Hi Gordon, Thanks for confirming my understanding that the extractor should not have to be defined for 0.10. However, I'm still experiencing the case where not using an extractor results in zero window triggers.
I've verified the timestamps in the Kafka records with the following command: bin/kafka-console-consumer.sh --bootstrap-server localhost:9092 --topic input --from-beginning --property print.timestamp=true Where the 'input' topic happens to consist of Strings representing the time the record was created. I get output such as the following: CreateTime:1494998828813 1494998828813 CreateTime:1494998828901 1494998828901 CreateTime:1494998828914 1494998828914 CreateTime:1494998828915 1494998828914 CreateTime:1494998828915 1494998828915 CreateTime:1494998829004 1494998829003 CreateTime:1494998829016 1494998829016 CreateTime:1494998829016 1494998829016 where CreateTime is the timestamp generated by Kafka and the second value is the record value. In this particular case that happens to be the time the record was created in the producer, which resides on the same machine as the Kafka broker (hence the identical values). -Jia On Tue, May 16, 2017 at 9:30 PM, Tzu-Li (Gordon) Tai <tzuli...@apache.org> wrote: > Hi Jia! > > This sounds a bit fishy. The docs mention that there is no need for a > timestamp / watermark extractor because with 0.10, the timestamps that come > with Kafka records can be used directly to produce watermarks for event > time. > > One quick clarification: did you also check whether the timestamps that > come with the Kafka 0.10 records are sound and reasonable? > > Cheers, > Gordon > > > On 17 May 2017 at 4:45:19 AM, Jia Teoh (jiat...@gmail.com) wrote: > > Hi, > > I'm trying to use KafkaConsumer010 as a source for a windowing job on > event time, as provided by Kafka. According to the kafka connector doc ( > link > <https://ci.apache.org/projects/flink/flink-docs-release-1.2/dev/connectors/kafka.html#using-kafka-timestamps-and-flink-event-time-in-kafka-010>), > I've set the time characteristic to event time ( > streamExecutionEnvironment.setStreamTimeCharacteristic( > TimeCharacteristic.EventTime)) and am using KafkaConsumer010 along with > Kafka 0.10.2. I've also set up windowing: "stream.windowAll( > TumblingEventTimeWindows.of(Time.milliseconds(windowMillis)))" (using > timeWindowAll appears to be equivalent as well). > > Using these configurations I can verify that data is read from Kafka. > However, the event time windows never trigger even when data is loaded for > much longer than the window size. Is there an additional configuration I am > missing? > > I have verified that the Kafka messages have timestamps. The docs mention > that there is no need for a timestamp extractor, but using one to > explicitly assign the current time does result in windows being triggered. > > Thanks, > Jia Teoh > >
