SUMMARY: lpd/printer stuck on large files

From: Fergus O'Reilly (fergus@capsogeti.fr)
Date: Thu Jul 18 1991 - 12:00:56 CDT


A SUMMARY OF THE PROBLEM (posted Tuesday July 16, 1991):

> System: Sun 3/180 running SunOS 4.1.1.
>
> Consider a Postscript printer---I've tried both a Dataproducts
> LZR-1260E and a Sun LW220 (i.e., Apple LaserWriter)---attached
> via a moderately long serial line to one of 16 ports on an ALM
> board.
> [...]
> Since the OS was upgraded from 4.0.3 the printer has died on
> large files. A certain number of pages are printed and then the
> printer stops and its status returns to IDLE. The truncated job
> just sits there at the top of the print queue and the lpd waits
> for Godot.

A SUMMARY OF THE SUGGESTED SOLUTIONS:

SOLUTION ONE. This problem was posed and solved on this list in
December, 1990. The fault lies with the ALM and its handling of
flow-control. There is a Sun patch (officially for 4.1 but it
applies to 4.1.1 also) and here is a bit of the README:
> Patch-ID# 100137-01
> Synopsis: ALM flow control patch
> Date: 09/Aug/90
> SunOS release: 4.1
> BugId's fixed with this patch: 1044721
> Architectures for which this patch is available: sun3
> Problem Description:
> This patch corrects the behavior described in Sun Bug No.
> 1044721. Systems running SunOS 4.1 with ALM (Systech MTI)
> boards can experience flow control problems on these serial
> ports. The system will not recognize an incoming XON
> character after receiving an XOFF character, which causes
> the port to appear frozen. Typing ^S^Q will reenable
> output.

[COMMENT: I cannot say if this will work since I have not yet
 transferred and installed the patch here (due a vital gateway being
 down). But it looks like what I need.]

SOLUTION TWO. Put `big nasty buffers' between things. This is a
general solution for flow-control problems between hardware-only
flow-control equipment attached to software-only flow-control
equipment.

[COMMENT: This would no-doubt work but if I can get things going
 again with the patch then I will be happy. I will add that it is
 usually possible to get into a flow control deadlock if data
 is spewed at a software-only flow-control device very fast and in
 large enough quantities. But my Sun 3/180 is not what you would
 call fast...]

SOLUTION THREE. I received some responses apropos of ALM-2 boards.
I omitted to mention that my board was an ALM-1. If anyone is
interested in receiving copies of these suggestions then please mail
me.

Many thanks to the following for their kind responses:

        David Fetrow <fetrow@biostat.washington.edu>
        Stefan Mochnacki <stefan@centaur.astro.utoronto.ca>
        Birger Wathne <birger@vest.sdata.no>
        Trish Stuckey <trish@trantor.harris-atd.com>
        John Polstra <jdp%polstra@uunet.UU.NET>

--
Fergus O'Reilly
fergus@capsogeti.fr



This archive was generated by hypermail 2.1.2 : Fri Sep 28 2001 - 23:06:20 CDT