[
https://issues.apache.org/jira/browse/KARAF-3205?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Guillaume Nodet updated KARAF-3205:
-----------------------------------
Description:
The create-dump command is a nice feature in theory.
However, in practice, you may need it while it's not available anymore, because
the needed bundles have been stopped somehow.
I wonder if we should refactor this idea and move it into a more low level such
as in main and have it called using the usual SIGHUP signal.
That does not necessarily mean we need to drop its extensibility via OSGi
services, as it could still discover extensions in it, but it should be very
robust in trying to access the OSGi registry and not expect everyhing to work.
was:
The create-dump command is a nice feature in theory.
However, in practice, you may need it while it's not available anymore, because
the needed bundles have been stopped somehow.
I wonder if we should refactor this idea and move it into a more low level such
as in main and have it called using the usual SIGQUIT signal (which usually
dumps threads to the console).
That does not necessarily mean we need to drop its extensibility via OSGi
services, as it could still discover extensions in it, but it should be very
robust in trying to access the OSGi registry and not expect everyhing to work.
> Refactor create-dump to be more low level
> -----------------------------------------
>
> Key: KARAF-3205
> URL: https://issues.apache.org/jira/browse/KARAF-3205
> Project: Karaf
> Issue Type: New Feature
> Reporter: Guillaume Nodet
> Fix For: 4.0.0
>
>
> The create-dump command is a nice feature in theory.
> However, in practice, you may need it while it's not available anymore,
> because the needed bundles have been stopped somehow.
> I wonder if we should refactor this idea and move it into a more low level
> such as in main and have it called using the usual SIGHUP signal.
> That does not necessarily mean we need to drop its extensibility via OSGi
> services, as it could still discover extensions in it, but it should be very
> robust in trying to access the OSGi registry and not expect everyhing to work.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
