Skip to main content

Efficiently backup your server – Rsnapshot

Backup are often overlooked or set in “Procrastination” mode (I’ll do it tomorrow…) but it is probably one of the most important task you need to do it, and do it right.

When a failure or hack happen, no one want to loose all its work (Hundreds articles written? Super important files? etc…).

I think we have to differentiate RAID and Backup system.

For me, a RAID system is not for backup, but for availability. If you got an issue, you can quickly rebuild the system. But if a bad copy is shared to all your drives through your RAID, well it’s too late.

The most popular way to back up a server is probably through RSYNC (file synchronization and file transfer program). Very simple to use, it will only transfer files if there is a difference (New files? modified files? Deleted files?, etc…). It can then save a lot of bandwidth, better than simply copy everything again and again.

I’ve been using Rsync for years for my backup system, but decided to change over Rsnapshot, mainly because of the lack (or not efficient) versioning. I used to create a script to keep 7 days back up of my MySQL databases and websites and an additional copy every month. But my backup folder started to explode after few months, so I had to delete the old one manually to free up some space.
Rsnapshot can actually fix it, that’s way I’ve decided to use it instead.Rsnapshot is an open source backup software written in Perl, that uses Rsync and SSH to create incremental backups.

The differences of backup are kept with hardlinks (Keep only 1 version of the file).You can set up rotative backups on a daily, weekly and monthly basis while backups can be from a local or a remote filesystem (Through SSH).

Let’s see how to use it.


Simply run in root (or with sudo)


Rsnapshot configuration

  • Local Backup

1) Create a specific folder to keep all backups

In my case, I have a SSD for the system and 1 HDD for big files. So I’ll back up all the important stuff from the SSD on my HDD.

2) Configure the /etc/rsnapshot.conf file

and configure:

– snapshot_root (Where to store your backups)

– Backup Intervals

Although the default settings are enough, you could modify this part to add a monthly backup for example.

Hourly 6, means you keep 6 copies of the hourly backup (So one every 4 hours)

Daily 7 means you keep 7 copies of the daily backup (So one every day)

You can uncomment the

(It’s not space it’s TAB between the values)

if you want monthly backup too.

– Local directories

Setup the folder your want to backup using the following syntax:

In my case I will backup my /home and /var/www.

Note that the path must finish by /

– MySQL databases

You will need the mysql-client to be installed in order to use mysqldump.

Then create a simple script in your /home/user/

and paste:

(replace the user/password)

and in the /etc/rsnapshot.conf at the localhost local directories, add:

(Still use TAB, not space, between the values)

3) Test the configuration

4) Run it

Give it a try by running

5) Set up cron tasks for the automatic back up

By default, rsnapshot comes with cron file under “/etc/cron.d/rsnapshot“ but are commented.

Simply uncomment them to enable hourly, daily, weekly and monthly backup.

The first line, will run every 4 hours the hourly backup. Rsnapshot will create the hourly directory and based on your configuration will keep “n” copies.

The second line will run daily at 3:30am and will create the daily directory.

The third will run the weekly backup every Monday at 3:00am

The forth will run the monthly backup at 2:30am.


  • Remote Backup

If you want to remotely backup another server/machine, here are the additional steps:

1) Create a SSH Password-less Login for the remote machine

On your local server, you will first need to create a RSA SSH key, simply run:

and leave the password empty.

Then, add your public SSH key to the authorized keys on your remote machine.

Still using the terminal of your local server, run:

If you have a message about the authenticity of the host, just say yes.

It will then ask you the user password you mentioned on the command. Once enter, it will add your SSH key automatically to the authorized keys.

Simply give it a try by running

2) Modify the /etc/rsnapshot.conf

– Enable remote backup over SSH:

Find and uncomment the line

– Change SSH port if needed

If your SSH port is not the default 22, you can change it at the line:

– Set the remote directories

At the end of the file you can add something similar:

to copy the remote host’s home under the folder ‘MyRemoteHost’ that will be created on your local backup folder.

If you want to backup your remote MySQL databases, similar than local one, you can create a script to dump the databases, or directly set the command inside the /etc/rsnapshot.conf such as:

As you can see, it needs to be broken in 2 parts, the first will run mysqldump on the remote host and we will use a “Unused” folder to avoid conflicts with others “used” folder like mysql in the second command. This will avoid the error:

The second command will simply copy the mysqldump.sql file to your local backup folder.

3) Give it a try

You can manually run the first backup to give it a shot:

or for some verbose, simply add -v


You should now have a working and proper backup system. Obviously you will still need to make some others copies on different supports/locations to minimize the risk of loosing your data.

If any question or unclarity, please let me know

Websync, web interface to manage your rsync tasks

Rsync is a great tool to replicate, sync some data on your computer. And I’m heavily relying on it to backup my server and to mirror some opensource projects and GNU/Linux Distributions.

But I’ve recently found a Web interface to manage all my rsync tasks called websync.

Websync is a web based rsync task manager where you can add, edit, clone, remove, scheduled,…. your rsync tasks while being able to have a remote host as source or destination of the task (With SSH RSA key too)

websync screenshot main tasks


Under the free license MIT, Websync has been developped by Sander Struijk and is still actively being maintained, as you can see on github forum. But it is still an early project, so if you face any issue, make sure to report them on the issue tracker.

Interested to give it a shot, here is how to install Websync!


You will first need to install node.js

1) Install nodejs and npm

– If you use Ubuntu, simple run (as root)

– If you are on Debian (Wheezy?) like me, you can install nodejs from wheezy-backports (modify your source.list accordingly) and run (in root):

and for npm, you can use the installer (still in root)

2) Install the global packages bower and gulp

In root, run,

3) Clone the git repository into the folder of your choice.

This time, you need to run these commands with your regular user.

4) Install Websync

If you have any missing package it will tell you. I had the error “Cannot find module ‘imagemin-gifsicle'”, I’ve simply run:

and rerun:

5) Start Websync


You should now be able to access to your Websync instance through http://YourIP:3000


sync turn signal