A simple differential backup script for systems maintained by package managers with single-file manifests
Project description
Introduction
This is a simple backup script for systems maintained by package management software providing a detailed manifest about installed files. It uses the repository of system-installed files to automatically exclude them from backup because those files can (and should) be reinstalled easily with the package manager. The program detects the package manager automatically and makes use of its manifest of installed files. At the moment, portage (on Gentoo Linux systems and derivations) and dpkg (Debian and derivations) are supported. More systems can be supported with plugins.
Configuration
Backup uses a configuration file (~/.local/etc/backup if not overridden on the command line) for its required information in order to be run with a simple command for daily use. The configuration file is in the YAML format. The format is simple enough to be described by an example:
roots: - / prune: - /tmp - /proc - /sys - /dev - /mnt - /var/tmp - /var/run - /var/cache - /run - /lost+found destination: user: backup password: hj983sh34 host: backuphost port: 22 directory: /backup/{HOSTNAME}/rsync-backup generations: 7
There are three sections (root, prune and destination). In the root section, a list of directories to be included in the backup are stated. Each list element is on a separate line and begins with “- “ (the space after the dash is required).
The prune section contains a list of directories to be excluded from the backup. If it is missing, the following directories are excluded by default: /tmp, /proc, /sys, /dev, /mnt, /var/tmp, /var/run, /var/cache, /run, /lost+found.
In the destination section, a couple of parameters specify where the backup has to be written to. Since rsync is capable of writing to remote hosts through the use of ssh, three parameters, user, host and port specify the necessary information for reaching a remote host. As of version 0.9, the host parameter is required, meaning only remote backups are possible. This will change in later versions, allowing local backups. The parameter directory specifies the base destination directory on the remote host. Parameter port has a default of 22. As you can see in the example, there can be variable substitution expressions in the form of {VARIABLE}. The expression will be replaced by the respective environment variable. The parameter generations specifies how many generations of old backups are to be retained before they are removed. The default is 30. If a private key is required to log in to the remote host, it can be specified with a key: key-file entry in the destination section.
Backup organisation
To enable differential backups, the backup is stored under a subdirectory of the destination directory specified in the destination section of the configuration. This subdirectory has a name resembling a time stamp with date and time, like 2014-02-12-03:12:26 where the first three numbers are the date in ISO-format and the last three numbers are the time in 24-hour-format.
The new backup is stored in reference to the last backup. All changed files are stored normally. Files which have not changed since the last backup are simply hard-linked to the corresponding file in the old backup. This way they don’t take any addition space on the backup medium (apart from the directory entry). This is done by rsync by the use of the --link-dest option.
Usage
The program is intended to be run by the simple command:
backup
There are some options for testing and one-time use:
- -h, --help
shows a help message and exits
- --config file, -c file
to specify the path to an alternative configuration file (default: ~/.backup)
- --version, -V
displays the program version
- --log-level level, -L level
to set the log level. Level is one of DEBUG, INFO, WARNING, ERROR, or CRITICAL. DEBUG shows the most detail while CRITICAL shows almost nothing. (default: INFO)
- --log-file file, -l file
to set log file (default: /var/log/backup)
Per-directory exclusion
To exclude a directory from backup, simply create a file named .no-backup in it.
Installation
The software can be installed easily from the Python software repository, either on the command line or by downloading the package and installing it explicitly.
Installation using pip
On the command line, type:
pip install --user jw.backup
Explicit Installation from a downloaded package
Download the package from https://pypi.python.org/pypi/jw.backup. Unpack it, cd into the unpacked directory and type the command:
python setup.py install --user
Bug reports
Please report bugs and enhancement requests to https://bitbucket.org/JohnnyWezel/jw.backup/issues.
Project details
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.