This is not the first time this has happened. It last happened about a month ago. Tonight it happened for a different reason. This time it wasn’t the size of the mbox (I’m not using mbox on this server).
This time it’s the size of the log.
$ ls -lha ~/.mailfilter.log* -rw-r--r-- 1 dan dan 52k Jan 11 00:18 /usr/home/dan/.mailfilter.log -rw------- 1 dan dan 48M Jan 10 21:03 /usr/home/dan/.mailfilter.log.1
Once I mv‘d the file away and did a postfix flush, everything was delivered.
It’s time for some file rotation.
Let’s fix that by adding a file to /usr/local/etc/newsyslog.conf.d/.
[dan@mailjail:/usr/local/etc/newsyslog.conf.d] $ cat mailfilter # logfilename [owner:group] mode count size when flags [/pid_file] [sig_num] /usr/home/dan/.mailfilter.log dan:dan 600 80 * @T07 B [dan@mailjail:/usr/local/etc/newsyslog.conf.d] $
That should deal to this annoying little problem.
FYI, I had a Nagios alert for the mail queue on this server. It was properly reporting the problem and it notified me. Via email.
The problem was also on my Nagios website, but I was so preoccupied with writing an article on Bacula that I didn’t notice.
The first entry in the log file was dated Sep 3 16:02:04 2014.