Clean up "incremental" (--link-dest) rsync backups

00_cleanup.rst

Backups with rsync's --link-dest are great, they let you keep backup daily versions of files with instant file level access while being fast and relatively space efficient.

The code below prunes duplicates etc. from large collections of --link-dest based incremental backups when the overhead of the directory structure etc. starts to consume too much space.

cleanlinks.sh

space() {  # convenience function to show space on this device
    df -BM . | tail -1 | tr -s ' ' | cut -d' ' -f4
}
min () {  # show minutes and seconds
    date '+%M:%S'
}

# find will complain if '.' is on the path when you use -execdir
PATH=/bin:/usr/bin

for ROOT in 201[34]-*; do
    echo $ROOT $(date) $(space)
    echo "  $(min) remove links > 2"
    
    find $ROOT -type f -links +1 -execdir rm \{\} \+
    
    echo "  $(min) remove broken sym links"
    
    find $ROOT -depth -type l -xtype l -execdir rm \{\} \+
    
    echo "  $(min) remove empty dirs"
    
    find $ROOT -depth -type d -empty -execdir rmdir \{\} \;
done