[cups.general] CUPS slow to start

[Kurt Pfeifle]

Minatra, Pat H. <pminatra at hsutx.edu> wrote (Wednesday 23 August 2006 16:39):

> In /var/spool/cups:
> 
> ls -l c* | wc -l = 8035
> ls -l d* | wc -l = 8034

Well, 8000 jobs in the job history, plus all of the job files preserved as
well -- that does entirely explain why you should see a very slow startup
of cupsd.

> Values of:
> PreserveJobFiles Yes
> PreserveJobHistory Yes
> MaxJobs 0
> AutoPurgeJobs No

If you set f.e.

  MaxJobs 1000

you'd keep the 1000 latest jobs and their history around without a need
to devise your own cleanup script.

> The reason these are set like they are is due to customer complaint
> sometimes the only thing that I have available to determine if a job was
> actually sent is to locate it in /var/spool/cups.  Now, what I can do is
> a cron job or something that after they get so old they are removed
> through cron but I am not quite sure that is the 'correct' way to handle
> it; 

Well, it's not an elegant way :-)  but you may have no other choice.

> something like say '/bin/find /var/spool/cups -mtime +60 -exec rm {} 
> \;' or something along that line.  

If you urgently need the jobs around for a long time, you could 
replace the "rm" by an "mv". That way you could push back any job 
(c* plus d* file) under investigation into the valid spool directory 
(but you'd need to restart cupsd to get it recognized).

Cheers,
Kurt

> This would allow me to keep anything 
> up to 60 days I think.
> 
> We only have 147 printers on one CUPS server on ONE Uznix box with one
> main application.
> 
> Am open to any and all suggestions.  Thank you so much.
> 
> -------------------------
> "Life is but a twinkle in the eye of eternity"
> "The shortest distance between a problem and a solution is the distance
> between your knees and the floor"
> "sorrow looks back - worry looks around - faith looks up"
> Regards,
> Pat H. Minatra - N5GJR
> (325) 670-5804 voice
> (325) 670-1570 fax
> Hardin*Simmons University
> www.hsutx.edu
> PO BOX 16040
> Abilene, TX  79698
> 
> -----Original Message-----
> From: cups-bounces at easysw.com [mailto:cups-bounces at easysw.com] On Behalf
> Of Kurt Pfeifle
> Sent: Wednesday, August 23, 2006 9:28 AM
> To: cups at easysw.com
> Subject: Re: [cups.general] CUPS slow to start
> 
> Minatra, Pat H. <pminatra at hsutx.edu> wrote (Wednesday 23 August 2006
> 14:34):
> 
>> 
>> We are on a Linux Red Hat OS running CUPS 1.1.22 and am unable to
>> upgrade to 1.2.2.
>> 
>> Item 1:
>> -------
>> Can anyone tell me as to why; when I am logged in as root and I run
> the
>> command '/etc/init.d/cups start' why it would take nearly 3 minutes to
>> start?  I am at a loss what to look for.  My FilterLimit in
>> /etc/cups.conf is set at 500 which is the only thing that rings a bell
>> when looking in the admin manual.  Am I hitting some hidden limiter
>> somewhere that is hard to find?
> 
> You may have lots and lots of files in the job history. What is the
> output of
> 
>   ls -l /var/spool/cups/c* | wc -l
>   ls -l /var/spool/cups/d* | wc -l
> 
> ? What are your settings in cupsd.conf for
> 
>   PreserveJobFiles
>   PreserveJobHistory
>   MaxJobs
>   AutoPurgeJobs
> 
> ?
> 
>> Item 2:
>> -------
>> This version of CUPS has run relatively well for us but we are
>> continually getting complaints from the user community that their jobs
>> are not printing; however, when I go to /var/spool/cups and look for
> the
>> output I find no output that parallels with their claim of printing so
> I
>> must presume from that there is not real print job.
> 
> You have a single CUPS server? And all print clients access that
> same CUPS server?
> 
>> Thank you for any information and have a GREAT day!
> 
> Cheers,
> Kurt
> _______________________________________________
> cups mailing list
> cups at easysw.com
> http://lists.easysw.com/mailman/listinfo/cups