[
https://issues.apache.org/jira/browse/CAMEL-17000?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17435316#comment-17435316
]
Karen Lease commented on CAMEL-17000:
-------------------------------------
I created a PR https://github.com/apache/camel/pull/6350 to upgrade to Jupiter
5.8.1 and set a custom strategy for parallel execution which is used in
camel-core tests only for now. Using this on my environment (Ubuntu 20.04.3,
16Gb RAM, 8 CPU, openjdk 11.0.11) I typically get around 30 flakes but no
errors.
Most of them appear to be related to heavy use of asynchronous parallel
processing in Camel itself. For example in MulticastParallelAllTimeoutAwareTest
it appears that the exchanges are not even sent to the multicast targets before
the timeout on the processing expires.
Typically the failing tests succeed on the second try because surefire reruns
them after the launching all tests the first time, and at this point, there is
much less contention for the processor.
I created CAMEL-17064 to further investigate flake issues.
> Upgrade to junit 5.8.x
> ----------------------
>
> Key: CAMEL-17000
> URL: https://issues.apache.org/jira/browse/CAMEL-17000
> Project: Camel
> Issue Type: Dependency upgrade
> Components: tests
> Reporter: Claus Ibsen
> Assignee: Karen Lease
> Priority: Major
> Fix For: 3.13.0
>
>
> The 5.8.x upgrade causes some tests in core to fail.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
