![](https://seccdn.libravatar.org/avatar/9435667f7160374bc34a8600b686aecd.jpg?s=120&d=mm&r=g)
В Mon, 09 Mar 2015 18:57:48 +0100 Per Jessen <per@computer.org> пишет:
Andrei Borzenkov wrote:
В Mon, 09 Mar 2015 17:43:47 +0100 Per Jessen <per@computer.org> пишет:
I've found the issue - it's a matter of the printk levels. On an openSUSE 11.0 system, the default was "1 4 1 7", on a 13.1 system it's "7 4 1 7". That's why I'm seeing so much output.
If that's intentional, it would sure make sense to direct it all to tty10. For myself, I'll adjust the printk levels.
"7" is default console level for printk. Before systemd it was usually adjusted in one of startup scripts (dmesg -n).
Aha, so this could be a regression. On my 11.0 systems, I see the following in /etc/init.d/boot.klog:
if test -x /usr/sbin/klogconsole ; then if test -x /sbin/showconsole ; then if test "`/sbin/showconsole`" = /dev/tty1 ; then ( test -c /dev/tty10 && > /dev/tty10 ) > /dev/null 2>&1 \ && /usr/sbin/klogconsole $KLOGCONSOLE_PARAMS -r10 else /usr/sbin/klogconsole $KLOGCONSOLE_PARAMS -r 0 fi fi elif test -x /bin/dmesg ; then /bin/dmesg -n 7 fi
I don't know how kernel.printk gets adjusted to "1 4 1 7".
One one of 13.2 printk level is "1" and on another it is "4". It is set to "1" by klogd.service. But I was not able to find out where it is set to "4". Also apparently openSUSE journald includes patch to redirect printk messages to /dev/tty10. This does not happen on at least one system either. One difference is, system where messages go to active tty has two console drivers. I suspect loading second driver resets setting made by journald earlier. This is real mess. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org