Here is the stack trace. It happens on tomcat shutdown.
Exception in thread "Thread-53" java.lang.NullPointerException
at org.apache.openejb.OpenEJB.destroy(OpenEJB.java:263)
at
org.apache.openejb.tomcat.catalina.TomcatLoader.destroy(TomcatLoader.java:187)
at
org.apache.openejb.tomcat.catalina.TomcatLoader$1.lifecycleEvent(TomcatLoader.java:160)
at
org.apache.catalina.util.LifecycleSupport.fireLifecycleEvent(LifecycleSupport.java:117)
at org.apache.catalina.core.StandardServer.stop(StandardServer.java:748)
at org.apache.catalina.startup.Catalina.stop(Catalina.java:628)
at
org.apache.catalina.startup.Catalina$CatalinaShutdownHook.run(Catalina.java:671)
Luis Fernando Planella Gonzalez
Em Segunda 06 Abril 2009, às 18:14:42, David Blevins escreveu:
>
> On Apr 6, 2009, at 12:55 PM, Luis Fernando Planella Gonzalez wrote:
>
> > I don't know if it's related, but I ALWAYS get a
> > NullPointerException in
> > OpenEJB.destroy(), line 263, which have the following:
> > Assembler assembler =
> > SystemInstance.get().getComponent(Assembler.class);
> > assembler.destroy(); //HERE: assembler is null
> >
> > I use openejb.war integrated in Tomcat.
> > Perhaps adding a try / catch or null test would be great.
> > Since we're in development and months away from production, it's not
> > an issue
> > for now, but one day it will be...
>
> If you can post a full stack trace, we can definitely take a look at
> it. Also, when are you getting this? (Tomcat shutdown, restart, or
> app shutdown, restart, etc.)
>
> -David
>
>
> >
> >
> > Luis Fernando Planella Gonzalez
> >
> >
> > Em Segunda 06 Abril 2009, às 16:26:17, David Blevins escreveu:
> >> On Apr 4, 2009, at 5:12 AM, [email protected] wrote:
> >>> Hi again!
> >>> Finally I have had some time to patch OpenEJB according to my
> >>> previous posting.
> >>> It turns out that the NullPointerException STILL occurs, despite the
> >>> constructor never being invoked with a null value parameter.
> >>
> >> That's crazy. If there is some way to get Tomcat to not attempt to
> >> aggressively clean references that would be by far the best option.
> >> Clearly the VM has already cleared the references and Tomcat is
> >> coming
> >> in afterwards and making a bit of a mess.
> >>
> >>> I also
> >>> tried to patch the catch block to return null.
> >>> Apart from that, it seems like the container (Tomcat) is left in an
> >>> inconsistent state after I have removed the application, requiring a
> >>> restart.
> >>> I will try to find alternative ways to use OpenEJB in my
> >>> application,
> >>> since it seems like the patch I suggested earlier do not really
> >>> solve
> >>> the problems I have had.
> >>
> >> The standard way of deploying the OpenEJB war is guaranteed to work.
> >> I don't recall what about that didn't work for you but maybe there's
> >> some thing we can offer in that regard.
> >>
> >> -David
> >
>
>
