[comp.mail.uucp] UUCP on a Sun SLC

deans@stpaul.gov (Dean R Schrimpf) (05/04/91)

From: deans@stpaul.gov (Dean R Schrimpf)
Newsgroups: Comp.mail.uucp
Subject: Re:UUCP on a Sun SLC
Organization: City of St. Paul Public Works


>My first UUCP connection using this new setup (Sun Sparc SLC, SunOS
>4.1) went smooth as silk.  Lots of files were transferred.  All was
>well in the land.  The *next* time, and all subsequent times, I get
>connected, the log shows that it's starting a transfer (from the
>remote to my site), and I immediately get "alarm n" messages until the
>converstaion fails.  I can send more detailed log info for anyone who
>thinks they can help.  The modem (if it matters) is a Telebit T1000.
>Everything worked fine (as far as UUCP went) on a '386 box running
>various flavors or Unix.  I anyone has *any* insight into this
>problem, or think they would if I gave them more info, please E-mail a
>response. 
>
>While I'm at it... does anyone know why the serial port (/dev/ttya) on
>the Sparc should want to hang (*completely*) so that I have to reboot
>to get it back?  When it gets in this state UUCP and tip can't touch
>it ("all devices busy") and stty < or > to/from it hang and I have to
>kill it with a QUIT signal.  Any info on this, too, would be
>appreciated.
>
>Thanks much!
>
>-EricF
>ericf@zurich.ai.mit.edu

We have had similar problems with or Sun SLC running SunOS 4.1.1.

What it sounds like to me is that you are using 7 data bits instead of 8.
Sun assumes 7 data bits even parity as default and UUCICO does not 
automatically change to 8 data bits no parity when invoked.

At first you might have only been sending ASCII text, and then you hit a 
binary file which you then stall on.

We discovered two ways to solve this problem.
First,  in your Dialers file add two things  put a "" P_ZERO "" in the 
front of your dialer script and at the end of your script put STTY=raw.
This will put UUCICO in 8 data bits no parity, along with some other
setup information that you can read about in your Administration manual.
Also, I found that even though you set STTY=raw which does set no parity
if the other site is activly checking for no parity this will not completely
solve the initial handshake problem, which P_ZERO does, and it does not 
definitely hurt.

Secondly, you can change your default configuration for 8 data bits no
parity, of the top of my head you need to put ap in place of something 7.
But I would not go all the way to STTY=raw, because this affects
everthing that might be connected to your SLC, I found out the hard way
with a WYSE terminal, becuase raw has no handshaking.

Hope I could be of some help.  Let me know if this solves your problem.


Dean Schrimpf
deans@pwcs.stpaul.gov
.

guy@auspex.auspex.com (Guy Harris) (05/07/91)

>What it sounds like to me is that you are using 7 data bits instead of 8.
>Sun assumes 7 data bits even parity as default and UUCICO does not 
>automatically change to 8 data bits no parity when invoked.
>
>At first you might have only been sending ASCII text, and then you hit a 
>binary file which you then stall on.

If your UUCP stalls on a binary file because, when using the "g"
protocol, it's using 7 data bits even parity, you have a Mutant UUCP
from Hell - you don't have the SunOS-pre-4.1 UUCP, nor the 4.1-and-later
UUCP, nor any other UUCP I know about.

The "g" protocol, on any sane UUCP, puts the port into 8 bits, no parity
mode, forcibly.

Now, *before* the "g" protocol is started up, although the serial port
hardware *is* in 8-bits-no-parity mode, UUCP will glue on a parity bit
as specified by the P_ flag, with the default being even parity.  Having
the wrong parity for the site to which you're connecting can cause
problems when logging in; it can't disrupt the "g" protocol, though. 
P_ZERO tells it to glue a zero bit on.