Reading data in the activate() method of your OSGI service is making your service unflexible. Whenever the content read in activate() changes, you have to restart the service/bundle/complete instance to make this updated content getting effective.
I blogged about it: Pattern: Configuration of OSGI services | Things on a content management system
Excellent community post.
Great post Jorg!! Good to know!!
Views
Replies
Total Likes
Awesome post Jörg Hoh . Have faced same situation once .
Views
Replies
Total Likes
What is your recommendation in reading configurations and assigning it in Activate method. I know it won't be a better solution , but just trying to understand. I saw the changes picked up in my local every time I add a new value to config node, but worried about PROD server.
Views
Replies
Total Likes
Do not mix that up with reading OSGI configuration, which is persisted in /apps/<myapp>/config/pid.of.my.service.xml; that's a different story and technically you are not reading it from the repository, but you are reading it from OSGI; and everything behind that is someone elses business :-)
I just wanted to say, do not use something like that:
because at that point, nothing happens when content at somepath changes.