||8 years ago|
|client||8 years ago|
|server||8 years ago|
|.gitignore||9 years ago|
|README.md||8 years ago|
|install.sh||8 years ago|
|uninstall.sh||8 years ago|
How it works
rs-backup-suite is designed for push backups, which means the client pushes its files to the server. This is ideal for computers which are not always on such as most desktop PCs.
It is also a user-centric backup system. That means each user creates his own backup on the NAS instead of root backing up the whole machine at once (although this is possible). That also means that each user has a UNIX account on the NAS. The NAS username is usually
<hostname>-<local user name> (e.g.
On the client machine(s) each user can create a file called
.rs-backup-include (name is configurable) inside his home directory which includes the list of files that should be considered by the backup. Additionally root can maintain a similar file located at
/usr/local/etc/rs-backup/include-files for the system files.
Setup (please read this carefully before performing any actions!)
rs-backup-suite is split into two parts: a client part for pushing the backup to the NAS and a server part which runs on the NAS itself.
For installing the server component run
sudo ./install.sh server
on your server machine. This installs all the necessary files into the right location on your system.
Tweaking the configuration file
If you need to tweak the server settings, simply edit
/etc/rs-backup/server-config to your needs. There you can configure the following directives:
BACKUP_ROOT: The directory under which the home directories of the backup users are stored. The default is
FILES_DIR: The directory under which the actual backups are kept (relative to the backup user's home directory). The default is
SET_QUOTA: Whether to set disk quota for the users or not (for Ext3/4 file systems). Default is
QUOTA_INODE_HARD_LIMIT: The individual limits for disk quota. Ignored, if
WARNING: Adjust these settings before you create backup users, because they won't be re-applied for already existing users!
Adding a backup user
A backup user is an unprivileged UNIX account on the server. Normally each user on each client has one corresponding backup user which he uses to log into the NAS. A backup user can be created by running
rs-add-user hostname username [ssh-public-key-file]
on the server where
hostname is the name of the client host and
username is the name of the user on that machine for whom this account is made. Of course you can use any other names for
username as well, but it's generally a good idea to stick to this naming convention. The resulting UNIX username will be the combination of both.
The optional third parameter specifies the path to the SSH public key file which the user will use to log into the NAS. If you don't specify it, the user won't be able to log in at all. But you can add one later at any time by running
rs-add-ssh-key hostname username ssh-public-key-file
username are the same as above and mandatory for identifying the user that should get the new key.
TIP: If you don't remember the parameters for all these commands, simply run them without any and you'll get simple usage instructions.
Making the chroot work
rs-backup-suite can chroot backup users into the backup home base directory. For this to work you need to add a few lines to your
/etc/fstab and run
mount -a afterwards (replace
/bkp with your backup path):
# Chroot /bin /bkp/bin none bind 0 0 /lib /bkp/lib none bind 0 0 /usr/bin /bkp/usr/bin none bind 0 0 /usr/lib /bkp/usr/lib none bind 0 0 /usr/share/perl5 /bkp/usr/share/perl5 none bind 0 0 /dev /bkp/dev none bind 0 0
NOTE: In Ubuntu the Perl modules are located at
/usr/share/perl instead of
/usr/share/perl5. Change that accordingly. Also note that if you are using Synology DSM, you also need to add a bind mount for /opt/bin to /bkp/opt/bin.
If your 64-bit system doesn't have a
/lib folder but only
/lib64 you may need to add this to your
/lib64 /bkp/lib64 none bind 0 0
/lib is symlinked to
Finally add this to the end of your
Match Group backup ChrootDirectory /bkp/
and restart OpenSSH. Your backup users are now chrooted into
NOTE: When using a chroot environment and you change anything in your user configuration (e.g. the username) you need to run
rs-update-passwd or your user might not be able to log in anymore.
Changing the rotation options/backup levels
To change how many increments of which level are kept, edit the file
/bkp/etc/rsnapshot.global.conf. This is the global configuration file for rsnapshot which will be included in each user-specific configuration. There you can tweak the names and numbers for all backup levels.
If you add or remove any backup levels, make sure you also update the cron scripts. By default three cron scripts are installed:
To set up the client you simply need to run
sudo ./install.sh client
on your client machine. Then copy the file
/etc/rs-backup/client-config, edit it as root and replace the value of
REMOTE_HOST with the hostname or IP address of your NAS.
On the client machines the script
/usr/bin/rs-backup-run is used for performing the backups. This script can either be run as root or as an unprivileged user. The behavior differs in both cases:
- If run as root, all files and folder specified in /etc/rs-backup/include-files
will be backed up. The backup user used for logging into the NAS ishostname-root
by default (wherehostname
is the hostname of the current machine). Additionally the home directories of all users will be scanned. If a home directory contains a file called.rs-backup-include
all files and folders specified inside that file will be backed up under this user's privileges. The username used for logging into the NAS ishostname-username
is again substituted for the hostname of the current machine andusername` for the user whose home directory is being backed up).
- If run as a normal user, only the files that are specified in your own
.rs-backup-includewill be backed up.
Changing the default configuration
All the client configuration options are defined in
/etc/rs-backup/client-config. You can edit the file as you wish. All parameters are documented clearly by comments. Most of these configuration options can also be overridden at runtime by passing command line arguments to
rs-backup-run. For a list and a description of all possible command line arguments run
Installing client and server on the same machine
You can of course also install server and client on the same machine. This may be useful if you want, e.g. save your data to an external USB drive instead of a real NAS. A shortcut for running both
sudo make server-install and
sudo make client-install is simply running
sudo ./install all
For uninstalling run ./uninstall.sh [all|server|client]. These remove all the scripts but preserve the data in
/bkp (or whatever your backup folder is).
The intended use case for rs-backup-suite is as follows: you set up the server part on your NAS. Then you create a backup user for each user on each client machine.
In the next step you edit the crontab for root on each client and add a job for running
/usr/bin/rs-backup-run at certain times. You can of course also create a shell script that calls
rs-backup-run and put it in
/etc/cron.daily to perform a global backup once a day.
After everything is set up that way you create the file
/etc/rs-backup/include-file and write to it a list of files and folders you want to back up as root (e.g. you can specify
/etc/*** to backup the whole
/etc directory and all its subdirectories). Furthermore each user creates a file called
.rs-backup-include inside his home directory that serves the same purpose for his own home directory instead of the global system. Such a file could look like this:
- /home/johndoe/.cache/*** /home /home/johndoe/***
Lines that start with a
- are treated as excludes, all other lines as includes. The three asterisks mean “Include this directory and everything below”. For more information about these globbing patterns read the FILTER RULES section of the rsync(1) man page.
NOTE: To include a directory you need to mark all parent directories for inclusion, too. For instance to include
/home/johndoe you also need to include
/home as shown above. But don't confuse
/home without the trailing slash only selects the (empty) directory itself, not its contents.
Restoring files from the NAS
To restore files from the NAS server simply run:
rsync -a -e ssh backupuser@remotehost::pull/source/path /destination/path
backupuser with the proper backup user (e.g.
remotehost with the hostname of the NAS.
/source/path is the file name on the remote side (e.g.
/destination/path is the local destination file name.
You can also log into the NAS using SFTP or SSHFS. This is probably more convenient for browsing available files.
Be aware that both access methods are strictly read-only! Write access is only granted via rsync through the
rsync -a -e ssh backupuser@remotehost::push/destination/path /source/path
Because rs-backup-suite uses rsync for the client-server communication you don't necessarily need both parts. As long as you have a working rsync server on your NAS you can use the client script to push files to it. On the other hand you can use the rs-backup-suite server part with any other rsync client, as well.
Warning to users of older versions
rs-backup used to reside in
/usr/local instead of
/usr. With the addition of a proper Makefile in version 0.2.0 this has changed. The consequence is that older setups won't work with the new version without modifications. In order to update your setup you need to update the path to
rs-run-ssh-cmd (now at
/usr/bin/rs-run-ssh-cmd) inside your users'
~/.ssh/authorized_keys files as well as the path to
/usr/bin/rs-rotate) inside their
rsync.conf files. Alternatively just create symlinks to the old locations.
/usr also means that for chroot setups the
/bkp/usr/local mountpoint is no longer needed.