Thursday, March 1, 2012

Find specific files then move them using a list [Linux]

In a previous post I mentioned finding and deleting certain files from Linux.
http://noveckg.blogspot.com/2011/01/find-and-delete-certain-files-in-centos.html

This time around, in order to maintain an audit trail and facilitate recovery options, a prudent approach would be to remove the files from the directory and place them somewhere else where disk util is not critical.

In this example, I used these scripts to clean up my MoodleData directory of old backup zipfiles. It saved tons of time so I didnt have to peruse over 8000 courses to clean this crap up. This may or may not apply to all moodle environments. Obviously, if you rely on this form of course backups, these scripts are not the solution for you. I'll reiterate what was said in that earlier post.


Basically what this command does is search a specified directory for a name/filetype and then move the files. 
If you run it from the top level directory (/), chances are something important may get moved and you can screw things up.

Command 1: Generate a list of backup zip files complete with filepath (datestamped for convenience). This can also be used as an index in case any files need to be replaced

find /path/to/moodledata/ -name \*backup\*.zip -exec ls -lh {} \; | awk '{ print $9}' > /path/to/course-backup-filelist-$(date +%Y-%m-%d).txt


 Command 2: read list line by line, and move file to alternative location

cat /path/to/course-backup-filelist-01-03-2012.txt | while read line; do mv "$line" /path/to/other/location; done


I managed to recover close to 20GB of storage from mostly unneeded files.

\m/ (rock on, moodlers)

-noveck