Rotating Backup with rsync

Productivity Sauce

rsync is arguably the best command-line tool for performing local and remote backups, and you can set up a perfectly workable backup solution using a one-line shell script like this:

#!/bin/bash
rsync -avh sourcedir/ tartgetdir

This solution has one major drawback, though. It simply mirrors the contents of the source directory, so if some files and documents in it get corrupted, the script will duly back up the broken data. One way to solve this problem is to use a rotating backup which takes a full backup archive and then backs up all changed files in separate archives. This means that if anything goes wrong, you can restore data from previous backup archives.

There are several rsync-based rotating backup scripts floating on the web, and my favorite one is this one: http://www.noah.org/wiki/Rsync_backup. The script features a one-week rotation, and files changed during the day are saved in designated directories. During the first run, the script creates seven directories for each day of the week. The first directory contains the full backup, while the rest holds only modified files. The clever part is that backups in these directories appear complete because identical files are hard-linked. Deploying the script couldn't be easier. Create a daily cron job that points to the script followed by source and target directories:

@daily /path/to/rsync_backup sourcedir targetdir

To tweak the default rsync parameters used by the script, you need to modify two lines that start with RSYNC_OPTS. For example, to exclude certain directories and files from backup, create a exclude.list file containing a list of directories and files to be excluded. Add then the --exclude-from exclude.list parameter to the script as follows:

RSYNC_OPTS="-a --delete -v --exclude-from exclude.list"
RSYNC_OPTS="-a --delete -q  --exclude-from exclude.list"

Comments