So, about a week ago, I noticed that I was getting an error from webalizer about a referrer field being too large when I ran it from the command line (“Warning: Truncating oversized referrer field”). I knew this was something I should take care of, but I didn’t feel any pressure to do it right away.
This morning, having been trying to set up formmail to work with my old bobist survey, I started up my qmail smtpd. I then watched as it delivered 2400 mail messages first to root, then to the user account that the root alias pointed to. 2400 messages from the cron daemon containing that webalizer error message.
I really never wanted to see Pine display the query “Expunge the 2400 deleted messages from “INBOX”?”
So, I set the cron entry for webalizer to kick all errors and whining to /dev/null (I could have just done in from within the webalizer.conf, but oh well). I then went through my access log trying to track down the two lines that were causing all this hell.
After a few minutes of playing with grep and webalizer, I narrowed it down to two normal looking lines. Here they are:
prometheus.cnmnetwork.com – – [01/Sep/2000:01:11:01 -0700] “GET /~goldcam/” HTTP/1.0″ 404 691 “-” “Mozilla/4.74 [en] (X11; U; FreeBSD 4.0-STABLE i386)”
prometheus.cnmnetwork.com – – [01/Sep/2000:01:12:55 -0700] “GET /~goldcam/” HTTP/1.0″ 404 691 “-” “Mozilla/4.74 [en] (X11; U; FreeBSD 4.0-STABLE i386)”
So, I cleaned them out of the access_log, ran webalizer and all is good again. That really sucked.