> Passivation should definitely not kill a stateful session bean.
> (Does JBoss do this?)
No.
It seems that it was (as you suggested) a problem with serialization of
references that my stateful SB was holding (albeit somewhat indiretly) to
entity beans.
The last comments Burkhard and you made were the missing link that I needed.
Thanks to Michael, yourself, and Burkhard for the help. I greatly appreciate
it.
But, humour me please... two related questions...
Q1
----
When I checked what was happening during passivation of the SFSB I was using
readObject() and writeObject() to serialize the handles of of any referenced
entity beans. A handle was obtained from an EB's remote interfaces via
EJBObject.getHandle(). The fix was to instead serialize both an EB's home
and primary key via EJBObject.getEJBHome() and EJBObject.getPrimaryKey().
Why does this work? I thought that we were supposed to use the handle to
pass EB references around.
What exactly should be saved to properly and portably restore the EB
reference?
And does it make a difference if the reference is being passed to a real
remote client instead of another bean (I would hope not)?
Q2
----
So JBoss is correctly passivating SFSBs once the <max-bean-age> is reached,
and will activate them correctly (if I don't screw it up) when later needed.
How then does JBoss know when to finally kill off a bean (eg. where the
client has died)? Is there a way to configure this?
Thanks again
Mike
===================
Mike Finn
Tactical Executive Systems
[EMAIL PROTECTED]
_______________________________________________
JBoss-user mailing list
[EMAIL PROTECTED]
http://lists.sourceforge.net/lists/listinfo/jboss-user
