Having the reuse number set to 1 also makes it a pretty unfair performance comparison between the new code I sent and the original JMeter code. My new code will currently always reuse the connections, so it's like having the reuse number set to infinity. Since it looks like the new code is going to work okay, I'll have to do some additional investigation to try to incorporate the reuse number or have some other way to do more fine tuning of the pool. Having a reuse count doesn't really fit too well with the new model, although it could probably be done. On the other hand, I think this new code is probably closer to how most J2EE DataSource implementations do connection pooling -- reusing a connection unless it times out or fails in some odd way.
Anyway, thanks for the input. Please let us know if you learn anything else, or if you encounter any problems with the new code. I'll do some more cleanup and investigate the above issues some more and then commit it to CVS if/when I get something workable.
Jeremy
[EMAIL PROTECTED] wrote:
Did you try upping the reuse number? Reuse of 1 defeats the whole point of having a database pool.
--------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
