scala > cc526ec933052b239f0e7ce43e76cd9d011d5bd9 > core/src/main/scala/kafka/consumer/ZookeeperConsumerConnector.scala > e7a692a1d23ca5a9ecf86e3cb34be418b9c0c943 > > Diff: https://reviews.apache.org/r/14041/diff/ > > > Testing > ------- > > unit tests > > > Thanks, > > Guozhang Wang > >
