pyxtrabackup

Percona Xtrabackup python wrapper


Keywords
mysql, mariadb, database, backup, percona, xtrabackup
License
Apache-2.0
Install
pip install pyxtrabackup==3.1.6

Documentation


Join the chat at https://gitter.im/deviantony/xtrabackup-scripts

pyxtrabackup

pyxtrabackup is a Python wrapper for the Percona Xtrabackup utility, see official homepage.

The tool allows you to take full and incremental backups using Xtrabackup.

Call for maintainers

The author deviantony currently doesn't actively use this tool and he's looking for awesome contributors to help maintain this project. Please don't hesitate to contact him directly or check the issue #40 for more information.

Installation

Use pip to install it:

$ pip install pyxtrabackup

Requirements

You'll need to install Percona Xtrabackup on your system in order to use the tool.

See: Installation documentation

Full backup and restoration

Backup

This tool can be used to create a backup of a local MySQL server. It will create a a compressed archive and move it into a timestamp named folder in the specified repository.

You will need a MySQL user with the appropriate permissions to create the backup, check the sql folder on the git repository for an example.

Example of use:

$ pyxtrabackup <PATH TO REPOSITORY> --user=<MYSQL USER> [ --password=<MYSQL PASSWORD> ]

Usage in a cron file:

@midnight    mysql    /usr/local/bin/pyxtrabackup /mnt/myrepo --user=backup-user --password=changeme

Additional options

You can also specify the following options:

  • --tmp-dir: Specify the temporary directory used by the script. (default: /tmp).
  • --log-file: Log file for the script (default: /var/log/mysql/pyxtrabackup.log).
  • --out-file: Log file for innobackupex output (default: /var/log/mysql/xtrabackup.out).
  • --backup-threads: You can specify more threads in order to backup quicker (default: 1).
  • --no-compress: Do not compress the backup archive.
  • --webhook: URL to send a POST request after the backup is finished. Will send the archive_path and archive_repository in JSON.

Restoration

The archive is containing a binary backup of a MySQL server, all you need to do in order to restore the backup is to extract the content of the archive in your MySQL datadir, setup the permissions for the files and start your server.

Note: If you are restoring a backup from another server, be sure that the server configuration of the new server is the same otherwise it could cause issues when trying to restart MySQL.

Stop the MySQL service first:

$ sudo service MySQL stop

Clean the MySQL datadir:

$ sudo rm -rf /path/to/mysql/datadir/*

If you compressed the archive, uncompress and extract it:

$ sudo tar xvpzf /path/to/backup_archive.tar.gz -C /path/to/mysql/datadir

Otherwise you just need to extract it:

$ sudo tar xvpf /path/to/backup_archive.tar -C /path/to/mysql/datadir

Then restart your MySQL server:

$ sudo service mysql start

Setup an incremental backup cycle

An incremental backup cycle is consisting of a full binary backup (the base backup) and one or more incremental backups containing only the data difference between it and the last backup.

See Wikipedia incremental backup for more information.

Backup

Use the pyxtrabackup-inc binary to create an incremental backup cycle.

First you'll need to take a full backup which will serve as the base backup: :

$ pyxtrabackup-inc <PATH TO REPOSITORY> --user=<MYSQL USER> [ --password=<MYSQL PASSWORD> ]

Now, you can start to add incremental backups to the cycle by using the --incremental option: :

$ pyxtrabackup-inc <PATH TO REPOSITORY> --incremental --user=<MYSQL USER> [ --password=<MYSQL PASSWORD> ]

NOTE: The cycle will be reset every time a base backup is created (without the --incremental option).

Additional options

You can also specify the following options:

  • --tmp-dir: Specify the temporary directory used by the script. (default: /tmp).
  • --log-file: Log file for the script (default: /var/log/mysql/pyxtrabackup-inc.log).
  • --out-file: Log file for innobackupex output (default: /var/log/mysql/xtrabackup.out).
  • --backup-threads: You can specify more threads in order to backup quicker (default: 1).
  • --no-compress: Do not compress the backup archives.

Restoration

WARNING: The folder structure and the file names created by the pyxtrabackup-inc binary needs to be respected in order to restore successfully:

  • TIMESTAMP_FOLDER/INC/base_backup_DATETIME.tar(.gz)
  • TIMESTAMP_FOLDER/INC/inc_1_backup_DATETIME.tar(.gz)
  • TIMESTAMP_FOLDER/INC/inc_N_backup_DATETIME.tar(.gz)

To restore an incremental backup, you'll need to use the pyxtrabackup-restore binary the following way: :

$ pyxtrabackup-restore --base-archive=<PATH TO BASE BACKUP> --incremental-archive=<PATH TO INCREMENTAL BACKUP> --user=<MYSQL USER>

Also, if you did use the --no-compress option with the backup tools, you'll need to specify the --uncompressed-archives option: :

$ pyxtrabackup-restore --base-archive=<PATH TO BASE BACKUP> --incremental-archive=<PATH TO INCREMENTAL BACKUP> --user=<MYSQL USER> --uncompressed-archives

The binary will stop the MySQL service, remove all files present in MySQL datadir and import all the incremental backups up to the specified last incremental backup.

For example, using the following parameters: :

$ pyxtrabackup-restore --base-archive=/tmp/repo/20140518/INC/base_backup_20140518_1700.tar.gz --incremental-archive=/tmp/repo/20140518/INC/inc_backup_5_20140518_2200.gz --user=backup-user

The script will restore the inc_N_backup_DATETIME.tar.gz from 1 to 5.

Additional options

You can also specify the following options:

  • --data-dir: MySQL datadir. (default: /var/lib/mysql)
  • --restart: Restart the MySQL service after restoration.
  • --tmp-dir: Specify the temporary directory used by the script. (default: /tmp).
  • --log-file: Log file for the script (default: /var/log/mysql/pyxtrabackup-restore.log).
  • --out-file: Log file for innobackupex output (default: /var/log/mysql/xtrabackup.out).
  • --backup-threads: You can specify more threads in order to backup quicker (default: 1).
  • --uncompressed-archives: Do not try to uncompress backup archives. Use this option if you used the backup tool with --no-compress.

Development

You can use the Dockerfile to build a development environment container with all pre-requisites: :

$ docker build -t pyxtrabackup .

Then you can use it to run the scripts: :

$ docker run --rm -it -v ${PWD}:/src pyxtrabackup zsh
$ cd /src
$ python xtrabackup/full_backup.py ...

Limitations

This tool supports the following versions of Percona Xtrabackup:

  • 2.2.x

It has been tested on the following OSes:

  • Ubuntu 12.04
  • Ubuntu 14.04

It has been tested against the following MySQL versions:

  • 5.5

It has been tested against the following Python versions:

  • Python 3.4