[cups.general] cups daemon dying

[Matt Broughton]

This reply also posted to the macosx-list at linuxprinting.org list as the 
OP has also posted the same question there. 

In article <23140-cups.general at news.easysw.com>, <marc at arctic.net> 
wrote:

> we are running cups on Mac hosts against a linux host running .23.  we are 
> using a 
> BrowsePoll server directive in the workstation cupsd.conf.  we have started 
> to 
> notice that in a number of cases we are seeing either the polld command 
> disappear 
> leaving a /usr/sbin/cupsd -L running or the polld command is still running 
> but 
> there is no cupsd process running,  and therefore no printers available.
> 
> The following appears in the console:
> 
> Nov 21 08:17:57 Macintosh /System/Library/CoreServices/mcxd.app/Contents/
> MacOS/mcxd: cupsd mach_msg error (ipc/send) timed out
> 
> Can someone help us figure out why we are getting these time outs and what we 
> can do to keep the daemon running during any loss of connection so we don't 
> have 
> to log in as ladmin to restart the daemon

Try changing your /etc/hostconfig file.  Change the CUPS entry from 
-AUTOMATIC- to -YES-.  This will prevent cupsd from starting with the -L 
option.  

I don't guarantee this will solve your problem, but it is a good bet.  
Beginning with Tiger, Apple added the -L option to cupsd.  This appears 
to be an Apple only "feature".  A separate compile and installation of 
CUPS 1.1.23 on a test Tiger partition showed the the original source 
code does not contain an -L option.  This option seems to start the 
cupsd only if a printer is found during the boot process.  My guess is 
that this feature is also timing out the cupsd leading to your problems.

I have not found any information  on this -L option in the developer 
section on the Apple website.  I'm sure it is mentioned somewhere at a 
level I would not understand.

Matt

-- 
Matt Broughton
Only relatives are absolute.