Re: [netatalk-admins] PAP end-of-job timeouts?


Subject: Re: [netatalk-admins] PAP end-of-job timeouts?
From: J. Guy Stalnaker (jstalnak@students.wisc.edu)
Date: Wed Mar 03 1999 - 19:35:39 EST


At 10:16 AM -0500 on 3/3/99, you wrote:

> Try adding "-E" to the pap command in lp2pap.sh. There are a variety
> of printers (I'm guessing the ones where Adobe wrote the PAP
> implementation) that don't send EOF on the channel when the job is done
> printing. This is a protocol violation, and very inconvenient for
> error reporting.
>
> psf calls pap with the -E option by default. Probably lp2pap.sh should
> do the same, since several common printers exhibit this problem.

Hiya,

Is it possible that I have a problem similar to Robert E.'s, except on
Linux 2.0.35, netatalk1.42b_asun2.1.6, Apple LaserWriter Select310
(parallel port /dev/lp0 polling)? I send an error handler postscript job
to the printer (errhandler.ps that comes with OzTeX) to print out a page
with any errors. Imagine my surprise on getting a document page printed
with eight inches of text *over written* by this:

ERROR: timeout
OFFENDING COMMAND: timeout
STACK: 1902.77

For me the jobs queue fine, they just don't all come out of the printer!
Here's what I can tell you (data gathered from actual print attempts):

Output of 'lpq -P PS_300dpi-raw'

raw is ready and printing
Rank Owner Job Files Total Size
active guy 12 Mac Job 142102 bytes
1st guy 13 Mac Job 95425 bytes

Output of 'ls /var/spool/lpd/PS_300dpi-raw':

acct
cfA012buckland
cfA013buckland
dfA012buckland
dfA013buckland
lock
log
status

Contents of /var/spool/lpd/PS_300dpi-raw/lock:

105
cfA012buckland

Contents of /var/spool/lpd/PS_300dpi-raw/cfA012buckland:

Hbuckland
Pguy
JMac Job
TMac Job
Cbuckland
Lguy
fdfA012buckland
UdfA012buckland
NMac Job

========

In this case *nothing* ever came out of the printer. It was later that I
found the errhandler.ps file to get the printer to print any errors
encountered. I am associating the timeout error with the failure of these
and many other print jobs to complete successfully. The problem, at
present, follows this pattern: one job queued will usually print (I don't
know about the filesize issue mentioned by other posters, but I'll be
keeping watch on it now); any job queued while the first job is active will
usually fail to print, as will all subsequent jobs (I've had six queued
jobs disappear this way--off to /dev/null-land apparently).

One final note: the ready light on the printer sometimes will blink for
*minutes* after a job is finished (makes me think that the printer doesn't
know the job it done or lpd hasn't told it the job is done).

GuyS

                                                 ____________
-------------------------------------------------\__________/-----
J. Guy Stalnaker \________/
jstalnak@facstaff.wisc.edu \______/
http://bagend.doit.wisc.edu/theguy \____/
Live your life so you've no regrets! \__/
------------------------------------------------------\/----------



This archive was generated by hypermail 2b28 : Sat Dec 18 1999 - 16:16:23 EST