Ah, there should definately be an option to turn that bloody feature off. THe important part of the preparedstatement happens on the server anyway. Oracle does its own caching of the query plans (or whatever) involved for your PreparedStatement, as does any other database where a prepared statement gives you any actual advantage. The only thing this actually pools are the java objects involved, and with modern garbage collectors that doesn't buy much.
-danch Jeffrey Wescott wrote: > In the newly-released JBoss 2.4.5, a patch was applied (#532376) to fix bug > #478882. The patch changes the behavior of the prepared statement pool > (which, I believe, is used by anyone using the XA wrapper DataSourceClass) > such that when close() is called on the prepared statement, it propogates > that close() call to the underlying implementation object. > > This patch, unfortunately, has broken things for people using JBoss / Castor / > Oracle. > > Below is my message to the castor-dev mailing list on the topic. > > ------------------------------------->8 cut here > Here's the idea: > In 2.4.4: > 1- Castor obtains a PREPARED STATEMENT from the JBoss PREPARED STATEMENT pool. > 2- Castor does work. > 3- Castor calls STMT.close() to close the prepared statement. > 4- JBoss returns the PREPARED STATEMENT to pool. > > In 2.4.5: > 1- Castor obtains a PREPARED STATEMENT from the JBoss PREPARED STATEMENT pool. > 2- Castor does work. > 3- Castor calls STMT.close() to close the prepared statement. > 4- JBoss returns the PREPARED STATEMENT to pool and closes it. > > I investigated this further and commenting out the lines in the close() method > of the PreparedStatementInPool class make it work as it did in 2.4.4. > > if( impl != null ) > impl.close(); > > It looks like this patch (#532376) was committed to fix bug #478882, but it > seems to have caused another problem. > > For what it's worth, I'm using Oracle 8.x with classes12.zip. Any idea > whether or not it does its own prepared statement pooling? Also, how do I > turn OFF the pooling in JBoss to try to get this stuff to work WITHOUT > changing code? > ------------------------------------->8 cut here > > My questions are: > > 1- Was there any testing done before this patch was applied? Though this > fixes something for people using MSSQL, it breaks something for people using > Oracle. > > 2- Conceptually, is it the correct thing to do? Previously, the prepared > statement pool just returned the object to the pool upon a call to close(). > Now it actually closes the underlying prepared statement. What good is > pooling a closed prepared statement? > > 3- As an Oracle user, is there some other way to get stuff working without > using the XA wrapper DataSourceClass? I've tried lots of combinations or the > OracleXADataSource class and Oracle Xid stuff, but none with any success. > Others in the JBoss forums seem to share my woes in this regard. > > ++jeff > > P.S. -- At this point, it is not an option for me to use another technology > besides Castor / Oracle. > > > _______________________________________________________________ > > Have big pipes? SourceForge.net is looking for download mirrors. We supply > the hardware. You get the recognition. Email Us: [EMAIL PROTECTED] > _______________________________________________ > Jboss-development mailing list > [EMAIL PROTECTED] > https://lists.sourceforge.net/lists/listinfo/jboss-development > _______________________________________________________________ Have big pipes? SourceForge.net is looking for download mirrors. We supply the hardware. You get the recognition. Email Us: [EMAIL PROTECTED] _______________________________________________ Jboss-development mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/jboss-development
