[
https://issues.apache.org/jira/browse/FLINK-15611?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17019877#comment-17019877
]
Jiangjie Qin commented on FLINK-15611:
--------------------------------------
[~gjy] sorry for the late reply. FLINK-15490 ensures that the Kafka producer
will not cause duplicate messages in Kafka itself during the test setup. But it
does not guarantee that the consumer won't consume the same message twice
without fully restore its state in between, which may cause the exactly once
check to fail.
I am trying check if there is other bug in the test / code that may cause the
problem. In any case, the test itself could be improved a little bit. For
example,
# There is no need to produce the messages in random order, which makes the
debug harder.
# Some additional logging would be helpful, such as what was the last
successful checkpoint and what state was restored to the
{{ValidatingExactlyOnceSink}}.
> KafkaITCase.testOneToOneSources fails on Travis
> -----------------------------------------------
>
> Key: FLINK-15611
> URL: https://issues.apache.org/jira/browse/FLINK-15611
> Project: Flink
> Issue Type: Bug
> Components: Connectors / Kafka, Tests
> Reporter: Yangze Guo
> Assignee: Jiangjie Qin
> Priority: Blocker
> Fix For: 1.10.0
>
>
> {{The test KafkaITCase.testOneToOneSources failed on Travis.}}
> {code:java}
> 03:15:02,019 INFO
> org.apache.flink.streaming.connectors.kafka.KafkaTestEnvironmentImpl -
> Deleting topic scale-down-before-first-checkpoint
> 03:15:02,037 INFO
> org.apache.flink.streaming.connectors.kafka.FlinkKafkaProducerITCase -
> --------------------------------------------------------------------------------
> Test
> testScaleDownBeforeFirstCheckpoint(org.apache.flink.streaming.connectors.kafka.FlinkKafkaProducerITCase)
> successfully run.
> ================================================================================
> 03:15:02,038 INFO org.apache.flink.streaming.connectors.kafka.KafkaTestBase
> - -------------------------------------------------------------------------
> 03:15:02,038 INFO org.apache.flink.streaming.connectors.kafka.KafkaTestBase
> - Shut down KafkaTestBase
> 03:15:02,038 INFO org.apache.flink.streaming.connectors.kafka.KafkaTestBase
> - -------------------------------------------------------------------------
> 03:15:25,728 INFO org.apache.flink.streaming.connectors.kafka.KafkaTestBase
> - -------------------------------------------------------------------------
> 03:15:25,728 INFO org.apache.flink.streaming.connectors.kafka.KafkaTestBase
> - KafkaTestBase finished
> 03:15:25,728 INFO org.apache.flink.streaming.connectors.kafka.KafkaTestBase
> - -------------------------------------------------------------------------
> 03:15:25.731 [INFO] Tests run: 12, Failures: 0, Errors: 0, Skipped: 0, Time
> elapsed: 245.845 s - in
> org.apache.flink.streaming.connectors.kafka.FlinkKafkaProducerITCase
> 03:15:26.099 [INFO]
> 03:15:26.099 [INFO] Results:
> 03:15:26.099 [INFO]
> 03:15:26.099 [ERROR] Failures:
> 03:15:26.099 [ERROR]
> KafkaITCase.testOneToOneSources:97->KafkaConsumerTestBase.runOneToOneExactlyOnceTest:862
> Test failed: Job execution failed.
> {code}
> https://api.travis-ci.com/v3/job/276124537/log.txt
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
