Hi, The broker is healthy with no exceptions but the OffsetRequest in getOffsetsBefore method fails with ClosedChannelException. This usually happens after 2-3 hours of the app restart. I have looked at the code of ConsumerOffsetChecker and I use the same logic. I do connect to the leader to issue request. I have also increased the socket-timeout-ms to a very high number in Simple consumer and the issue still persists.
Did anyone face this issue before? Any help in this would be great. Thanks. On Tue, May 12, 2015 at 11:47 AM, Achanta Vamsi Subhash < achanta.va...@flipkart.com> wrote: > Hi James, > > The thread was different from the one I was asking. When I issue a > OffsetFetchRequest to the channel, I am getting a ClosedChannelException > and all of them seem to come from a single broker. I don't see anything > problematic with the broker (config, exceptions.etc.) but requests for > OffsetFetch fail. The same request to other brokers succeeds. > > How can we debug this? > > On Mon, May 11, 2015 at 11:15 PM, James Cheng <jch...@tivo.com> wrote: > >> Vamsi, >> >> There is another thread going on right now about this exact topic: >> >> "Is there a way to know when I've reached the end of a partition >> (consumed all messages) when using the high-level consumer?" >> http://search-hadoop.com/m/uyzND1Eb3e42NMCWl >> >> -James >> >> On May 10, 2015, at 11:48 PM, Achanta Vamsi Subhash < >> achanta.va...@flipkart.com> wrote: >> >> > Hi, >> > >> > What is the best way for finding out the log end offset for a topic? >> > Currently I am using the SimpleConsumer getLastOffset logic mentioned >> in: >> > >> https://cwiki.apache.org/confluence/display/KAFKA/0.8.0+SimpleConsumer+Example >> > >> > But we are running into ClosedChannelException for some of the topics. >> We >> > use Kafka for offset storage and version 0.8.2.1. >> > >> > What is the ideal way to compute the topic log end offset? >> > >> > -- >> > Regards >> > Vamsi Subhash >> >> > > > -- > Regards > Vamsi Subhash > -- Regards Vamsi Subhash
