Hi David,
> Resending without the stupid HTML attachments, shame on me!
> >> Out of curiosity, what command are you using to restart klogd?
> I think >>I have the same problem with one of my boxes.
> >
> >/etc/rc.d/init.d/syslog restart
> >
> >kill -HUP does not do the trick.
>
> It has been my experience that 'kill -HUP syslogd' actually
> CAUSES klogd to stop working in both stock 7.0 and 7.1.
>
> By default klogd sends it's output to the syslog and it doen't
> seem to like syslogd being restarted, only stopping and starting
> klogd will get it going again (klogd halts on a SIGHUP). Issuing
> '/etc/rc.d/init.d/syslog restart' fixes the problem because it
> stops and starts syslogd and klogd in turn.
>
> So, it is a possibility that your problem arises from 'killall
> -HUP syslogd' getting called via cron scripts or elsewhere. The
> syslog logrotate file is one culprit. If you make no alteration
> to the default cron/logrotate scripts, then kernel logging will
> work ok from boot until the first weekend and then stop working
> until klogd is restarted.
>
> Getting klogd to log direct to a file might be worth trying too,
> (klogd -f file).
>
At last an explanation! Thanks for this, I'll adjust my cron jobs
accordingly.
Regards,
Tony
