[
https://issues.apache.org/jira/browse/ACE-79?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12989630#comment-12989630
]
Bram de Kruijff commented on ACE-79:
------------------------------------
Sounds a lot like the issues we have over at amdatu
(http://jira.amdatu.org/jira/browse/AMDATU-262). What we know
(http://jira.codehaus.org/browse/SUREFIRE-524) is that surefire does not kill
forked processes (which IMHO is reasonable) so its up to the testrunner. It
seems hudson even has some kill mechanism in place which makes me wonder.... At
Amdatu we now have a "kill all pax processes"cgi in place as a poor workaround.
The real problem seems to be that the forked process either pax exam or the
osgi container and in the end the jvm is not shutting down properly after test.
I thought things might be deadlocking when shutting down during startup (eg.
https://issues.apache.org/jira/browse/FELIX-1816) but I'm not sure. In any case
I can get the exact same two buildfailures using Equinox 3.6.0 as a container.
> Port all integration tests over to the new structure
> ----------------------------------------------------
>
> Key: ACE-79
> URL: https://issues.apache.org/jira/browse/ACE-79
> Project: Ace
> Issue Type: Sub-task
> Reporter: Marcel Offermans
> Assignee: Angelo van der Sijpt
> Attachments: ACE-79-mvnpaxconfig.patch
>
>
--
This message is automatically generated by JIRA.
-
For more information on JIRA, see: http://www.atlassian.com/software/jira
