Thanks Jaime - That worked However, I am still not able to locate the files within the zookeeper that stores the offset information. Would be of great help if I can locate those
On Sat, Nov 7, 2015 at 4:24 AM, Jaime Solano <[email protected]> wrote: > Hi Birendra, > > What type of bolts are you using in your topology? If your extending from > BaseRichBolt, you have to explicitily ack the tuples, otherwise your > KafkaSpout will try to consume them again after 30 secs. > > -Jaime > On Nov 6, 2015 5:38 PM, "Birendra Kumar Singh" <[email protected]> wrote: > >> In fact, what I see is that once all messages are being consumed - it >> starts consuming again all of them from the beginning.. >> >> On Sat, Nov 7, 2015 at 3:49 AM, Birendra Kumar Singh <[email protected]> >> wrote: >> >>> Can you guys please help me understand how to zk for existence of offset. >>> I tried this on a separate zkroot as well, but the problem still >>> persists. >>> >>> Thanks >>> Birendra >>> >>> On Fri, Nov 6, 2015 at 11:08 PM, Santosh Pingale < >>> [email protected]> wrote: >>> >>>> To add to Stephen's answer, you might want to change your zkroot from >>>> /broker to something else as Kafka uses it to maintain the metadata. >>>> >>>> Nevertheless, you will need to check zk for existence of offset and not >>>> local directories of Kafka cluster. >>>> >>> >>> >>
