You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@flink.apache.org by "Shengkai Fang (Jira)" <ji...@apache.org> on 2020/11/09 09:12:00 UTC

[jira] [Created] (FLINK-20058) Improve kafka per-partition-watermark tests

Shengkai Fang created FLINK-20058:
-------------------------------------

             Summary: Improve kafka per-partition-watermark tests
                 Key: FLINK-20058
                 URL: https://issues.apache.org/jira/browse/FLINK-20058
             Project: Flink
          Issue Type: Improvement
          Components: Connectors / Kafka, Table SQL / API
    Affects Versions: 1.12.0
            Reporter: Shengkai Fang
             Fix For: 1.12.0


Currenlty, we have already pushed watermark into the kafka table source. If the consumer doesn't get data from each partition, it will not output a valid watermark. In the test {{KafkaTableITCase#testPerPartitionWatermark}} the test sink will drop the record whose rowtime is late than watermark. If the consumer reads the partition where number of records in the partition is 1, the sink will receive the valid watermark when all data has been emitted. Therefore, we'd better to keep the data number in all partitions are more than 1.

 

Considering, we have already pushed watermark into the source, the issue FLINK-19878 has been solved.



--
This message was sent by Atlassian Jira
(v8.3.4#803005)