I'm having the same problem. I applied a fix to the mailman logrotate configuration which solved the problem of the mailman directory getting massive, however every monday morning my logrotate is still going wild for several hours, taking 99% of my cpu.
Anybody know why it would run like this for hours? I get the same output from the "ps ax -eo pid,args --forest" command as noted above but don't really like the solution suggested (looks too much like a hack, but then again maybe I just don't understand it properly)
One thing to note is that based on the /var/lib/logrotate.status command, it "looks" like it is working on /var/log/cron but it is hard to tell as that file has no timestamps in it. The cron file is about 1.5 Mb but it shouldn't take hours to process it...
Any ideas?
NOTE: the fix to the mailman logrotate config file I mentioned is to change the file /etc/logrotate.d/mailman from:
/var/log/mailman/* {
missingok
}
to be:
/var/log/mailman/bounce {
missingok
}
/var/log/mailman/digest {
missingok
}
/var/log/mailman/error {
missingok
}
/var/log/mailman/post {
missingok
}
/var/log/mailman/smtp-failure {
missingok
}
/var/log/mailman/locks {
missingok
}
/var/log/mailman/fromusenet {
missingok
}
/var/log/mailman/subscribe {
missingok
}
/var/log/mailman/vette {
missingok
}
From what i can tell the logrotate program interprets the original file such that it trys to "rotate" all the already rotated files and basically gets into a rescursive nightmare.