Backup Incrementals On Mysql

Home » CentOS » Backup Incrementals On Mysql
CentOS 5 Comments

Hi all.

I usally make backups of databases mysql.

I make buckups of all datbase for example:

mysqldump -u user -ppassword name_db > backups.sql

also I make backups just its schema for example

mysqldump -u user -ppassword name_db –no-data > backups.sql

but now I need a backups incrementals, because the size of DB is very big
(500 mb)

How to make this?

regards

—
*Atte. Rodrigo Pichiñual N.*
*Ingeniero Administrador de Sistemas Linux*
*rodrigo.pichinual@gmail.com *
*+56 9 87272971*
*@Roodrigo0461*

*http://cl.linkedin.com/in/rodrigopichinual
*

5 thoughts on - Backup Incrementals On Mysql

  • I’ve seen somewhere script that does the following. It dumps all databases, then commits them into some version control system. CVS and subversion come to my mind. As the last do diff of text files 9and this way keep changes from version to version), and database dump is ASCII
    text, this will fulfill pretty well what you need. I must confess, I never came to the point of setting it that way myself: with database sizes we have, and the space we can devote we can handle a week worth of daily full dumps, and a couple of Months of backups of these… but one day I’ll do it this way. The following may not be what I originally saw, but seems to be doing exactly this:

    http://rocketmodule.com/blog/database-backup-dump-and-svn-commit-script-drupal-workflow

    Valeri

    ++++++++++++++++++++++++++++++++++++++++
    Valeri Galtsev Sr System Administrator Department of Astronomy and Astrophysics Kavli Institute for Cosmological Physics University of Chicago Phone: 773-702-4247
    ++++++++++++++++++++++++++++++++++++++++

  • Hola,

    If size’s a concern, just export your DBs gziped :

    mysqldump -u user database | gzip > backup_database.sql

    Also if you are concerned about time to compress, you can enable multithreaded parallel gzip compress with pigz (available in EPEL) :

    mysqldump -u user database | pigz > backup_database.sql

    TBH, 500 MB databases aren’t big enough to seek a more complex approach (unless you have more ambitious RTO/RPO requirements)

    HTH

  • I would stay away from compression. Compression results in binary file. Even though compressed result is smaller, when you will try to keep several versions back, you will have multiple compressed file versions hitting your backup storage. The original poster’s intent is better: to keep diff of ASCII dump files. If one commits dumps into some version control system, even though its files may be treated as binary, you will need only one latest version on backup (as it contains all versions of database).

    Just my $0.02.

    Valeri

    ++++++++++++++++++++++++++++++++++++++++
    Valeri Galtsev Sr System Administrator Department of Astronomy and Astrophysics Kavli Institute for Cosmological Physics University of Chicago Phone: 773-702-4247
    ++++++++++++++++++++++++++++++++++++++++