When I query actlog, if the interruption goes long enough so that the
session is terminated, I get the following:
ANR0481W Session 1485 for node FTC-IS (Netware) terminated - client did =
not
respond in 600 seconds.

Otherwise, nothing.  Where else might I look?

>>> James SPORER <james.sporer AT CCMAIL.ADP.WISC DOT EDU> 12/04 3:28 PM >>>
     What messages are you seeing on the AIX ADSM server side?
     Jim Sporer



______________________________ Reply Separator
_________________________________
Subject: TCP/IP 'twixt NetWare 4.11 and AIX ADSM server
Author:  "ADSM: Dist Stor Manager" <ADSM-L AT VM.MARIST DOT EDU>  at IPNET
Date:    12/4/98 2:07 PM


I am attempting to backup NetWare 4.11.  We have a FDDI backbone to which
the
AIX ADSM server attaches directly.  NetWare boxes which use an Ethernet
connection have no problem.  However, NetWare boxes with FDDI cards lose
connection to the AIX ADSM server during a backup.  The interruption can
last
several minutes, after which another connection is usually established.

This happens repeatedly during the incremental.  The ADSM error log on the
NW
server gives the following message, apparantly for each time the connection=

is
lost:

12/01/1998 20:56:25 TcpFlush: Error 60     sending data on Tcp/Ip socket =
8.
12/01/1998 20:56:25 sessSendVerb:     Error sending Verb, rc: -50
12/01/1998 20:56:25 TcpFlush: Error 32     sending data on Tcp/Ip socket =
8.
12/01/1998 20:56:25 ANS1809E Session     is lost; initializing session
reopen procedure.
12/01/1998 20:56:25 ANS1809E Session is lost; initializing session reopen
procedure.
12/01/1998 20:56:40 ANS1810E ADSM session has been reestablished.

During the disconnection hiatus, you can't ping from either the NetWare
client
to the AIX server or from the AIX to the NetWare client it can't, but each
can
ping other addresses.

I have tried manipulating the TCPWINDOWSIZE, TCPBUFFSIZE, and
LARGECOMMBUFFERS
parameters to solve the problem, to no avail.

Anybody run into this problem before?  What do these error messages mean?