“Any Linux system will generate many log files by default, containing various information about the operation of the system. If no rotation would occur on the various log files, then they will just grow bigger and bigger, filling up the space. This little how-to will show you how the default log rotation works, based on syslog and logrotate.”
What I’d like to know is who is the brain-dead moron who decided to set the defaults to rotate the logs so frequently (most are daily, some are weekly, none are monthly), and to keep the rotated logs in the same directory are the in-use logs??
A much nicer, more sane, default would be to put the rotated logs into /var/log/backups (or something similarly named), and to only rotate the logs monthly.
Keep the old logs separate from the in-user logs.
And what is up with the config file format? Why is it that every Linux distro (actually, every program developed on Linux) has gone to the “1 config file per sub-unit” format with all the config files under a conf.d/ sub-directory?
What’s wrong with “1 config file per program” and “1 line per sub-unit” in that config file?
/etc/logrotate/conf.d, /etc/apache2/conf.d, everywhere you look is a friggin’ conf.d/ sub-directory. It takes too long to find information when it’s split across a dozen files in two directories … and it makes editing them with a console text editor a royal pain.
KISS used to be such a nice principal, and the bulk of Linux developers followed it. Now it appears to have been replaced with “spread the info all over the place, it’ll be better in the end, I swear”.
Edited 2006-06-07 21:36