* Jon R. Kibler <[EMAIL PROTECTED]> > The only thing I have left to suggest is on the machine that isn't > logging, run syslogd in debug mode and see if it is ever getting the > log requests.
Have the usual syslogd caveats been covered yet? 1. Some syslog varients *really* want tabs, not spaces in /etc/syslog.conf between the facility/priority statement and the destination. $ sudo perl -i.orig -pe 's/ /\t/g' /etc/syslog.conf $ diff /etc/syslog.conf.orig /etc/syslog.conf If different (and diff is not ignoring whitespace changes), restart syslogd. 2. Some syslog varients will not log if the appropriate logfile did not exist when the syslog daemon started up. No /var/log/maillog, no logs. Create the logfile with proper permissions, then restart syslog. Currently I replace the default syslog daemon with syslog-ng if possible, which allows logging over TCP (no lost UDP logs: good for logging from a limited number of core servers, as opposed to lots of clients), and other benefits such as built in logfile rotation (no stupid conveyer belt of logfiles and syslogd restarting) and logging of the facility and priority. _______________________________________________ Visit http://www.mimedefang.org and http://www.canit.ca MIMEDefang mailing list [EMAIL PROTECTED] http://lists.roaringpenguin.com/mailman/listinfo/mimedefang
