|-----Original Message-----
|From: [EMAIL PROTECTED]
|[mailto:[EMAIL PROTECTED]]On Behalf Of Jason Dillon
|Sent: Friday, December 01, 2000 3:49 PM
|To: [EMAIL PROTECTED]
|Subject: [jBoss-User] Re: Entities, Transactions & Commits (Oh my...
|HELP!)
|
|
|Hello again.  I ran an even simplier example, with one bean.  I tried to
|create it using the minerva datasource and the record 'never' got inserted.
|All of the beans methods were marked as 'Supports' transactions.  I changed
|them to 'Required' and it works fine.
|
|Is there something wrong with 'Supports'?
|

supports means that if there is a transaction it will take it if not nothing
is there.

If you have no transaction then nothing is commited.

The moment you declare "required" it works.  Kosher.

Hey, is it just me but I thought "supports" wasn't a valid tag for entities
since you need to work in CMT with transaction by definition? Can one of the
bible freaks set me straight?  If so isnt' this in the deployer verifier?

marc

|--jason
|
|
|On Thu, 30 Nov 2000, Jason Dillon wrote:
|
|> Hello, I have been banging my head for three days now trying to
|deploy our
|> application (which runs just fine under WebLogic 5.1) in jBoss v2.0.  It
|> seems that most of my problems are database related.  My
|entities don't get
|> persisted properly.
|>
|> In the basic scenario I have three beans (lets call them A, B & C)
|>
|>   A is a stateless session bean, which creates B
|>   B is an entity bean
|>   C is an entity bean created after A (in the same TX), and has
|a foriegn key
|>     constraint to B.id (which is a java.lang.Long).
|>
|> I originally was using the 'org.jboss.minerva.xa.XADataSourceImpl' data
|> source, which worked well for reading data, but failed to properly write
|> data.  With this configuration an oracle exception is throw:
|>
|>   javax.ejb.CreateException: Could not create
|entity:java.sql.SQLException:
|>   ORA-02291: integrity constraint (PLEXUS.FK_B_ID) violated -
|parent key not
|>   found
|>
|> This looks like there was never an insert into the B table and
|thus the FK
|> constraint on C.b_id fails.  Interesting though If I change the
|transaction
|> isolation on A.createBandC to 'Supports', the same thing happends, which
|> looks like nothing is every being commited.  I also had a sql
|prompt opened,
|> so I could check if any records were inserted and found nothing (with the
|> server running and after killing it).
|>
|> I tried changing the data-source to
|> 'oracle.jdbc.xa.client.OracleXADataSource', but with this configuration I
|> get:
|>
|>   java.sql.SQLException: ORA-00060: deadlock detected while waiting for
|>   resource
|>
|> In my sql prompt, rows would not actually show up until I ctrl-c'd the
|> server.
|>
|> Both entities use the 'Standard CMP EntityBean' configuration, which has
|> been modified in 'conf/default/standardjboss.xml':
|>
|>   Optimized: False
|>   Commit-option: C (which is a shared db I believe)
|>
|> The session is a 'Standard Stateless SessionBean' and that
|configuration has
|> not been modified.
|>
|> The assembly descriptor for A & B looks like this:
|>
|> ----8<----
|>
|>  <assembly-descriptor>
|>
|>     <!-- Methods with 'Supports' TIO -->
|>     <container-transaction>
|>       <method>
|>      <ejb-name>A</ejb-name>
|>      <method-name>*</method-name>
|>       </method>
|>
|>       <method>
|>      <ejb-name>B</ejb-name>
|>      <method-name>*</method-name>
|>       </method>
|>
|>       <trans-attribute>Supports</trans-attribute>
|>     </container-transaction>
|>
|>     <!-- Methods with 'Required' TIO -->
|>     <container-transaction>
|>       <method>
|>      <ejb-name>A</ejb-name>
|>      <method-name>createBandC</method-name>
|>       </method>
|>
|>       <trans-attribute>Required</trans-attribute>
|>     </container-transaction>
|>   </assembly-descriptor>
|>
|> ---->8----
|>
|> So the call to A.createBandC(), should have a transaction
|created if there
|> isn't already one.  In my case there isn't.
|>
|> When I run the 'same' code under WebLogic (with the jndi context
|changed via
|> a system property) every thing runs fine.
|>
|> Does anyone have any clues as to why jBoss would behave like this?
|>
|> --jason
|>
|>
|
|
|
|--
|--------------------------------------------------------------
|To subscribe:        [EMAIL PROTECTED]
|To unsubscribe:      [EMAIL PROTECTED]
|Problems?:           [EMAIL PROTECTED]
|
|



--
--------------------------------------------------------------
To subscribe:        [EMAIL PROTECTED]
To unsubscribe:      [EMAIL PROTECTED]
Problems?:           [EMAIL PROTECTED]

Reply via email to