Nope, linux doesn't work. Let me debug why it's not triggered.
On Tue, Mar 25, 2014 at 9:21 AM, Bae, Jae Hyeon <metac...@gmail.com> wrote: > Hm... I cannot reproduce in my local, I downloaded kafka_2.8.0-0.8.1 > package but it didn't work. Let me try in my linux machine. > > > On Mon, Mar 24, 2014 at 6:11 PM, Neha Narkhede <neha.narkh...@gmail.com>wrote: > >> I think you are trying to introduce a session expiration, then could you >> try to do the following and see if you can reproduce the session >> expiration? >> >> ./bin/kafka-server-start.sh config/server.properties >> kill -SIGSTOP <kafka_server_pid> >> sleep 6s >> >> At this point, the session will be expired and the node will disappear >> from >> zookeeper. Then you can do the following - >> >> kill -SIGCONT <kafka_server_pid> >> >> At this point, you should see the following log message from inside the >> handleNewSession() method - >> >> INFO re-registering broker info in ZK for broker 0 >> (kafka.server.KafkaHealthcheck) >> >> Hope that helps. >> >> Thanks, >> Neha >> >> >> >> On Mon, Mar 24, 2014 at 3:19 PM, Bae, Jae Hyeon <metac...@gmail.com> >> wrote: >> >> > Hi >> > >> > On zookeeper session timeout due to some stopping the world long GC >> pause >> > or zookeeper server outage, Ephemeral nodes on kafka broker and consumer >> > should be recreated but in my test environment, handleNewSession() is >> not >> > called. >> > >> > My test scenario is, starting kafka broker locally and put a breakpoint >> > somewhere to simulate long pause, and then, I expected >> handleNewSession() >> > should be called, but it was not and I saw kafka broker zk registration >> is >> > gone. >> > >> > Previously, to avoid this problem, I overrode zkclient implementation >> > internally to replace createEphemeral() function call with Apache >> Curator's >> > PersistentEphemeralNodes and for reinstating watchers, I implemented >> > ConnectionStateListener to reinstate all watchers when RECONNECTED >> happens. >> > >> > Do you know why I cannot reproduce handleNewSession()? >> > >> > Thank you >> > Best, Jae >> > >> > >