I can't argue with the idea that Aaron might possibly have logged things a
bit heavily, however my banktest completed successfully after 2860.064
seconds. What form did this deadlock take? Is running 50 threads through
10 connections with ~10,000 transactions a reasonable unit test?
david jencks
On 2001.08.16 21:57:21 -0400 Scott M Stark wrote:
> Start with the fact that the bank unit test deadlocks and the server is
> spewing tons
> of messages at info level to the console:
>
> [DefaultDS] Pool
> org.jboss.pool.connector.jdbc.JDBCManagedConnectionFactory-1 is full
> (10/10)!
> [DefaultDS] Pool
> org.jboss.pool.connector.jdbc.JDBCManagedConnectionFactory-1 is full
> (10/10)!
> [DefaultDS] Pool
> org.jboss.pool.connector.jdbc.JDBCManagedConnectionFactory-1 [10/10/10]
> waiting for a free object
> [DefaultDS] Pool
> org.jboss.pool.connector.jdbc.JDBCManagedConnectionFactory-1 [10/10/10]
> waiting for a free object
> [DefaultDS] Pool
> org.jboss.pool.connector.jdbc.JDBCManagedConnectionFactory-1 [9/10/10]
> returned object org.jboss.pool.connector.jdbc.JDBCManagedConnection@7124
> af to the pool.
> [DefaultDS] Pool
> org.jboss.pool.connector.jdbc.JDBCManagedConnectionFactory-1 [8/10/10]
> gave
> out pooled object: org.jboss.pool.connector.jdbc.JDBCManagedConnect
> ion@7124af
> [DefaultDS] Pool
> org.jboss.pool.connector.jdbc.JDBCManagedConnectionFactory-1 [8/10/10]
> returned object org.jboss.pool.connector.jdbc.JDBCManagedConnection@44cb
> be to the pool.
> [Default] local transaction committed
>
> Now its not clear if the test is failing because it is taking to long to
> run
> due this absurd level
> of logging or there is another problem. David, look at the bank test and
> fix
> the logging.
>
> ----- Original Message -----
> From: "Jason Dillon" <[EMAIL PROTECTED]>
> To: <[EMAIL PROTECTED]>
> Sent: Thursday, August 16, 2001 6:01 PM
> Subject: [JBoss-dev] jca changes & loaders
>
>
> > It looks like the recent jca & datasource loaders have introduced
> > inconsistency in the transaction integrety of the entity system.
> >
> > I have not yet been able to track this down, but based on an
> application
> > that was working the day before yesterday, which is now non-functional
> after
> > more than one run, I would guess that there was a change between now
> and
> > then that has broken something.
> >
> > I am seeing problems commiting transactions that invole entity beans.
> Where
> > a client calls a method, which updates a field, sets a modified flag.
> > Sometimes the isModified() method will be called, sometimes it won't.
> There
> > are a few different beans getting called in a specific order, but I am
> not
> > seeing some of the latter JDBCCommand* logs showing that the database
> is
> > actually getting updated.
> >
> > My application has not changed and will appear to function with small
> > message quantities (1) the first time. After running it again it will
> not
> > work, even though the functionality is performed, the database just
> never
> > gets up dated and thus my clients are left in the dark (pooling,
> waiting
> for
> > the dogs...).
> >
> > I think it would be a really good idea if we could stabale and branch a
> 2.6
> > before any other changes are added. 2.5 (or pre-3.0 as it is called)
> is
> > much, much better than 2.4... if we would stop breaking things trying
> to
> > make it even better.
> >
> > Personally I would like to see this happen... so I could stop chasing
> > around bugs that break my app and actually start working on other JBoss
> > bits (like documentation of the build system, integration of the
> testsuite
> > and other such things).
> >
> > JBossMQ has come leaps and bounds, so has the locking system. Lets
> tidy
> up
> > what we have then branch and keep going for full RH in 3.0!
> >
> > It is nice to troll through the code, but I ernestly just want
> something
> > working. I have been playing "chase the bug" for over a month now. It
> is
> > getting closer...
> >
> > If someone knows what is wrong with the tx stuff please let me know.
> My
> > guess is that David might have a clue =P Or I guess I could have
> broken
> > something, but I don't think I did.
> >
> > =|
> >
> > --jason
> >
> >
> > _______________________________________________
> > Jboss-development mailing list
> > [EMAIL PROTECTED]
> > http://lists.sourceforge.net/lists/listinfo/jboss-development
> >
>
>
> _______________________________________________
> Jboss-development mailing list
> [EMAIL PROTECTED]
> http://lists.sourceforge.net/lists/listinfo/jboss-development
>
>
_______________________________________________
Jboss-development mailing list
[EMAIL PROTECTED]
http://lists.sourceforge.net/lists/listinfo/jboss-development
