[cups.general] Re: lprm <PID> => Not all processes killed

Felix E. Klee felix.klee at inka.de
Sun Aug 29 09:53:24 PDT 2004


On Sun, 29 Aug 2004 10:18:35 -0400 Gene Heskett wrote:
> Thats not the desired action Michael in 99% of the cases here. When 
> the printer has lost its init, and is doing what is essentially a hex 
> dump of the raster data being sent, and that might take 5000 pages of 
> paper to finish it, we should be able to kill it unconditionally, in 
> its tracks so to speak. 

Well, that the printer looses its - as you say - init usually only
happens because of hardware failure or someone fiddled around with the
hardware (i.e. turned the printer off). 

If a job is killed with lprm and the printer is not touched then there
are no problems, I guess. The printer will finish printing the pages it
has in the buffer and uses the information provided by that stray
process to terminate the last page properly.

I, however, often turn off the printer after or before I kill a job with
lprm. The reason is that the printer usually has some more pages in the
buffer and I don't want them to be printed. In this case the stray job
turns into an annoying beast that keeps sending data which often results
in loads of plain text pages.

The latter case is especially frustrating for novices that don't know
how to kill the stray process. This is an additional reason why, on
several systems, I'd like to configure CUPS to always kill all relevant
processes upon issuing "lprm <ID>". Is this possible?

> I've been in that situation several times, and the only way out is to
> reboot which will kill that process about 95% of the time. 

A "kill <process_id>" should be sufficient.

Felix





More information about the cups mailing list