[cups] Print Queue out of order

Johannes Meixner jsmeix at suse.de
Thu Feb 18 01:33:18 PST 2016



On Feb 17 10:38 Helge Blischke wrote (excerpt):
>> Am 17.02.2016 um 09:30 schrieb Meyer Klaus <kmeyer at dohle.com>:
...
> D [26/Jan/2016:11:45:02 +0100] [Job 5375616] Process is dying
>  with "Caught termination signal: Job canceled
> D [26/Jan/2016:11:45:02 +0100] [Job 5375616] ", exit stat 0
> D [26/Jan/2016:11:45:02 +0100] [Job 5375616] Cleaning up...
> D [26/Jan/2016:11:45:02 +0100] [Job 5375616] Killing pdf-to-ps
> D [26/Jan/2016:11:45:09 +0100] [CGI] cgiSetArray: job_id[0]="5375616"
> D [26/Jan/2016:11:45:10 +0100] [Job 5375616] Killing kid3
>
> Obviously the foomatic-rip filter cancels the job in response
> to the queue being stopped
> (I have no idea why you need to restart cups to get the
> queue working again).
>
> I think you need to file a bug for the cups-filters package
> from open printing.org (or revert to the PostScript centric
> workflow you had with the 1.3.9 version of cups).

I don't know if the foomatic-rip filter cancels the job
in response to the queue being stopped.
>From the log is seems the foomatic-rip filter only exits
in response to the queue being stopped which means that
the foomatic-rip filter kills all its child-processes
to cleanly exit.
I don't see from the log that the foomatic-rip filter
cancels the print job.

>From "man cupsdisable" in CUPS 1.7.5
http://cups.org/documentation.php/doc-1.7/man-cupsenable.html
it seems that stopping a print queue does not by default
cancel its jobs.
Accordingly I think that filters, backends, or the cupsd
should not by default cancel jobs when a queue is stopped.

I think the primary issue is what Klaus Meyer wrote:
  "we are using cups 1.7.5 ... We notice, when a printer runs
   out of paper or a paper jam occurs the printjob is killed."

I would not expect that the printjob gets killed when
there is a temporary issue inside the printer device.

I would expect that by default the backend keeps running
for infinite time unless there is an issue that clearly
can never go away - only then the backend can exit with
appropriate non-zero exit code (cf. "man 7 backend").

Regarding what Klaus Meyer wrote:
  "printers ... connected via socket
   Before cups 1.7.5 we we were using cups 1.3.9 with the
   same printer drivers and we never had this problem"

The CUPS 1.3.9 socket backend did not do SNMP communication
with the printer.

Therefore I assume the CUPS 1.3.9 socket backend did not
notice any out of paper or paper jam issues iside the
printer.

But the CUPS 1.7.5 socket backend does SNMP communication
with the printer so that it can notice issues iside the
printer.

Now it depends on the particular details of the particular
printer devices if out of paper or paper jam issues iside
the printer appear via SNMP as issues that let the
CUPS 1.7.5 socket backend "think" something went serverely
wrong so that the whole print job must be killed.

This is only a blind guess off the top of my head.

To check whether or not CUPS backend SNMP communication
is related to the issue, see
http://cups.org/documentation.php/doc-1.7/network.html
how SNMP communication can be switched off for the
CUPS 1.7.5 socket backend (excerpt):
-------------------------------------------------------------
AppSocket Protocol
...
socket://ip-address-or-hostname/?option=value&option=value
...
The "contimeout" option controls the number of seconds that
the backend will wait to obtain a connection to the printer.
The default is 1 week.

The "snmp" option controls whether the socket backend queries
for supply and page count information via SNMP.

The "waiteof" option controls whether the socket backend
waits for the printer to complete the printing of the job.
The default is to wait.
...
Internet Printing Protocol (IPP)
...
snmp=false  Specifies that SNMP supply and page count
             queries should not be performed.
...
Line Printer Daemon (LPD) Protocol
...
contimeout=seconds  Specifies the number of seconds to wait
                     for the connection to the server to
                     complete.
-------------------------------------------------------------


Kind Regards
Johannes Meixner
-- 
SUSE LINUX GmbH - GF: Felix Imendoerffer, Jane Smithard,
Graham Norton - HRB 21284 (AG Nuernberg)




More information about the cups mailing list