[jboss-jira] [JBoss JIRA] (WFCORE-553) ModelControllerClientOperationHandler doesn't manage the clientRequestExecutor properly

[Brian Stansberry (JIRA)]

    [ https://issues.jboss.org/browse/WFCORE-553?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13041623#comment-13041623 ] 

Brian Stansberry commented on WFCORE-553:
-----------------------------------------

The WFLY-3225 fix is what introduced WFCORE-553. Prior to that fix the pool was scoped to the factory service, not the handler.

> ModelControllerClientOperationHandler doesn't manage the clientRequestExecutor properly 
> ----------------------------------------------------------------------------------------
>
>                 Key: WFCORE-553
>                 URL: https://issues.jboss.org/browse/WFCORE-553
>             Project: WildFly Core
>          Issue Type: Bug
>          Components: Domain Management
>    Affects Versions: 1.0.0.Alpha1, 1.0.0.Alpha18
>            Reporter: Brian Stansberry
>
> ModelControllerClientOperationHandler's constructor creates a ThreadPoolExecutor for handling client requests and then the class doesn't clean it up.
> In addition, an instance of ModelControllerClientOperationHandler is created per channel, not one per ModelControllerClientOperationHandlerFactoryService. I know I at least thought of the thread pool as being per remote management interface, not per channel.
> Making it be per ModelControllerClientOperationHandlerFactoryService and cleaning it up in that service's stop would be the easiest fix, but the pool settings may not be appropriate if we do that, so tread carefully.



--
This message was sent by Atlassian JIRA
(v6.3.11#6341)