scale days to clean by usage percentage

[fordmcdonald]

Incorporate Disk Usage to Improve Cleanup Cronjob

This PR adjusts the timeframe of files to delete according to the current disk usage. It takes the output of df -h /data/xnat/cache and scales the -mtime parameter of find by the percentage of disk space used. If disk usage exceeds 75% for truenas mounted directories, DAYS_TO_CLEAN will be scaled accordingly.

NOTE: df -h /data/xnat/cache will print the disk usage percentage of the entire truenas mount, not only the cache directory.

Bash division will resolve to an integer, so DAYS_TO_CLEAN will always be a valid input to find.

I also added -mindepth 1 to the find commands, so that the top level directory would not be deleted if empty.

counter=0
df -h /$PARENT_PATH/cache | while IFS= read -r line; do
        ((counter++))
        if [ $counter -eq 2 ]; then
                usage=$(echo "$line" | awk '{print $5}')
                usage=${usage%\%}
                if [ $usage -gt 75 ]; then
                    DAYS_TO_CLEAN=$((DAYS_TO_CLEAN / usage))
                fi
        fi
done

[mirestrepo]

if df -h /data/xnat/cache is the same as df -h /data/xnat/ than why not just use the later one?

also, should we also clean other directories like /prearchive and /build?

Another way is to get usage, remove iterate (if the logic of keeping file count across directories gets messy)

[kmilo9999]

Hi Ford,
I left a commend in the bash script in case you want to apply it.