We have been experiencing some very unusual telnet hangs since I
installed the Telnet "I" patch PTDFDT3A, although there are one or
two possible other culprits.  I hesitate to point the finger, but it
is looking like the "I" patch more and more.  The problem:

Users are in Reflection, doing character mode data entry.  While
entering data (keystrokes themselves, not after hitting return), the
session simply "hangs" - no further echo of input, etc.  Windows is
not hung (other apps are fine), the application itself is not hung
(command bar menu drop-downs work and function, etc).  There are *NO*
network errors logged on the console like the older NetIPC error 64s and
such.  The user can close out Reflection or do whatever they want, it
doesn't respond on the client or the 3000.  The session remains
connected, if they don't close out/exit Reflection, the session hangs
indefinitely (known to be true for hours).  If a session is reported
hung, and they call before closing Reflection, I can do an :abortjob of
their session -- they do NOT get the "session aborted by system
management" but they DO get a "<Telnet session has been terminated>"
indicative that the PC and host still have an established TCP connection
prior to the abort.

NETTOOL.NET "resource;display;quit" indicates nothing close to the
limits.  LINKCONTROL statistics indicate no errors (100TX LANIC, full
duplex, speed/duplex hard coded in NMMGR and on switch on the other
end).  No errors recorded on the switch.  No NMLGxxxx recorded errors.

I restarted the network tonight, will see if it persists tomorrow; but
if it does, I'm thinking about backing out to my previous stage area to
see if that makes a difference.

If it matters, our other network patch levels are NMSFDQ9A, NSTLX57A.

Jeff Kell <[log in to unmask]>