https://github.com/r3c/arshive
Simple rotation backup tool written in pure shell
https://github.com/r3c/arshive
archive backup rotate shell
Last synced: 5 months ago
JSON representation
Simple rotation backup tool written in pure shell
- Host: GitHub
- URL: https://github.com/r3c/arshive
- Owner: r3c
- License: mit
- Created: 2012-02-23T19:22:15.000Z (over 14 years ago)
- Default Branch: master
- Last Pushed: 2026-01-16T13:59:26.000Z (5 months ago)
- Last Synced: 2026-01-17T04:22:18.720Z (5 months ago)
- Topics: archive, backup, rotate, shell
- Language: Shell
- Homepage:
- Size: 43 KB
- Stars: 1
- Watchers: 1
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: readme.md
- License: license.md
Awesome Lists containing this project
README
# Arshive backup script
[](https://github.com/r3c/arshive/actions/workflows/verify.yml)
[](https://opensource.org/licenses/MIT)
## Overview
Arshive is a simple backup tool written in pure shell script. It is designed to
perform periodic snapshots of any software data and rotate them to keep only
the last N archives. Being written as a shell script, Arshive can be added to
your crontab without any extra dependency.
## How to use
Download [latest release](https://github.com/r3c/arshive/releases/latest) or
clone repository, then place contents of the `src` directory anywhere in your
file system e.g. in `/opt/arshive/`:
```sh
mkdir -p /opt/arshive &&
wget -qO - http://github.com/r3c/arshive/tarball/master/ |
tar xvz --strip-components=2 --wildcards -C /opt/arshive '*/src/'
```
Open and edit configuration file `arshive.conf` to declare one or more rule
file(s) (see `rules` option) defining backup rules. Other options can be
modified in this file to tweak the script behavior.
Open and edit rule file(s) (see sample `arshive.rule` file) to define how, when
and for how long backup files should be created. Each rule file can contain
one or more backup task declaration(s) to define how backup files should be
created and maintained.
Here is a sample rule declaring a backup task named "web-data" creating a
gzipped tar archive of everything within `/var/www` every two hour, and keeping
the three most recent backups created that way:
```sh
web-data: cd /var/www && tar -czf {.tgz} *
interval=7200
keep=3
```
Run `/your/directory/archive.sh -q` every 5 minutes or so from your crontab.
The "-q" flag will suppress any non-error output so the tool will take care of
executing any due backup task and warn about anything that went bad.
## Rule options
Following rule options can be used to tweak a backup task:
- `interval=N`: create a backup only if latest one is more than N seconds old
or if no backup was successfully created yet. Default value is 86400 (1 day).
- `keep=N`: keep the last N successfully created backup files and delete older
ones. Default value is 7.
- `max_size=N`: log a warning if size of created backup file is larger than N
bytes.
- `max_size_ratio=N`: log a warning if size of created backup file is more than
N times larger than previous backup size. N can be a decimal number, default
value is 10.0.
- `min_size=N`: log a warning if size of created backup file is smaller than N
bytes.
- `min_size_ratio=N`: log a warning if size of created backup file is less than
N times smaller than previous backup size. N can be a decimal number, default
value is 0.1.
## Resource
- Contact: v.github.com+arshive [at] mirari [dot] fr
- License: [license.md](license.md)