[
https://issues.apache.org/jira/browse/KAFKA-5100?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15978863#comment-15978863
]
Jun Rao commented on KAFKA-5100:
--------------------------------
It seems the issue is that we now print the total before closing the producer.
This means that some intermediate reporting result may show up after the total,
which breaks the parsing.
> ProducerPerformanceService failing due to parsing error
> -------------------------------------------------------
>
> Key: KAFKA-5100
> URL: https://issues.apache.org/jira/browse/KAFKA-5100
> Project: Kafka
> Issue Type: Bug
> Reporter: Ismael Juma
> Priority: Blocker
> Fix For: 0.11.0.0
>
>
> 16 tests that use ProducerPerformanceService failed with:
> {'ProducerPerformanceService-0-139930177129936-worker-1': Exception(u'Unable
> to parse aggregate performance statistics on node 1: 150 records sent, 29.3
> records/sec (0.08 MB/sec), 6027.2 ms avg latency, 9028.0 max latency.\n',)}
> https://jenkins.confluent.io/job/system-test-kafka/579/consoleFull
> Logs available via links in the following page:
> http://confluent-kafka-system-test-results.s3-us-west-2.amazonaws.com/2017-04-21--001.1492775811--apache--trunk--f18a14a/report.html
> Looking at recent commits, the following seems like the most likely one:
> https://github.com/apache/kafka/commit/609e9b0b2f46ce72ed91965f7e43c512b26a609b
> cc [~huxi_2b] [~junrao]
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)
