Re: BEA and Java Connection Pooling

  • From: Phil Singer <psinger1@xxxxxxxxxxxxx>
  • To: Oracle-L <oracle-l@xxxxxxxxxxxxx>
  • Date: Thu, 20 Oct 2005 22:51:30 -0700

Michael Fontana wrote:

This is, of course, as I am told by our development management, a "DATABASE PROBLEM". Given that Oracle has similar methods to accomplish the same things as the java connection pools are supposed to, does anyone have any suggestions/experience/advice about the use of these techniques?

I've had to deal with similar problems using ADO connection pooling. But , first, let me (re)tell a little story:

A man was walking down the street. He saw another man on his knees under a streetlight. He asked what the problem was. The 2nd man said he had lost his wedding ring. He knelt down to help him look.

After an hour he asked exactly where he was when he lost it. The response was "over there, across the street".

"So, why are you looking over here?"

"Because the light is so much better."

I am convinced that the reason the DATABASE gets blamed (or is made the focus for the troubleshooting) is that Oracle makes it so easy to look at a single connection, that no one wants to look where the problem actually is. Troubleshooting web server logs and network traces are difficult. Querying the data dictionary (and even a sql*net trace) is duck soup.

Having said that, we found that getting on the latest version of Oracle Net services fixed a great many things. And I also think (not that it's worth much) that putting up with these issues is still better than trying to deal with Oracle Net Shared Servers.

Other related posts: