Now I know I've seen some really useful looking documentation on
interbase.log somewhere... wish I could find it now.

I am troubleshooting a strange (so far un-reproducible) problem where
transactions seem to be rolled back unexpectedly, without any exception
being generated. While sniffing around for clues I found these errors in
interbase.log:

MHS_SRV01 (Server) Thu Mar 01 10:38:31 2001
INET/inet_error: read errno = 10054

MHS_SRV01 (Server) Thu Mar 01 10:38:31 2001
INET/inet_error: read errno = 10054

MHS_SRV01 (Server) Thu Mar 01 12:10:52 2001
INET/inet_error: read errno = 10054

MHS_SRV01 (Server) Thu Mar 01 12:10:52 2001
INET/inet_error: read errno = 10054

The times seem to correspond to when my user had problems, which makes me
wonder what the error means.

The site is IB 6.01 running on NT4 sp6 server with Win 2000 clients. The
site runs both LAN and WAN clients but the problem has so far only occurred
on the LAN clients.

My client application uses IBO and IBX components.

I have found a knowledgebase article that lists 10054 as WSAECONNRESET, I
wonder what that means in real life :-)

Windows SDK says:
WSAECONNRESET
(10054)
Connection reset by peer.
A existing connection was forcibly closed by the remote host. This normally
results if the peer application on the remote host is suddenly stopped, the
host is rebooted, or the remote host used a "hard close" (see setsockopt for
more information on the SO_LINGER option on the remote socket.)

But I am still left wondering who that might relate to my application.

Thanks in advance for any insight,

Brian