[jboss-user] [EJB 3.0] - Re: problems with optimistic locking/persistence data in EJB

[wiggy]

wayne 

thats not all complete problem - 

as the remote debugger reported problems with the transaction failure when remote debugging into the server, i rewrote the removeLinks method in the local application client using an application scoped EM.

that way i could run the debugger in local mode.


  | 	public static boolean removeLink (Node fromNode, Node remoteNode)
  | 	{
  | 		EntityManager em;
  | 		EntityManagerFactory emf ;
  | 		List<Link> res;
  | 		boolean result = false;
  | 
  | 		emf = Persistence.createEntityManagerFactory("embedDS");
  | 
  | 		em = emf.createEntityManager();
  | 
  | 		try
  | 		{
  | 		EntityTransaction t = em.getTransaction(); 
  | 
  | 		t.begin();
  | 		fromNode = em.merge(fromNode);
  | 		remoteNode = em.merge(remoteNode);
  | 		Query q = em.createNamedQuery("findLinksBetweenNodes");
  | 		q.setParameter("toNode", remoteNode);
  | 		q.setParameter("fromNode", fromNode);
  | 				
  | 		res = (List<Link>)q.getResultList();
  | 		if (res != null && res.size () == 1)
  | 		{
  | 			Link link =  res.get(0);
  | 			remoteNode.deleteLinkFrom(link);
  | 			fromNode.deleteLinkTo(link);
  | 			em.remove(link);
  | 			result = true;
  | 		}
  | 		else
  | 			{result = false;}
  | 
  | 		t.commit();
  | 		return result;
  | 		} 
  | 		finally 
  | 		{
  | 			em.close();
  | 			emf.close();
  | 		}
  | 
  | 	}
  | 


this essentally throws the same optimistic lock error - so its something to do with the transaction that this lives in updates the version for the attached elements.  when this completes the next step in the main application path 



  | 
  | 			//local if (removeLink(uNode, zNode))
  | 			if (nodeEAO.removeLink(uNode, zNode))
  | 			{
  | 
  | 				System.out.println ("deleted link  : \n");
  | 			}
  | 
  | 			//update to latest version number
  | 			yNode = nodeEAO.refresh(uNode);
  | 

the code errors with optimistic lock when i try and do an em.merge within the nodeEAO.refresh routine.

what this kind of implies i think is that i have to issue another "find" operation to reread a fresh copy of the entity back to avoid this rather than being able to "merge" the starting uNode back into the persistence context.

Is this the only/correct way to proceed - or am i not understanding properly and theres a easy 'correct' fix that would allow me to avoid the exception (I havnt tried to take the version annonation of yet - and see if it handles it without the optimistic lock force enabled.

 

View the original post : http://www.jboss.com/index.html?module=bb&op=viewtopic&p=4119816#4119816

Reply to the post : http://www.jboss.com/index.html?module=bb&op=posting&mode=reply&p=4119816