You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@felix.apache.org by "Magnus Jungsbluth (JIRA)" <ji...@apache.org> on 2016/12/01 20:53:58 UTC

[jira] [Created] (FELIX-5438) Allow for parallel calls on ManagedService(Factory)#update

Magnus Jungsbluth created FELIX-5438:
----------------------------------------

             Summary: Allow for parallel calls on ManagedService(Factory)#update
                 Key: FELIX-5438
                 URL: https://issues.apache.org/jira/browse/FELIX-5438
             Project: Felix
          Issue Type: Improvement
          Components: Configuration Admin
    Affects Versions: configadmin-1.8.12
            Reporter: Magnus Jungsbluth


I have been using the ConfigAdmin implementation a while now and I was always under the assumption that ManagedService(Factory)S were called in parallel (with the restriction that per MSF instance updated() cannot be called concurrently). 

However after looking at UpdateThread.java it became clear that ALL updates to ManagedService(Factory)s are sequentially applied. In our system this led to a situation were a timeout of a very early ManagedServiceFactory triggered a cascade of other timeouts that made it very unclear what the root cause was especially cause these parts of the system were completely independent (despite all using ManagedServiceFactoryS).

I also suspect that this can impact startup performance when using MS(F) services. 

Is there any design rationale why this is the case?  



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)