Hi, Please also note that the “auto.offset.reset” property is only respected when there is no offsets under the same consumer group in ZK. So, currently, in order to make sure you read from the latest / earliest offsets every time you restart your Flink application, you’d have to use an unique groupId on each restart.
We’re currently working on new configuration for the Kafka consumer to explicitly configure the starting offset / position without respecting existing offsets in ZK. You can follow the corresponding JIRA here: https://issues.apache.org/jira/browse/FLINK-4280. Regards, Gordon On August 5, 2016 at 8:47:32 PM, Stefan Richter (s.rich...@data-artisans.com) wrote: Sorry, I think you are actually asking for the largest offset in the Kafka source, which makes it setProperty("auto.offset.reset", "largest"). Am 05.08.2016 um 14:44 schrieb Stefan Richter <s.rich...@data-artisans.com>: Hi, I think passing properties with setProperty("auto.offset.reset", "smallest“) to the Kafka consumer should do what you want. Best, Stefan Am 05.08.2016 um 14:36 schrieb Mao, Wei <wei....@intel.com>: I am doing some performance tests with Flink (1.0.3 )+ Kafka (0.8.2.2). And I noticed that when I restarted my Flink application, it reads records starting from the latest offset that I consumed last time, but not the latest offsets of that topic in Kafka. So Is there any way to make it read from last offsets of broker/MyTopic instead of consumer/MyTopic in Flink? Thanks, William