uucp
Ken Benson
Our polling server is an OpenServer 5.0.5 machine. From the machine, we
poll, using uucp, several hundred customer machines on a regular (often
daily) basis.
Some historical background: Initially, both the customer machines - and our
polling machines were based on Unix 2.3.4 and/or Xenix machines.
Now, almost all of our customer base is on OpenServer 5.0.5.
However, we have a couple of customers that have upgraded to OpenServer
5.0.6 - and therein lies our problem.
This server polls successfully over 500 Openserver 5.0.5 machines on a
nightly basis - but when I try to poll an OpenServer 5.0.6 machine - with
all other things set up exactly the same as the OpenServer 5.0.5 machines, I
get communication garbage as in the login sequence.
I've already looked on the technical articles and pursued those things and
the result is EXACTLY the same - no change in baud rate has been effective,
yet. Instead of the login prompt from the remote (client) machine, I get a
random string of characters.
Is is possible to use the uucico that comes with OpenServer 5.0.5 on the
5.0.6 machine? Was there a significant change in the structure or setup of
the uucp package between these two versions?
I appreciate any help that can be provided.
Ken Benson
Ben Rosenthal
How refreshing to hear someone is using uucp. The problem on the 5.0.6 machines
may be in the dialers setup and not uucp. I ran into this pre 5.0.6 so I am not
sure this will work on 5.0.6. The machine would throw an error to the console
when the modem would answer. (I do not remember the error "getty something or
other"). Be sure the atdialer file is linked correctly to the modem setup
file. I do not have access to an SCO machine this second, so I can not be more
specific that that. This was discussed on this group before so try searching
archives for atdialer. Can you login as a user other then uucp?
Ben
Stephen M. Dunn
$I'm trying to trouble shoot a uucp problem ... here's the situation -
$This server polls successfully over 500 Openserver 5.0.5 machines on a
$nightly basis - but when I try to poll an OpenServer 5.0.6 machine - with
$all other things set up exactly the same as the OpenServer 5.0.5 machines, I
$get communication garbage as in the login sequence.
Does the same thing happen with cu or some other terminal emulator?
Is the whole thing garbage, or does some of the normal login prompt
come up as you'd expect?
Double-check the dialer you're using on the 5.0.6 machines. Maybe
the modem definition file you're using (assuming you're using the
atdialer) changed between versions. Also go through the inittab
and gettydefs lines with a fine-toothed comb. If it's possible
(it would involve having someone at the remote site on the phone
with you while you try logging in manually), get the output of
stty -a < /dev/tty1A (or whatever port the modem is on on the remote
system) from a working system and a non-working system and compare
them.
My guess is that the 5.0.6 box has a problem with either its modem
settings/initialization strings or its serial settings. I doubt it's a
problem on the 5.0.5 box since it can communicate perfectly well with
lots of other machines. If it's happening during the login sequence,
then it's also not uucico on the remote side, because uucico doesn't
start up at that point - you first go through getty, then login, and
only once you've satisfied login do you get uucico.
--
Stephen M. Dunn <ste...@stevedunn.ca>
>>>----------------> http://www.stevedunn.ca/ <----------------<<<
------------------------------------------------------------------
Say hi to my cat -- http://www.stevedunn.ca/photos/toby/