we basically have a nasty issue with various backup files being created of random stuff and left there..

I want to implement a standardized naming convention, and have a cron task go through some specific directories and find junk thats older than so many days that match a specific description.

Now this might be something thats already been done and I'm reinventing the wheel, if so.. please do enlighten me.

Step one is to implement a small bash function in user's profiles that they wll use instead of just cp -pr..

It would cp the file with a specified retention and user information like so:

filename.backupdate_username_retentioninmonths 

it needs to use $LOGNAME, and check for $SUDO_USER and use that instead

by default it would use a 6mo retention, but if the -r mo flag is passed it would use that (in months).

Here is the start of that bash function that I slammed together (it doesnt have the optional retention stuff in yet).

backup() { if [ -n "$SUDO_USER" ]; then USER=$SUDO_USER else USER=$LOGNAME fi # @TODO: accept -r <mo> argument for dynamic retention retention=6 for file in "$@"; do local new=${file}.$(date '+%Y%m%d')_${USER}_${retention}mo while [[ -f $new ]]; do new+="~"; done; printf "copying '%s' to '%s'\n" "$file" "$new"; \cp -ipr "$file" "$new"; done } 

edit: it would also be cool if the function could accept a source AND target dir.. basically mirrior what cp does, just appending the username, backup date, and retention info..