[
https://issues.apache.org/jira/browse/KARAF-5528?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16283229#comment-16283229
]
ASF subversion and git services commented on KARAF-5528:
--------------------------------------------------------
Commit 2da30f975f1403a324d43db219ccc2dd585a2956 in karaf's branch
refs/heads/master from [~gzres]
[ https://gitbox.apache.org/repos/asf?p=karaf.git;h=2da30f9 ]
Revert "[KARAF-5528] Stop refreshed bundles together with updated and deleted
ones"
This reverts commit ee2b0933fdb5a0c86ac183d38d99025e32569678.
> Karaf feature deployer should stop refreshed bundles together with the
> updated ones
> -----------------------------------------------------------------------------------
>
> Key: KARAF-5528
> URL: https://issues.apache.org/jira/browse/KARAF-5528
> Project: Karaf
> Issue Type: Bug
> Components: karaf-feature
> Affects Versions: 4.0.10, 4.1.3, 4.2.0.M1
> Reporter: Grzegorz Grzybek
> Assignee: Grzegorz Grzybek
> Priority: Critical
> Fix For: 4.0.11, 4.1.4, 4.2.0
>
>
> I had a case when features service was provisioned and there were two wired
> bundles:
> * bundle A which was updated
> * bundle B which had camel context defined in blueprint, depending on OSGi
> service from bundle A.
> Stopping bundle with came context defined in blueprint, requires access to
> all {{<reference>}}-d services, otherwise blueprint proxies may wait for
> specific timeout for the services. Having proper start levels set, everything
> works fine.
> But when during deployment, bundle A was updated, it was stopped first. Then,
> bundle A and B were properly stopped during refresh, but bundle A was already
> stopped (and updated). bundle B hanged during stop.
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)
