[
https://issues.apache.org/jira/browse/TWILL-27?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13882142#comment-13882142
]
Terence Yim commented on TWILL-27:
----------------------------------
Seems like the failure is not caused by Kafka upgrade but rather some timing
issue on the test. Do you still have the surefire report file? It can be found
in "twill-yarn/target/surefire-reports/". The .txt file would contains the test
run stdout, which could help figuring out what causing the problem.
In the first test case failure, the test runner sleep for 5 seconds and then
check the number of containers running from the report. It's possible that, if
the machine is slow/busy, it could take more than 5 seconds for events being
happened and detected by AM.
Terence
> Upgrade to Kafka-0.8
> --------------------
>
> Key: TWILL-27
> URL: https://issues.apache.org/jira/browse/TWILL-27
> Project: Apache Twill
> Issue Type: Sub-task
> Reporter: Terence Yim
> Assignee: Terence Yim
> Attachments: TWILL-27-1.patch
>
> Original Estimate: 48h
> Time Spent: 4h
> Remaining Estimate: 1h
>
> Twill currently uses Kafka-0.7 which is not available in public maven repo,
> hence need to package with the binary, which would hinder the Twill release.
--
This message was sent by Atlassian JIRA
(v6.1.5#6160)
