[cups.general] Sense and nonsense of RIPCache

Gene Heskett gene.heskett at gmail.com
Tue Jul 27 10:52:32 PDT 2010


On Tuesday, July 27, 2010 01:40:20 pm Till Kamppeter did opine:

> On 07/27/2010 05:10 PM, Gene Heskett wrote:
> >> cupsctl RIPCache=auto
> > 
> > Ahh, so my install does suffer from this.  Does this need to be put in
> > ones rc.local, and will it survive a cupsd restart?
> > 
> > Or does it take a restart to implement it once the above command has
> > been issued?
> > 
> > And since I have two printers, is it printer specific?
> 
> You run the command once and it changes cupsd.conf and afterwards
> restarts CUPS. so the change will be active immediately and stay after
> reboots.
> 
Great, then its done.

> > The only problem I have had so far was sending a targetz pdf to it
> > from acroread or its recent substitute telling it I wanted 50 copies,
> > then sending another I wanted 20 copies of.  My Brother laser ran out
> > of paper, which caused dbus-notify to go to 99% of one core.  Adding
> > paper, the printer finished the job ok, but dbus-notify stayed out in
> > lala land, and killing it killed kmail and much of x.  It was easier
> > to reboot, but rebooting to 2.6.35-rc6 failed to find the network, so
> > I had to do a full power down and rebooted to the PClinuxOS kernel as
> > the ati drivers work faster than the radeon drivers.  By about 2000 
> > fps.
> 
> This looks like a bug in dbus-notify and not like a problem of
> rendering. Also multiple copies do not change the memory consumption of
> Ghostscript. Ghostscript does never hold more than one page in memory.
> 
>     Till
That was how I understood it myself, Till.

Most of the *-notify utilities all seem to be renamed versions of each 
other, duz they all seem to have the same, grab a cpu & kill it, bugs.

knotify4 running on mdv seems to get several copies started at boot time, 
but I always had to go around with htop and kill the top listed one because 
it was using 99% of a core.  The rest seemed to be able to carry on w/o it 
so I never tried to trace it down.  But it appears there is only one copy 
of dbus-notify running.

And I believe I have been quoting the wrong name, it is 
"dbus-daemon --fork --print-id 6 --print-address 8 --session" that hung.  
There is apparently one incarnation running per printer.

According to htop.

-- 
Cheers, Gene
"There are four boxes to be used in defense of liberty:
 soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
* woot smiles serenely.
<woot> I don't want to seem over eager about getting into knghtbrd's
       siglist.





More information about the cups mailing list