Thanks Tom and Stephen for your inputs. The latest sync has brought in the new UUID feature downstream for us, and the script which used to overwrite the feature list in the features.cfg file failed

Karaf has always behaved that way. To get changes to be processed on next startup, you delete the bundle cache under data/cache.

Hi Faseela, Stephen Kitt <skitt@...> wrote: The other issues you’ve been having are caused by the fact that featuresBoot is only processed on first boot, so you can’t start a Karaf

Hi, Faseela K <faseela.k@...> wrote: The UUID-named feature is just a wrapper around the boot features in the distribution. The featuresBoot configuration directive works as previously, with

+odlparent-dev(Hope that is the right platform to ask this). Thanks, Faseela

Hi again, Here's some progress, but no success yet. I created one patch for controller and one for

Thanks for tracking it down. I guess the IT suite needs improvements to its UT :) Regards, Robert

Thank you both, very helpful, I'm tracking down all the paces where I need to update the code. Controller project has the base class for out IT, so I will file a patch for it when I can successfully

We should also review current documentation referencing log4j, for example if we have passages explaining how to increase log level for a specific project or class. I'll take a look and file patches

We should probably expose more this change as it may also impact people using ODL.

+aaa-dev: ==> https://jira.opendaylight.org/browse/AAA-166 ==> https://jira.opendaylight.org/browse/CONTROLLER-1815?

This error [0] is in logs since 17.1.2018 across multiple project There is also [1] in all tests using jolokia. They can be seen for example here [2]. As far as I can tell they don't affect the

Hi Daniel, Daniel Farrell <dfarrell@...> wrote: It wouldn’t — the existence of this plan is one of the reasons we’re getting all our ducks in a row in order to release the “base”

Michael Vorburger <vorburger@...> wrote: The rule is that any breaking change, i.e. any change which *in and of itself* could cause downstreams to break, goes in the next major release, which

Robert Varga <nite@...> wrote: I haven’t been able to reproduce it with a plain Karaf installation though, so it might be related to something we’re doing... Regards, Stephen

Hey Stephen, As we talked about on the Kernel call, remember that we're asking non-SR Managed Projects to have their releases ready for Managed Projects no later than two weeks after the start of a

Here is a patch to csit where we made the updates. Notice also in addition to changing from log4j to log4j2 that the module and level settings have changed. https://git.opendaylight.org/gerrit/67749

Sounds like it should be reported upstream as a critical issues. Regards, Robert

Hi Lori, Lori Jakab <lorand.jakab+odl@...> wrote: This could well be related to the switch from Log4J to Log4J2 as the default logging provider. If you’re

Hi Jamo, Jamo Luhrsen <jluhrsen@...> wrote: Thanks for the investigation! This feels like a Karaf bug... There have been a number of improvements to Karaf’s thread handling in the 4.2