A script for backing up data using ssh(1), rsync(1), and scp(1). Can handle backup generations on btrfs and ZFS.

Copyright (c)2008-2022 Alexander Barton (alex@barton.de)
Homepage: https://github.com/alexbarton/backup-script

Call the scripts located in ./bin directly from the source folder, or run make install to install them to /usr/local/sbin.

You can set PREFIX to use an other path prefix than /usr/local like this: make PREFIX=/opt/backup-script install.

Run all or individual backup jobs.

Usage: backup-script [<options>] [<job> [<job> [...]]]

Options:

• -n, --dry-run: Test run only, don't copy any data.
• -p, --progress: Show progress, see rsync(1).
• -t TAG, --tag TAG: Only run jobs with tag TAG (see "tags" variable below).
• -x, --no-exec: Don't run global setup-, pre-, and post-exec commands.

When no job is given, all defined jobs are run.

Backup all systems ("run all jobs") and mail the report to "root".

Usage: backup-script-wrapper [<backup-script-options-and-job-names ...>]

Show information about backups.

Usage:

• backup-status [--errors|--latest] [--quick] [<job> [<job> [...]]]
• backup-status --running

Options:

• -e, --errors: only show current backups with errors (implies --latest).
• -l, --latest: only show latest backup generations.
• -q, --quick: quick mode, don't calculate backup sizes.
• -r, --running: check if an backup-script task is currently running.

When no job is given, all defined jobs are listed.

Show "relevant" differences in system configuration between backup generations.

Usage: backup-audit [-q] [-v] [<job> [<job> [...]]]

Options:

• -d, --dirs: compare two backup directories (not jobs).
• -q, --quiet: quite mode, only list jobs with changes or errors.
• -v, --verbose: verbose mode, show all checks that are run.

When no job is given, all defined jobs are checked.

All default configuration variables are read from the first file found of this list: /usr/local/etc/backup-script.conf, /etc/backup-script.conf or from /etc/backup-script.d/backup-script.conf (deprecated).

All jobs that should be backed-up are configured using individual files in the configuration directory, which is /usr/local/etc/backup-script.d/ or /etc/backup-script.d/ by default (whichever is found first), and can be specified using the conf_d variable in the main configuration file.

The must be one job file for each system to backup (files ending in *.sh are skipped, as well as files named backup-script.conf). Please avoid spaces and other "special" characters! The filename is used as hostname for the system by default, but this can be overwritten using the system configuration variable.

Variables in backup-script.conf must be prefixed with default_ to define default values for all jobs.

All defaults can be overwritten in the individual job configuration files.

For example:

• /etc/backups-script.conf: defaults for all hosts
• /etc/backups-script.d/host01.example.net: configuration for host 1
• /etc/backups-script.d/clientXY.example.com: configuration for host 2

The following global configuration options exist:

• setup_exec: Script to run before creating the lock file etc.
• pre_exec: Pre-execution script, run before all jobs.
• post_exec: Post-execution script, run after all jobs.

In Addition, all job configuration options (see below) that have a "default_XXX" variant can be used and define default values for all jobs that don't overwrite them individually.

System host name. Default: file name.

Note: There is no default_system variable!

Backup type to use. Default: rsync.

• rsync: system backup using rsync(1). Use source_root to specify the root directory to save.

• scp: file backup using scp(1). Use files to specify the files to copy.

• disabled: job is disabled and will not be run. This becomes accounted as "success" in the summary and exit code of the backup script.

Please note that neither ssh_args_add, rsync_args_add, compress, nor any "exclude" parameters are supported when using the "scp" backup type! And There "scp" backup type never deletes files from the backup store; so if you reduce the list of files to backup, old files will still be kept, because they were already saved in an older generation (but no longer updated).

Remote user. Default: root.

Remote root directory, must end with a slash ("/") character! Default: "/".

When saving the whole (remote) system ("/"), default excludes are set up automatically, which exclude standard system directories like /sys and /proc.

Space separated list of files to copy when using the "scp" backup_type. Default: "running-config".

Local backup directory. The backup data of each job is stored in a folder named like the job (see system variable) inside of this target directory.

Note: There is no default, you have to specify this variable, for example as default_target in the backups-script.conf file!

Additional parameters for ssh. Default: none.

Additional parameters for rsync. Default: none.

Additional (exclude) parameters for rsync. Default: none.

Deprecated! Use "exclude_dirs_add" instead!

Additional directory path names to exclude from the backup. Use full path names separated by spaces. Default: none.

Enable (1) or disable (0) rsync transfer compression. Default: 1 (on).

Enable (1) or disable (0) ping'ing the target system to make sure that it is reachable before calling rsync. Default: 1 (on).

Enable (1) or disable (0) local mode: when local mode is in effect, rsync is called without using ssh, which is a good idea when saving the local system. Default: 0 (off; use ssh).

Number of generations to keep. Default: 0 (none).

On a suitable target file systems (see target variable), this script can generate generations using snapshots: the script creates a new snapshot named with the time stamp for each generation inside of the system directory inside of the target directory.

Supported file systems are:

• btrfs: All generations are btrfs subvolumes and named after the date and time.
• ZFS: All generations are ZFS file systems. Latest generation is named current, elders are links to the ZFS snapshot directories.

The latest snapshot is always reachable using a symlink named latest inside the system directory.

The maximum I/O timeout in seconds. If no data is transferred for the specified time then rsync will abort. Default: 1800 (30 minutes).

Comma-separated list of tags of this job. All uppercase tag names are reserved and become set automatically on runtime:

• NONE: Jobs with no other tags at all.
• ALL: Matches all jobs, regardless of their tags (see -t/--tags option).
• LOCAL: All jobs running on "localhost".

Default: NONE.

Optional script to execute before rsync starts. Default: none.

When the job_pre_exec script returns an error (exit code is not 0), the backup run is skipped!

Optional script to execute after rsync exited. Default: none.

The following job configurations variables used by the outdated backup-pull(1) script in job definition files are automatically mapped to the new backup-script variables:

• host -> system
• source -> source_root
• pre_exec -> job_pre_exec
• post_exec -> job_post_exec

• 0: No error, success.
• 1: Unspecific Error!
• 2: Usage information has been shown.
• 3: Can't read job definition
• 4: PID-file exists!
• 5: Pre-exec command failed!
• 6: There have been jobs with errors!
• 7: Not all jobs were run!
• 9: Aborted (CTRL-C)!