Hey,
> Hi Simon!
> Thanks for your answer. That will probably solve the problem. I can't
> test it right now, because we just discovered a new bug that
> prevents us
> from getting through the process :-(. I have two more questions now:
>
> 1) If the bean is not marked reentrant and I try to call it in the
> described way, shouldn't there be an other error message than just
> TIMEOUT ? Perhaps we could add the hint, that it might be a
> problem with
> reentrant calls.
EJB Spec says we should throw RemoteException in such a case. Due to a
marginal problem we are fixing, we disabled this for now.
> 2) If I mark a bean as reentrant, which effects do I have to expect ?
> Does this also mean the bean could be visited from two or more callers
> at the same time ?
See EJB 1.1 Spec 9.1.12
Bye
Simon
>
> Thanks,
>
> Tobias
>
>
> "Bordet, Simone" wrote:
> >
> > Hey,
> >
> > > Hi!
> > > I wonder if the following thing is allowed by the EJB specs 1.1:
> > >
> > > We have: bean Alice and bean Bob deployed on the container.
> > >
> > > 1. AliceHome.create()
> > > 2. -> AliceBean.ejbCreate(), done
> > > 3. -> AliceBean.ejbPostCreate()
> > > 4. -> BobHome.create(Alice)
> > > 5. -> BobBean.ejbCreate(Alice), done
> > > 6. -> BobBean.ejbPostCreate(Alice)
> > > 7. -> Alice.getSomeAttribut(), done
> > > 8. 6. done
> > > 9. 4. done
> > > 10. 3. done
> > > 11. 1. done
> > >
> > > Is this allowed by the specs ? How SHOULD jBoss react ?
> Right now the
> > > bean hangs directly at step 7 (BobBean is left but Alice is
> > > not reached
> > > !) in my code.
> >
> > This kind of beans (Alice) is called reentrant; jboss
> supports them of
> > course, simply specify in ejb-jar.xml that the bean is
> reentrant (see
> > ejb-jar documentation on where in the xml tree specify
> this), cause the
> > default is that they aren't (and hence your problem).
> > Remember though that use of this kind of beans is
> discouraged by the EJB1.1
> > spec.
> >
> > HTH,
> >
> > Simon
> >
> > >
> > > After waiting a long time I get a TIMEOUT exception (->
> > > ROLLBACK). There
> > > is also a bean mentionend in the trace with som XidImpl stuff that
> > > should not be involved in the proccess at all.
> > >
> > > Any ideas ?
> > >
> > > Tobias
> > >
> > >
>
>
> --
> --------------------------------------------------------------
> To subscribe: [EMAIL PROTECTED]
> To unsubscribe: [EMAIL PROTECTED]
> Problems?: [EMAIL PROTECTED]
>
--
--------------------------------------------------------------
To subscribe: [EMAIL PROTECTED]
To unsubscribe: [EMAIL PROTECTED]
Problems?: [EMAIL PROTECTED]
