urmcore README file:

every script i have ever seen that takes care of removing old core dumps has
used `find` to traverse the entire filesystem looking for them.  on a system
that uses slocate and runs the updatedb command frequently (most linux
distributions do this by default), this is a waste of effort since the
whole filesystem has been traversed and entered into a database.

this script uses `slocate` or `locate` to access that database of the
filesystem and find the core files.  it then checks to make sure the file
really is a core dump (another feature which i've never seen) using the
`file` command.  finally, it checks to make sure the core dump is old enough
to delete (4 days by default) and then removes it.

this script is intended to be dropped into cron's .daily directory to be run
nightly, but can be run from the command line if you like.  for cron, it needs
to be run after `updatedb` (or slocate -u) has been run to see the most recent 
slocate database of the file system.  on my redhat system, this is run by
the slocate.cron file (which comes alphabetically before 'urmcore' :P ), so
just dropping it in the cron.daily directory is fine.  your system may be
different. 

if you wish to change the age after which the core files are deleted, change
the MAX_AGE_DEFAULT variable at the beginning of the urmcore shell script or
modify the urmcore.cron shell script (placed in /etc/cron.daily by default):

example (/etc/cron.daily/urmcore.cron):

======================================================================
$! /bin/sh

/path/to/urmcore 7
======================================================================

this above will delete core files that haven't been accessed in a week (7 days)

- Uberdog <uberdog@linuxbox.com>

