Do you know if it is "safe" to have a stateful session cache the connection
and session object if it is only invoked over and over my a single client
thread? Or do I have to create a new session for each method call?
--jason
On Tue, 7 Aug 2001 [EMAIL PROTECTED] wrote:
> On 7 Aug, Jason Dillon wrote:
> > A minor note on top of this, I changed the pool to not limit the maximum
> > size, then ran a large test against it and ended up with:
> >
> > 2001-08-06 19:24:55,475 JmsXA [Thread-11] INFO - Pool
> > org.jboss.jms.ra.JmsManagedConnectionFactory-1 [999/999/Unlimited] gave
> > out new object: org.jboss.jms.ra.JmsManagedConnection@389071
>
> Well it sounds as if the JmsManagedConnection (or the handle to it
> JmsSession) is not closed properly or upon close is not properly
> returned to the pool.
>
> As far as I remember the connections was properly returned when I wrote
> the stuff, but I may have missed something. I can see 3 possible places
> where there are errors:
>
> 1. Application code (the JMS sessions are not closed properly)
> 2. RA code (the close on the session is not sent to the
> ConnectionEventListener.
> 3. The JBoss pool stuff (it does not properly return the connection
> (JmsManagedConnection) to the pool when it have recived a connection
> close event.
>
> I may ofcourse have something to do with the transaction handling, but
> what I am really not shure of.
>
>
> //Peter
> >
> > 999 ... is a lot of managed connections. Perhaps I am missing something on
> > the application side. I will look into this some more, but if you have any
> > insight it would be helpful.
> >
> > --jason
> >
> > On Mon, 6 Aug 2001, Jason Dillon wrote:
> >> Where does a JMS RA managed connection get returned? I just ran into:
> >>
> >> JmsXA [RMI TCP Connection(12)-192.168.0.124] INFO - Pool
> >> org.jboss.jms.ra.JmsManagedConnectionFactory-1 is full (10/10)!
> >>
> >> JmsXA [RMI TCP Connection(12)-192.168.0.124] INFO - Pool
> >> org.jboss.jms.ra.JmsManagedConnectionFactory-1 [10/10/10] waiting for a
> >> free object
> >>
> >> Which I have never seen before. I looked over the ra classes briefly but I
> >> can't seem to find out where they are returned. After a while the
> >> transaction timed out and now the process is hung waiting for a connection.
> >>
> >> On a separate yet related note should/can the container signal the thread
> >> running in a transaction that it has timed out... or just leave it
> >> potentially locked forever?
> >>
> >> --jason
> >>
> >>
> >> _______________________________________________
> >> Jboss-development mailing list
> >> [EMAIL PROTECTED]
> >> http://lists.sourceforge.net/lists/listinfo/jboss-development
> >>
> >
>
>
_______________________________________________
Jboss-development mailing list
[EMAIL PROTECTED]
http://lists.sourceforge.net/lists/listinfo/jboss-development
