> On July 24, 2015, 4:13 p.m., Jun Rao wrote: > > core/src/main/scala/kafka/network/SocketServer.scala, line 264 > > <https://reviews.apache.org/r/36652/diff/3/?file=1020607#file1020607line264> > > > > Not sure if it's better to keep the thread alive on any throwable. For > > unexpected exceptions, it seems it's better to just propagate the > > exception, log it and then kill the thread. This is already done through > > Utils.newThread. If we want to clean things up (e.g. countdown) before > > propagating the exception, we can do that in a finally clause.
Hi Jun, I think only letting the acceptor thread die might cause more issue. The broker in that case will not serve new connections. This essentially makes all the partitions on this broker become unavailable. If we let the accpetor thread exit, maybe we should shutdown the broker completely. - Jiangjie ----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/36652/#review92920 ----------------------------------------------------------- On July 24, 2015, 4:36 a.m., Mayuresh Gharat wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > https://reviews.apache.org/r/36652/ > ----------------------------------------------------------- > > (Updated July 24, 2015, 4:36 a.m.) > > > Review request for kafka. > > > Bugs: KAFKA-2351 > https://issues.apache.org/jira/browse/KAFKA-2351 > > > Repository: kafka > > > Description > ------- > > Added a try-catch to catch any exceptions thrown by the nioSelector > > > Addressed comments on the Jira ticket > > > Diffs > ----- > > core/src/main/scala/kafka/network/SocketServer.scala > 91319fa010b140cca632e5fa8050509bd2295fc9 > > Diff: https://reviews.apache.org/r/36652/diff/ > > > Testing > ------- > > > Thanks, > > Mayuresh Gharat > >
