> So - firstly, any ideas why the fbserver is dying up there?

There can be a number of reasons. Poorly written UDF libraries (even
non-theadsafe code) can crash the engine, as can corruption within the
database file, as can unexpected file system level locks on the
database file.

If you have inhouse UDFs, that would be a good place to start
investigation. I would also use gfix to validate the database(s). If
the crashes happen at the same time every day, it may be because gbak
is encountering corruption. If you are using xcopy (or any other such
file system level copy tool) to backup, then don't, not while the
server is active.

The 100?? errors are winsock errors that you should be able to google
for an explanation.

> And
> secondly, what can i do about it short of reprogramming all interfaces
> to only make db connection when needed/close off db connection when

idle?

It is throwing a table cloth over the mess, not a real solution, but
if you use the Classic architecture a separate server instance is
created for each connection. So if one connection calls a faulty UDF,
only that connection will crash, not everyone. It may at least help
you isolate the problem.

When you install Classic, make sure to uninstall Guardian.

Adam