--- In firebird-support@yahoogroups.com, Steve Wiser <steve@s...> wrote:

> From what we
> could tell it was not a problem with the db server or the java server,
> but with our code and how we were cleaning up the db connections in the
> destroy method.
>

Thanks for your suggestions.. I considered that, but what makes
that idea hard to consider is that I have the exact same database
(same OS, same hardware) running the exact same web front ends..

Here is the teste I'm running..

3 days ago I rebooted both machines (problem machine and current
machine) to insure a fresh start. Both are new installs of CentOS 4.2
with the latest Firebird download as of last week.

Problem machine - all the same databases and they are being access
using DatabaseWorkbench.. (Just holding each database open, there are
eight)

Current machine - same databases but they are being access by the
front end web applications... (use dbConnectionBroker as a pooler)

Status Today

Problem Machine - Most of the connections are set at TIME_WAIT with a
few set at ESTABLISHED

Current Machine - All connections are set at ESTABLISHED


This would seem to indicate to me that the web front end connection
code is not at fault.. and that something else hardware related is
causing a problem.

Anyone have another interpretation..???

John..