----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/16335/#review30592 -----------------------------------------------------------
core/src/main/scala/kafka/controller/ControllerChannelManager.scala <https://reviews.apache.org/r/16335/#comment58600> If connectToBroker timed out here, the thread will die here. So shall we move it into doWork, and let the first attempt trying to connect? core/src/main/scala/kafka/controller/ControllerChannelManager.scala <https://reviews.apache.org/r/16335/#comment58598> Could we also log the whole request content? Since this happens rarely this should not pollute the logs. core/src/main/scala/kafka/controller/ControllerChannelManager.scala <https://reviews.apache.org/r/16335/#comment58597> In cases when the toBroker is not responsive for a long time, the connectToBroker can still throw an exception and hence not being able to send the request. Will this be an issue? - Guozhang Wang On Dec. 18, 2013, 12:29 a.m., Neha Narkhede wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > https://reviews.apache.org/r/16335/ > ----------------------------------------------------------- > > (Updated Dec. 18, 2013, 12:29 a.m.) > > > Review request for kafka. > > > Bugs: KAFKA-1187 > https://issues.apache.org/jira/browse/KAFKA-1187 > > > Repository: kafka > > > Description > ------- > > Changes include 1) Handled error connecting while adding broker in controller > 2) Retry sending state change requests in controller's background thread > > > Diffs > ----- > > core/src/main/scala/kafka/controller/ControllerChannelManager.scala > 7991e42aac4e1d9aaa4c8b185b55fd6172a43f83 > core/src/main/scala/kafka/producer/SyncProducer.scala > 419156eb143fb04a305c91c964307a89ba5a82fa > > Diff: https://reviews.apache.org/r/16335/diff/ > > > Testing > ------- > > > Thanks, > > Neha Narkhede > >
