Skip to main content
learner_monk
New Participant
April 19, 2018

Duplicate OSGi configs with every installation

  • April 19, 2018
  • 4 replies
  • 5456 views

Hi All,

We are using few OSGi factory configurations in our authoring environment (AEM 6.2) like below -

Apache Sling Thread Pool Configuration

Apache Sling Job Queue Configuration

(these are factory configs) and, we have checked these in our source control system as below -

/apps/myProject/config.author/org.apache.sling.event.jobs.QueueConfiguration-abc.xml

/apps/myProject/config.author/org.apache.sling.event.jobs.QueueConfiguration-def.xml

What I am noticing that with each code deployment a set of duplicate OSGi configs start appearing from Felix console like below with different unique id but same name -

org.apache.sling.event.jobs.QueueConfiguration.<uniqueid>

org.apache.sling.event.jobs.QueueConfiguration.<uniqueid>

Please suggest.

This post is no longer active and is closed to new replies. Need help? Start a new post to ask your question.

4 replies

AnkurAhlawat-1
New Participant
November 8, 2019

I checked and the resolution for mentioned issue is provided here. http://aemcq5tutorials.com/tutorials/duplicated-osgi-configurations-upgrade/

joerghoh
Employee
November 10, 2018

If it's something like described in OSGI configuration entries duplicated on boot , I would recommend you to raise a Daycare ticket.

New Participant
November 13, 2018

I checked, and the issue we are having is related to that problem. (A more direct link: [SLING-6313] Sling installer can duplicate factory configurations if an invalid RegisteredResourceList.ser is found - AS… )  Looks like our root cause is AEM crashing on issues with the application code or infrastructure, which results in an unclean shutdown of AEM.  In our Dev environments, this is a common scenario.  Production is stable and has not seen the problem with the duplicate OSGI configs.  I will open a Daycare ticket to see if there is any resolution to this.

Thanks!

New Participant
November 10, 2018

We have this same problem, but not at every deployment.  It's only happening sometimes, and I can't find the reason or cause.  I too would like to know how to prevent it.

BTW, I created a quick command line to clean up the bad OSGI config if you have many to delete.  Copy the UIDS from OSGI console and clean them up in a text editor to look something like this:

com.corp.s3.authentication.impl.S3AuthenticationConfigurationImpl.0bd45398-bef0-43cc-9b4b-f3dff57b8db9

com.corp.s3.authentication.impl.S3AuthenticationConfigurationImpl.65bfa9ee-88e2-47b0-8a00-97864b279e33

com.corp.s3.authentication.impl.S3AuthenticationConfigurationImpl.6fc2aa28-410d-46d9-bd16-ba60688957d7

com.corp.s3.authentication.impl.S3AuthenticationConfigurationImpl.7abd8492-9f97-487c-ad70-4f302e9e32d2

com.corp.s3.authentication.impl.S3AuthenticationConfigurationImpl.7ff34f09-123f-4f18-9b60-dddd93e8692a

com.corp.s3.authentication.impl.S3AuthenticationConfigurationImpl.cf407026-692c-4b91-a334-2672543dc363

com.corp.s3.authentication.impl.S3AuthenticationConfigurationImpl.f021dcc8-3622-4548-8f32-9e2532269367

com.corp.s3.authentication.impl.S3AuthenticationConfigurationImpl.f53f4c42-3612-473e-a5ec-b8e194bd38b2

Save that to a text file, then run this command:

for osgiconfig in $(cat bad-osgi.txt); do curl -X POST -d "apply=1&delete=1" -u 'admin:password' "http://localhost:4503/system/console/configMgr/${osgiconfig}"; done

You should get something like this as output:

{ "status": true }{ "status": true }{ "status": true }{ "status": true }{ "status": true }{ "status": true }{ "status": true }{ "status": true }

Then reinstall your configs.