I am attempting to backup an OS/2 client over a 56k WAN.  I receive
an error on the screen that looks like:

date time filename .. Sent
date time filename ......................................Sent

SYS 1811:
The process has stopped. The software diagnostics code (trap number)
is 000D.

I'm not an OS/2 expert, but I was told this message is "common"
for OS/2.

The dsmerror.log has the following entries.

07/18/1996 22:37:28  TcpOpen: TCP/IP error connecting to server.
07/18/1996 22:37:28  TcpFlush: Error 0 sending data on Tcp/Ip socket 66.
07/18/1996 22:37:28  sessRecvVerb: Error -50 from call to 'readRtn'.
07/18/1996 22:37:28  TcpFlush: Error 0 sending data on Tcp/Ip socket 66.
07/18/1996 22:37:28  cuIdentifyResp: Error -50 reading IdentifyResp from server.
07/18/1996 22:37:28  sessOpen: Error -50 receiving IdentifyResp verb from server
07/18/1996 22:37:28  ANS4017E Session rejected: TCP/IP connection failure
07/18/1996 22:39:28  TcpOpen: TCP/IP error connecting to server.
07/18/1996 22:39:28  TcpFlush: Error 0 sending data on Tcp/Ip socket 67.
07/18/1996 22:39:28  sessRecvVerb: Error -50 from call to 'readRtn'.
07/18/1996 22:39:28  TcpFlush: Error 0 sending data on Tcp/Ip socket 67.
07/18/1996 22:39:28  cuIdentifyResp: Error -50 reading IdentifyResp from server.
07/18/1996 22:39:28  sessOpen: Error -50 receiving IdentifyResp verb from server
07/18/1996 22:39:28  ANS4017E Session rejected: TCP/IP connection failure


Can anyone offer any suggestions?

I have played with tcpbufsize and tcpwindowsize, and this is
quite consistent.  The last file that was backed up was a
.MDB file, so I was leaning toward a file locking problem,
but this database is not even in use at night.