I upgraded to consumer 0.9.0.1 as I was running into https://issues.apache.org/jira/browse/KAFKA-2978.
However, I still see this wired issue when I run (the below command) on my consumer group: bin/kafka-consumer-groups.sh --describe --group consumer-gp-1 --new-consumer --bootstrap-server localhost:9092 I notice that for certain partitions the "current offset" will actually go back between subsequent runs of the command. See an example below(Each line represents the result of running the above command at different times (~10 sec apart) for a particular partition): GROUP, TOPIC, PARTITION, CURRENT OFFSET, LOG END OFFSET, LAG, OWNER run1: consumer-gp-1, topic-1, 54, 6682915, 8264593, 1581678, consumer-5_/ 172.18.5.49 run2: consumer-gp-1, topic-1, 54, 8265001, 8265027, 26, consumer-5_/ 172.18.5.49 run3: consumer-gp-1, topic-1, 54, 6682915, 8265252, 1582337, consumer-5_/ 172.18.5.49 run4: consumer-gp-1, topic-1, 54, 8265010, 8265457, 47, consumer-5_/ 172.18.5.49 This is really wired. How can the current offset go back in time and again become current this often