[
https://issues.apache.org/jira/browse/OPENJPA-1824?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Rohit Dilip Kelapure updated OPENJPA-1824:
------------------------------------------
The way OpenBooks build dependencies are setup, one cannot inject an
emf/PersistenceContext without modifying the pom.xml to pull in JEE apis.
Therefore submitting a patch for this issue is difficult as it necessitates
making changes to the build.
Ideally I would like to make ServiceFactory a stateless EJB and Inject a
PersistenceContext in a server JEE environment so that one could get a server
manged emf rather than creating one explicitly through
Persistence.createEntityManagerFactory...
>From Pinaki Poddar .....
the PersistenceService should more be like a delegation than an inheritence
-then you can create a 'business service' and inject a EMF into it easier
-whether that EMF itself is injected by a JEE container or hand-constructed in
JSE
-will be orthogonal to the 'business service'
-that just needs a EMF and somehow expects a session-per-thread model
--- OpenBooks design constraint is that a business serivice must compile
without JEE libs ---
> OpenBooks should used container managed persistence when deployed in an
> application server.
> -------------------------------------------------------------------------------------------
>
> Key: OPENJPA-1824
> URL: https://issues.apache.org/jira/browse/OPENJPA-1824
> Project: OpenJPA
> Issue Type: Bug
> Components: samples
> Affects Versions: 2.0.1
> Environment: Windows Sever 2003
> Reporter: Rohit Dilip Kelapure
> Priority: Minor
>
> OpenBooks is using an app managed persistence context and it isn't cleaning
> up properly. openbook.server.ServiceFactory.getService(...) creates an EMF,
> but it is never closed.
> OpenBooks application is using an app managed persistence context which is
> NOT cleaned up correctly when deployed in a JEE server.
> This results in the JEE container JPA Runtime does NOT calling
> DataCacheManager.close() on the DataCacheManager plugin.
> The container JPA Runtime does NOT call openjpa.DataCacheManager.close() when
> the application is stopped.
> This does not give a chance to any OpenJPA L2 cache provider to cleanup their
> resources and remove cache instances from static hashmaps. utlimately
> resulting in a memory leak.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
