Newsgroups: comp.sys.apollo
Path: utzoo!utgpu!news-server.csri.toronto.edu!helios.physics.utoronto.ca!alchemy.chem.utoronto.ca!system
From: system@alchemy.chem.utoronto.ca (System Admin (Mike Peterson))
Subject: Re: lpr, lpd , ... errors
Message-ID: <1991Feb21.143951.6576@alchemy.chem.utoronto.ca>
Organization: University of Toronto Chemistry Department
References: <9102211140.AA12209@zeus.>
Date: Thu, 21 Feb 1991 14:39:51 GMT

In article <9102211140.AA12209@zeus.> kniveton%zeus%motcid%uunet%samsung@umich.UUCP (Andy Kniveton) writes:
>   <... much excellent advice deleted ...>
>So if the above files don't have the right permissions lpr, etc .. Won't be 
>able to kick off the second lpd process.

You must also correct permissions on /usr/spool/lpd, else lpd will get
upset and just die (quietly):

alchemy:/# ls -lgR /usr/spool/lpd
total 20
drwxr-xr-x   1 daemon   daemon      16384 Feb 21 08:23 imagen

/usr/spool/lpd/imagen:
total 32
-rw-r----x   1 daemon   daemon          4 Feb 21 08:21 .seq
-rw-r--r--   1 daemon   daemon         19 Feb 21 08:23 lock
-rw-r--r--   1 daemon   daemon       8041 Feb 21 08:24 log
-rw-rw-r--   1 daemon   daemon         55 Feb 21 08:23 status

After changing permissions, kill lpd (not with lpc) and start another.
Each node the wants to print should have its own /usr/spool/lpd; they
must not be links to a master node since the sequence numbers are node
specific (this is better in any case since nodes can still accept print
requests even if the printer node is down).

You only need subdirectories for the printers you really have (not
all the ones Apollo sets up for you).

Lpd does indeed use NCS to make connections between Apollos, but also
listens on the standard TCP port for foreign systems.

Older versions of Domain/OS, especially on DN10000's, have lpd problems
where the daemon will not start (ever). All is OK as of SR10.2{,.p},
as long as you don't use any of your own filters. We can not start lpd
from /etc/rc on the DN10000, but it works fine if started from cron.

To catch errors, run 'syslogd' (this is a good idea if any UNIX daemons
are run, since most only log errors via syslog, and if it's not running,
you either get no error reports, or sometimes no daemon at all).
The syslog.conf file we use is:

#	@(#)syslog.conf	4.2 - 88/11/18
#	Sample syslog.conf
#	Format is type.priority[;type.priority]<white space>file_or_users
#	Blank lines and lines beginning with '#' are comments
#
# *.err;auth.notice				/dev/console
# daemon,auth.notice;*.err;mail.crit		/usr/adm/messages
# lpr.debug					/usr/adm/lpd-errs
# mail.debug					/usr/spool/mqueue/syslog

# *.alert;daemon.err				operator
# *.alert						root

# *.emerg						*

*.notice;kern.debug				/usr/adm/syslog
*.crit						/usr/adm/critical
lpr.debug					/usr/spool/lpd/imagen/log
*.emerg						*
*.alert;kern.err;auth.notice			system,root

which uses 2 files in /usr/adm (which must already exist) for logging of
normal and critical messages. These files must be cleaned up regularly,
as they can grow quite quickly. We also direct lpd to put its' errors
into the same log file (on a by-printer basis).

Mike.
-- 
Mike Peterson, System Administrator, U/Toronto Department of Chemistry
E-mail: system@alchemy.chem.utoronto.ca
Tel: (416) 978-7094                  Fax: (416) 978-8775
