-
1. Re: Funky config needed for extended PersistenceContext repl
bill.burke Feb 17, 2007 5:43 PM (in response to brian.stansberry)Seems like a HEM problem, i'll ping emmanuel.
-
2. Re: Funky config needed for extended PersistenceContext repl
epbernard Feb 19, 2007 9:13 AM (in response to brian.stansberry)The reason why I do not set the name automatically is because the name is an actual JNDI name. What could be a portable name(space) across all JavaEE app servers? Or if I don't have a JNDI available (Tomcat).
I expected the JBoss container to set the property and pass the value to HEM.
WDYT? -
3. Re: Funky config needed for extended PersistenceContext repl
brian.stansberry Mar 22, 2007 3:56 PM (in response to brian.stansberry)Sorry, Emmanuel, I didn't reply before because I really didn't have a clue; my understanding of this just came from following code from the exception stack trace.
But, now QE has stumbled onto this and raised a blocker JIRA (http://jira.jboss.com/jira/browse/JBPAPP-90).
Are you saying hibernate.session_factory_name is the JNDI name? Or the value of the persistence-unit 'name' attribute? -
4. Re: Funky config needed for extended PersistenceContext repl
epbernard Mar 22, 2007 4:01 PM (in response to brian.stansberry)this is the JNDI name
-
5. Re: Funky config needed for extended PersistenceContext repl
brian.stansberry Mar 22, 2007 4:15 PM (in response to brian.stansberry)???
-
6. Re: Funky config needed for extended PersistenceContext repl
epbernard Mar 22, 2007 4:25 PM (in response to brian.stansberry)hibernate.session_factory_name is the JNDI name where the session factory should be put to and read from.
http://jira.jboss.com/jira/browse/EJBTHREE-893 propose to set up a pertinent default (for JBoss EJB3) if the user does not provide one -
7. Re: Funky config needed for extended PersistenceContext repl
brian.stansberry Mar 22, 2007 4:40 PM (in response to brian.stansberry)Great, thanks.
[OT] BTW, if an SFSB has a ref to an XPC and the bean fails over, Hibernate logs a WARN when the bean context is deserialized from the cache:14:36:32,244 WARN [SessionFactoryObjectFactory] Not found: ff808081117aea3701117aea37d40000
If the session factory has a name, then the deserialization continues on and succeeds.
The WARN isn't a huge problem, since it only happens on failover, when a little bit of unusual stuff in the logs isn't terrible. But it's a rather cryptic message that's likely to lead to support questions. -
8. Re: Funky config needed for extended PersistenceContext repl
epbernard Mar 22, 2007 4:49 PM (in response to brian.stansberry)open a JIRA issue for that, It needs to be tracked down