ailed to lazily initialize a collection - no session or sess

otis · **Joined:** Wed Dec 03, 2003 10:59 am **Posts:** 47

Hibernate version: 2.1.4

Hello,

I have an app where I see the following error every once in a while:

Code:

2004-11-19 16:01:45,359 FATAL [     ajp-3]                                        ErrLog Failed to lazily initialize a collection - no session or session was closed
net.sf.hibernate.LazyInitializationException: Failed to lazily initialize a collection - no session or session was closed
        at net.sf.hibernate.collection.PersistentCollection.initialize(PersistentCollection.java:209)
        at net.sf.hibernate.collection.PersistentCollection.write(PersistentCollection.java:84)
        at net.sf.hibernate.collection.Bag.add(Bag.java:281)

...
...

2004-11-19 16:01:45,360 FATAL [     ajp-3]                                        ErrLog HIBERNATE SESSION: net.sf.hibernate.impl.SessionImpl@126078f
2004-11-19 16:01:45,360 FATAL [     ajp-3]                                        ErrLog HIBERNATE SESSION IS CONNECTED: true

The last part above is from my app - my attempt to see what state the Hibernate Session is in. And this is where I see a conflict. Hibernate is saying that it either has no Session, or that the Session was closed. However, you can see from the above that the Session instance _does_ exist, and it is even connected to the DB (can it be connected but closed somehow?)

This happens so sporadically, so randomly, I don't know how to trace the source of the problem.

Is there something else I can dump or log on the Hibernate level that would let me peek inside Hibernate and see what is going on with the Session in question?

Thanks,
Otis

run4fun · **Joined:** Sat May 29, 2004 2:16 pm **Posts:** 81

try session.update()

anthony · **Posted:** Fri Nov 19, 2004 6:13 pm

Quote:

This happens so sporadically, so randomly, I don't know how to trace the source of the problem.

you are wrong with session management, remind session is not threadsafe, you must use something like threadlocal pattern with an interceptor (filter for example) to _bind_ it to the httpRequest cycle.

i'm pretty sure you're linking the session to a threadlocal and doing nothing else, by chance 2 successives httprequests may be executed in the same thread... but it may not, and in this case, you have your bug.
In production you will have really big problems.

It may also be something else, use your debugger