[ 
https://issues.apache.org/jira/browse/CXF-719?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#action_12503787
 ] 

willem Jiang edited comment on CXF-719 at 6/12/07 2:18 AM:
-----------------------------------------------------------

This issue is introduced by the ManagedEndpoint which will not unregister it 
from the ServerLifeCycleManagerImpl's listeners which is a ArrayList when 
Server.stop() is called.
ServerLifeCycleManagerImpl is referenced by BusImpl, and the the listener holds 
the reference of the ManagedEndpoint.
 ManagedEndpoint also holds the reference of the EndpointImpl, ServerImpl  ***, 
so it cases the memory leak.

And This issue has been resolved in the revision  546431
 


 was:
This issue is introduced by the ManagedEndpoint which will not unregister it 
from the ServerLifeCycleManagerImpl's listeners which is a ArrayList when 
Server.stop() is called.
ServerLifeCycleManagerImpl is referenced by BusImpl, and the the listener holds 
the reference of the ManagedEndpoint.
 ManagedEndpoint also holds the reference of the EndpointImpl, ServerImpl  ***, 
so it cases the memory leak.

And This issue has be resolved in the revision  546431
 

> Endpoint.stop() does not release the resources for this endpoint, which might 
> cause an OutOfMemory error.
> ---------------------------------------------------------------------------------------------------------
>
>                 Key: CXF-719
>                 URL: https://issues.apache.org/jira/browse/CXF-719
>             Project: CXF
>          Issue Type: Bug
>          Components: JAX-WS Runtime
>            Reporter: Jervis Liu
>            Assignee: Jervis Liu
>
> User reported:
> "I am using the the jaxws:endpoint Spring bean which when destroyed calls 
> Endpoint.stop(). This means when  the container removes an application, i.e. 
> stops an endpoint,  the resources for this endpoint will not be cleaned up. A 
>  user may think they are removing an application when in fact they are not, 
> which after a period of time will cause problems. We need to either change 
> the semantics of Endpoint.stop() or add some kind of destroy() method to 
> EndpointImpl.
> I can get out of memory error after as little as 14 times endpoint 
> publishing/stopping."

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.

Reply via email to