[jboss-jira] [JBoss JIRA] Commented: (EJBTHREE-1500) Discrepancies between removing a bean in cache vs bean passivated

[Galder Zamarreno (JIRA)]

    [ https://jira.jboss.org/jira/browse/EJBTHREE-1500?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12431340#action_12431340 ] 

Galder Zamarreno commented on EJBTHREE-1500:
--------------------------------------------

Hmmm, let's clarify all this:

1.- When timed out bean from passivated state is removed, we currently do:
- a) invoke pre destroy method
- b) update removeCount
- c) remove context from cache map

2.- When timed out bean from ready state is removed, we currently do:
- a) remove context from cache map.

Now, taking in account your statement, we should be fixing this:

- 1a) needs to be removed as per 4.4.3.3
- We need to update removeCount for 2.
- Optionally: call pre destroy when bean removed in 2 as per 4.4

> Discrepancies between removing a bean in cache vs bean passivated
> -----------------------------------------------------------------
>
>                 Key: EJBTHREE-1500
>                 URL: https://jira.jboss.org/jira/browse/EJBTHREE-1500
>             Project: EJB 3.0
>          Issue Type: Sub-task
>    Affects Versions: AS 4.2.3.GA, 1.0.0-Beta3
>            Reporter: Galder Zamarreno
>            Assignee: Galder Zamarreno
>
> There's a discrepancy between removing a bean from the cache 
> after a timeout when bean is in cache compared to when bean is 
> passivated:
> Example, when bean is in cache and has to be removed, this is called:
> if (now - centry.lastUsed >= removalTimeout * 1000)
> {
>   synchronized (centry)
>   {                                                                    
>     it.remove();
>   }
> }
> Now, when bean is passivated, this is called:
> if (now - centry.lastUsed >= removalTimeout * 1000)
> {
>   get(centry.getId(), false);
>   remove(centry.getId());
> }   
> And remove() method calls:
>       if(log.isTraceEnabled())
>       {
>          log.trace("Removing context " + key);
>       }
>       StatefulBeanContext ctx = null;
>       synchronized (cacheMap)
>       {
>          ctx = (StatefulBeanContext) cacheMap.get(key);
>       }
>       if(ctx == null)
>          throw new NoSuchEJBException("Could not find Stateful bean: " + key);
>       if (!ctx.isRemoved())
>          container.destroy(ctx);
>       
>       ++removeCount;
>       
>       if (ctx.getCanRemoveFromCache())
>       {
>          synchronized (cacheMap)
>          {
>             cacheMap.remove(key);
>          }
>       }
> What this means is that when a bean in cache is removed:
> 1.- removeCount is not updated
> 2.- container.destroy(ctx); is not executed and hence the @PreDestroy
> method is not executed either. With regards to this, spec says:
> Spec 4.4 "Note that a container can also invoke the PreDestroy method 
> on the instance without a client call to remove the session object after 
> the lifetime of the EJB object has expired."

-- 
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators: https://jira.jboss.org/jira/secure/Administrators.jspa
-
For more information on JIRA, see: http://www.atlassian.com/software/jira