Hi,
I think I've solved it, it was my error, after all. The second session
bean returned a newly created and persisted entity. Because the
interface for the session bean was remote, that instance became
detached. By making the interface a local interface the problem was solved.
Thanks for all the help,
Joerg
PS: I'm not sure, but shouldn't the EntityManager have thrown an
IllegalArgumentException?
Joerg Lauer schrieb:
> Hi,
>
> answers within.
>
> Craig L Russell schrieb:
>> Hi Joerg,
>>
>> To help problem determination here, (without actually sending a
>> complete test case) can you please provide:
>>
>> A description of the artificial home (what's artificial about it?)
>
> It's a stateless session bean implementing the methods offered by the
> home interface of an ejb 1.1 entity bean.
>
>> The methods or technique used for transaction completion
>
> None by me. As far as I understood it, the transaction completed, when
> the business method returns.
>
>> The type of EntityManagerFactory
>
> I don't use any EntityManagerFactory.
>
>> The transaction attributes of the business methods you call
>
> REQUIRED (actually I'm using the default, which should be REQUIRED)
>
>> The way you remove the entity
>
> by calling remove on the EntityManager
>
>> The way you acquire the EntityManager in the business methods
>
> Dependency Injection
>
>>
>> Thanks,
>>
>> Craig
>>
>> On Jul 10, 2006, at 8:33 AM, Joerg Lauer wrote:
>>
>>> Hi,
>>>
>>> while running some performance tests I noticed that some entities
>>> weren't removed, when they were created in the same transaction.
>>>
>>> The entities were persisted, in another business method of another
>>> bean, and afterwards removed. (I was using a fake home interface, for
>>> this test)
>>>
>>> After the transaction finished, the removed entities were inserted
>>> into the database.
>>>
>>> If I skip the artificial home and persist and remove in the same
>>> method (or even call methods on the same bean) everything works okay.
>>>
>>> I'd like to know whether I'm making a fundamental mistake, or it's an
>>> issue in toplink or glassfish.
>>>
>>> I'm currently running Appserver 9, but glassfish v2 Build 08 showed
>>> the same problem, both using toplink and JavaDB.
>>>
>>> Regards
>>>
>>> Joerg Lauer
>>
>> Craig Russell
>> Architect, Sun Java Enterprise System http://java.sun.com/products/jdo
>> 408 276-5638 mailto:Craig.Russell_at_sun.com
>> P.S. A good JDO? O, Gasp!
>>