Am Tue, 11 Feb 2014 13:39:42 -0800 schrieb Philip Guenther <[email protected]>:
> On Tue, Feb 11, 2014 at 1:14 PM, <[email protected]> wrote: > > I have a strange behaviour of starting cupsd via rc.d-system: > > > > For a long time I use ~current on my old Lenovo T60: > > OpenBSD 5.5-beta (GENERIC.MP) #233: Fri Feb 7 12:14:13 MST 2014 > > [email protected]:/usr/src/sys/arch/i386/compile/GENERIC.MP > > (full dmesg at the end) > > You're keeping current with the kernel; you're keeping the base and > packages up to date too? "pkg_add -u" is your friend! > > > > Since a few month cupsd won't start by the rc.d-system. > > So you waited a few months to make it...harder to track down and less > likely to be fixed before the 5.5 release? A few months ago was the > update to cups 1.7.0 and there have been 6 changes to just the > Makefile since then... > > > > I have the > > following line in my rc.conf.local: > > $ cat /etc/rc.conf.local | grep cups > > pkg_scripts="adsuck clamd freshclam dbus_daemon avahi_daemon cupsd" > ... > > Well as you might already guess - cupsd won't start. The log-line > > reads quote "cupsd (failed)" > > That's just the console output. Does it write anything to the logs > under /var/logs/ ? > > > Philip Guenther > One more info on my setup: I start OpenBSD usually with /usr mounted read-only (and /var with noexec). While this would certainly be a nuisance for developers, for a production system this gives a little more protection - right? (Just another part of the safety-puzzle...) Does for any reason cups require /usr to be mounted read-write if started by rd.d-system? Seems to be rather unlikely but I didn't want to hide that piece of information. STEFAN
