Adam,
Thanks for the reply. First things first, the programs we are using
have been in use (with a number of upgrades and the usual
maintenance) for several years, and the problems we are seeing now
arose only after we upgraded to FB 1.5.3 from 1.0.3 and moved from an
outdated Win2000 server to a much upgraded Win2003 server. We do plan
in the (hopefully near) future to move away from the BDE because of
its numerous problems.

Secondly, the problem is only occurring to individual users - the
Firebird service is running just fine. The applications also run as
expected, unless the user leaves the app open and connected for an
hour. I also had this happen to me today when I had Database
Workbench connected with an active SQL Query window open. When I
closed DBW I got the "connection was forcibly closed" error, which is
exactly what is happening to the clients. If you reopen the app, it
connects and operates just fine.

Thanks,
Scott

--- In firebird-support@yahoogroups.com, "Adam" <s3057043@...> wrote:
>
> Scott,
>
> I will preface this advice by noting that there is no BDE version
that
> is considered stable for connection to any Firebird Server version.
I
> will have to assume that this BDE issue has nothing to do with your
> problem, but obviously that may not be the case, it may well be your
> problem.
>
> The symptoms are consistent with the action when the server process
> has been restarted. This may be because it was explicitly restarted
or
> perhaps it crashed and was restarted by FB Guardian. There are many
> different causes for the engine to crash, mostly it is a result of
it
> panicing because things aren't as they should be. If it encounters
> something it shouldn't, the server will favour a shutdown over
> potentially corrupting the database. The usual culprits are poorly
> coded, incorrectly defined or non threadsafe UDF calls, faulty
> hardware, or corrupt database files. The place to start your
> investigation would be Firebird.log to see why the server process
> shutdown. Hopefully it was able to write the log file to disk before
> it died.
>
> Adam
>