Hi Rick, should I make a Jira for that?
Cheers, John > -----Original Message----- > From: Rick Curtis [mailto:curti...@gmail.com] > Sent: Thursday, October 10, 2013 2:53 PM > To: users > Subject: Re: Initialization Errors when unsing multiple PersistenceUnis in > persistence.xml > > > Is there a reason that we need to process persistence-unit definitions > which are not requested? > The DynamicEnhancementAgent was added as a feature to assist 'new' users > get up and running with minimal effort. Unfortunately this feature > wasn't/isn't > 100% solid, and this is clearly a bug. In the event that an invalid PU is > encountered, a message should be logged and the error should be tolerated. > > I would advise setting the DynamicEnhancementAgent to false to avoid this > bug, and obviously you're enhancing via another method. > > Thanks, > Rick > > > On Thu, Oct 10, 2013 at 2:33 AM, Boblitz John > <john.bobl...@bertschi.com>wrote: > > > Hello, > > > > I have a rather "simple" system configuration and wanted to simplify > > my persistence configuration by adding persistence-unit defintions to > > the persistence.xml. I basically have one pu definition for each > > environment (dev/tst/prd). > > > > On each system, there are several services which use the same pu > > definition but have their own instance of the pu. These communicate > > with each other via activeMQ to keep the caches in synch. > > > > Until now, I had one PU and in each service/environment the > > persistence configuration was included in a conf file and passed to the emf. > > > > The multiple pu's work just fine, as long as there is no error in ANY > > of them. > > > > However, if ANY one of the pu's has an error in the > > RemoteCommitProvider definition - jpa will not start. > > > > This occurs in the PCEnhancerAgent.registerClassLoadEnhancer which > > appears to loop over each pu and create a OpenJPAConfigurationImpl. > > In the loop the > > instantiateAll() method is invoked on each conf - whether I wanted to > > load that pu or not! - that in tune causes the RemoteCommitProvider to > > be created ... > > > > IF, when starting my development environment, the test environment is > > not online (or reachable) the start will fail. > > > > > > Is there a reason that we need to process persistence-unit definitions > > which are not requested? > > > > Setting DynamicEnhancementAgent to false resolves the problem .... > > > > > > > > John > > ---- > > Who is General Failure, and why is he reading my hard disk? > > > > > > > > > -- > *Rick Curtis*
