Skip to content

niol/sya

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

62 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Sya

Sya is a very basic frontend to borg. Its goal is to provide easy management of multiple independant backup tasks.

Overview

borg is a deduplicating backup program supporting encryption. A repository contains archives which may refer to multiple machine being backed up or multiple backups going back in time of the same machine, or a mix of both.

sya is a script aimed at easing the management of backup jobs. It is a basic frontend to the borg command line. It consists of the main script and a configuration directory containing a main config file and additional elements.

Configuration

General

The configuration directory (default is /etc/sya) contains the main configuration file config. IT follows an INI-style syntax.

sya section

This section makes it easy to configure general runtime items.

[sya]
verbose = true

Backup task section

The backup task section contains a backup task definition. The following configuration values are accepted in a task file :

  • run_this : This globally enables or disables the task.
  • repository : the path to the repository to backup to. Prefix with host:
  • ssh_key : the path to the private key that should be used to authenticate over SSH.
  • include_file : a full path (or relative to the configuration directory) to a file that lists what paths to include in the backup.
  • exclude_file : files to exclude from the backup. See borg patterns.
  • pre : execute this in the bash shell before executing the backup task.
  • post : execute this after the backup task.
  • keep-daily : how many daily archives to keep when pruning.
  • keep-weekly : how many weekly archoves to keep when pruning.
  • keep-monthly : how many monthly archives to keep when pruning.
  • passphrase_file : a file containing on the first line the passphrase used to encrypt the backup repository (borg init -e repokey)
  • remote-path : the path to the borg executable on the remote machine.

Example task section :

[example]
run_this = true # enable this task
#pre = "mount -o remount,rw /backu"p
#post = "mount -o remount,ro /backup"
remote-path = "/usr/local/bin/borg1/borg"
keep-daily = 30 # keep 30 daily archives
keep-weekly = 8 # keep 8 weekly archives
keep-monthly = 8 # keep 8 monthly archives
repository = "/tmp/local.borg"
include_file = "tobackup.include"
passphrase_file = "example.passphrase"
#exclude_file = "local.exclude"

Example exclude file /etc/sya/local.exclude :

/var
/proc
/bin
*.log

Global pre and post scripts

There is also the possibility of running global pre and post scripts, i.e. scripts that run before and after the task batch. Those are to put in the configuration directory under the names of pre.sh and post.sh.

Example pre.sh :

#!/bin/bash
# This script is run before the backup task batch begins.
# Feel free to customize it

# Dump mysql data to a file
mysqldump --user root --password=pass --all-databases > /var/backups/mysqldata base.dump

Usage

sya accepts the following command-line options :

-h, --help            show this help message and exit
-v, --verbose         Be verbose and print stats.
-p, --progress        Show progress.
-c, --check           Perform a repository check for consistency.
-d CONFDIR, --config-dir=CONFDIR
                        Configuration directory, default is /etc/sya.
-t TASK, --task=TASK  Task to run, default is all.
-n, --dry-run         Do not run backup, don't act.

Installation (short)

sya may run in any directory, but it was made to be run in the following way :

  • Create the directory /etc/sya and config files.
  • Drop the sya script in a cron directory if applicable, for example on Debian, I dropped it in /etc/cron.daily to make it run once a day. After a couple of days, you may want to put verbose = False in the config file to stop getting cron report emails.

Installation (my setup)

This section describes my setup. The goal is to backup from ripley to backup1 and backup2.

Password-less SSH logins

The first thing to do is to setup password-less SSH logins for the root user to the hosts you want to backup to (or the hosts you want to backup from, but I'll describe here how to use remote backup hosts to backup to). First create a key pair for this purpose.

# cd
# mkdir .ssh
# chmod 700 .ssh
# ssh-keygen
(no passphrase)

If, like me, you have aliases for your hosts and special ports, you may want to use a .ssh/config file :

Host backup1
User backupuser
Port 400
HostName backupfool.foo1.org

Host backup2
User backupuser
Port 400
HostName backupguy.foo2.org

Now, copy your IDs to the remote hosts (repeat for backup2) :

# ssh-copy-id -i ~/.ssh/id_rsa.pub backupuser@backup1
Password:

Now you should be able to login with the local root user on localhost:

# ssh backup1
(no password prompted)
[email protected] $

You can also make this key restricted to borg backups using an SSH forced command, see borg documentation.

sya configuration

So, here is the plan : I want to backup several selected directories to a dedicated /backup partition and I want to backup only a part of those selected directories to the remote hosts.

I run a mySQL server on ripley. Thus, I use /etc/sya/pre.sh to dump the database contents to a file :

#!/bin/bash
mysqldump --user root --password=unbreakable_password --all-databases > /var/backups/mysqldatabase.dump

Then I described /etc/sya/local.include which is a list of the things on ripley that I want to backup :

/etc
- **tmp
/home/niol
/root

Basically, I want to backup /etc, no files that end with tmp, my personal directory.

But when backing up to a remot host, I want to omit some files (some content that is too big) which I described in /etc/sya/remote.exclude:

/var/fileserv/

This was the most difficult part because you need to know what is on your computer. But now, the only remaining thing is to describe sya tasks.

/etc/sya/config :

[sya]
verbose = true

[local]
run_this = true
keep-daily = 30
keep-weekly = 8
keep-monthly = 8
repository = "/backup/local.borg"
include_file = "tobackup.include"
exclude_file = "local.exclude"

[backup1]
run_this = true
keep-daily = 10
keep-weekly = 2
keep-monthly = 1
repository = "backup1:/home/niol/fromripley.borg/"
include_file = "tobackup.include"
passphrase_file = "backup1.passphrase"

[backup2]
run_this = true
keep-daily = 10
keep-weekly = 2
keep-monthly = 1
repository = "backup2:/home/niol/fromripley.borg/"
include_file = "tobackup.include"
passphrase_file = "backup2.passphrase"

This should be self explainatory and this should show what sya is about.

Setting up the repositories

borg needs the repositories to be initialized:

# borg init -e none /backup/local.borg
# borg init -e repokey backup1:/home/niol/fromripley.borg
# borg init -e repokey backup2:/home/niol/fromripley.borg

Running the script

The first run should be a bit long :

# /etc/cron.daily/sya -v -p

From now on, you can check on the remote hosts which archives are available:

# borg list backup1:/home/niol/fromripley.borg
ripley-2016-09-25_05:11:26           Sun, 2016-09-25 05:11:28
ripley-2016-09-29_04:31:02           Thu, 2016-09-29 04:31:05
ripley-2016-09-30_04:29:02           Fri, 2016-09-30 04:29:04
ripley-2016-10-01_10:30:38           Sat, 2016-10-01 10:30:39
ripley-2016-10-02_04:28:56           Sun, 2016-10-02 04:28:58
ripley-2016-10-03_04:45:06           Mon, 2016-10-03 04:45:08
ripley-2016-10-04_04:30:18           Tue, 2016-10-04 04:30:21
ripley-2016-10-05_04:30:20           Wed, 2016-10-05 04:30:23
ripley-2016-10-06_04:36:09           Thu, 2016-10-06 04:36:13
ripley-2016-10-07_04:32:25           Fri, 2016-10-07 04:32:29
ripley-2016-10-08_04:34:33           Sat, 2016-10-08 04:34:36

And which files are available in thos archives:

# borg list backup1:/home/niol/fromripley.borg::ripley-2016-09-25_05:11:26 | less

And recover an old file:

# borg extract backup1:/home/niol/fromripley.borg::ripley-2016-09-25_05:11:26 etc/hosts

Tips: migrating a remote backup by taking advantage of deduplication

If you want to migrate a remote backup to this solution, deduplication can really help, for instance from an rdiff-backup backup located in fromripley to a borg repository located in fromripley.borg/:

(if not done before) $ borg init -e repokey fromripley.borg
user@backup1$ borg create -v -s -p fromripley.borg/::1stbackup fromripley/* --exclude fromripley/rdiff-backup-data

Synchronizing chunks cache...
Archives: 0, w/ cached Idx: 0, w/ outdated Idx: 0, w/o cached Idx: 0.
Done.
------------------------------------------------------------------------------  
Archive name: 1stbackup
Archive fingerprint: 6457fd3b13f37b9803c615f9cd4a96e43389d615ec2bfe292ea38d39a25c36c9
Time (start): Sat, 2016-09-24 20:52:04
Time (end):   Sat, 2016-09-24 21:36:26
Duration: 44 minutes 21.45 seconds
Number of files: 56608
------------------------------------------------------------------------------
                       Original size      Compressed size    Deduplicated size
This archive:               42.82 GB             42.82 GB             41.18 GB
All archives:               42.82 GB             42.82 GB             41.18 GB

                       Unique chunks         Total chunks
Chunk index:                   58201                69187
------------------------------------------------------------------------------
user@backup1$ rm -r .cache/borg
user@backup1$ rm -r fromripley 

Then, you won't need to transfer those 40Gib on the WAN because known chunks from your import of the rdiff-backup data will be deduplicated:

root@ripley # /etc/cron.daily/sya -t backup1 -v
------------------------------------------------------------------------------
Archive name: ripley-2016-09-25_04:28:12
Archive fingerprint: d0e1a9ce562fa90f34e68a2d8a0c78fb34a94626fe1f6569dc0f00e35c58c59b
Time (start): Sun, 2016-09-25 04:28:30
Time (end):   Sun, 2016-09-25 05:10:58
Duration: 42 minutes 28.58 seconds
Number of files: 62390
------------------------------------------------------------------------------
                    Original size      Compressed size    Deduplicated size
This archive:               43.10 GB             43.10 GB            313.95 MB
All archives:               85.92 GB             85.93 GB             41.50 GB

                    Unique chunks         Total chunks
Chunk index:                   63868               144211
------------------------------------------------------------------------------