[cups] "Aborting job because it has no files."

[Johannes Meixner]

Hello,

On Nov 23 21:30 Helge Blischke wrote (excerpt):
>> Am 23.11.2016 um 17:14 schrieb Szabolcs Csermák <csermak.szabolcs at gmail.com>:
...
>> I [23/Nov/2016:17:02:32 +0100] [Job 18094] Job submission timed out.
...
>> E [23/Nov/2016:17:02:32 +0100] [Job 18094] Aborting job because it has no files.
...
>> D [23/Nov/2016:17:02:32 +0100] Securely removing "/var/spool/cups/a18094".
>> 
>> What does "job submission timed out" means?

In the documenmtation of your particular CUPS version
see the various "timeout" related settings for cupsd.conf
perhaps you have a too short timeout setting somewhere?


>> We are using "lp" command in a loop, printing one ticket at every
>> time. So I we want to print 150 tickets, the lp command executes 150
>> times.
...
> Perhaps your print job submissions occur too rapidly, so that some
> thresholds may be reached (max jobs e.g.).

In the past I did a few tests with hundreds of thousands of jobs
that I submitted by a bash script as fast as possible to
hundreds of print queues that printout into /dev/null or into
files (via "FileDevice Yes" or via selfmade special backends)
but I never experienced a failure like the one described here.

When thresholds like MaxJobs/MaxJobsPerPrinter/MaxJobsPerUser
are reached, the lp command fails to submit the job (which is
the expected and documented behaviour).
But when the lp command succeeds, I never experienced a failed
job afterwards.

I did my above described tests with CUPS 1.5.4 and CUPS 1.7.5
(i.e. I did not test mass printing throughput with CUPS 2.x).

To avoid misunderstandings:
With real printouts on real printer devices in real networks
there could be arbitrary weird failures for mass printing.


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