[QUOTE=RDove;3786293]I think it has to do with your logrotation. Please see: /etc/cron.daily/logrotate and /etc/cron.daily/logrotate. Once a log fills up, it restarts syslogd. Here is an example from my box:
Code:
[root@ronalddove log]# grep syslogd *
messages:Dec 6 04:02:23 ronalddove syslogd 1.4.1: restart.
messages.1:Nov 29 04:02:05 ronalddove syslogd 1.4.1: restart.
messages.2:Nov 22 04:02:06 ronalddove syslogd 1.4.1: restart.
messages.3:Nov 15 04:02:08 ronalddove syslogd 1.4.1: restart.
messages.4:Nov 8 04:02:08 ronalddove syslogd 1.4.1: restart.
[root@ronalddove log]# w
02:46:07 up 31 days, 14:56, 4 users, load average: 1.66, 1.74, 1.79
As you can see, each new messages.X log has one syslogd restart. That is because a rotation took place. Also what is interesting is they all occur at the same time too.. around 04:02 on my box, so I guess you should see something similar on yours.