EmergencyEMERGENCY? Get 24/7 Help Now!

Announcing Percona XtraBackup 2.0.0 GA

 | April 4, 2012 |  Posted In: MySQL


I’m really excited to today announce the first GA (Generally Available; i.e. stable) release of Percona XtraBackup 2.0. We have worked hard since our last major release on improving the reliability and user experience of Percona XtraBackup as well as adding features to make it better suited to more environments.

The 2.0.0 release contains no changes from the last beta (1.9.2), only the version number has changed.

New since 1.6:

  • Percona XtraBackup can now save Galera replication information while performing a backup when given the --galera-info option to innobackupex.
  • XtraBackup now supports compressed backups. These backups can be done in a parallel way, thus utilizing multiple CPU cores if needed. In previous versions, compression was only possible with streaming backups + external (usually single-threaded) compression utilities, which also had a number of other limitations (e.g. could not be used with parallel file copying, it was required to uncompress the entire backup to restore a single table).
  • Percona XtraBackup now supports streaming incremental backups. In previous versions streaming backups were performed by the innobackupex script but incremental backups were done by the xtrabackup binary which calculated deltas by scanning data files. Which meant those two feature were mutually exclusive, i.e. one couldn’t do streaming incremental backups.
  • As part of the backup, the LRU dump is now included as well.
  • Support for compiling and running Percona XtraBackup against debug versions of InnoDB. This is only for very advanced users.

The new features in Percona XtraBackup will be part of the topics at the upcoming Percona Live MySQL Conference and Expo (register now!). Specifically, there are the following sessions at the conference on Percona XtraBackup and tools around it:

You can download Percona XtraBackup 2.0.0 from our downloads site or install it using our YUM or APT repositories.

Stewart Smith

Stewart Smith has a deep background in database internals including MySQL, MySQL Cluster, Drizzle, InnoDB and HailDB. he is also one of the founding core developers of the Drizzle database server. He served at Percona from 2011-2014. He is a former Percona employee.


  • Does XtraBackup take advantage of a filesystem that can take consistent read-only snapshots of the database?

    I’m considering placing my databases on a ploop filesystem (a new tool in OpenVZ). Seems to me you could uses ploop snapshots to bring the time the database is locked (or unavailable to clients) to a very small period. I’m also looking into upgrading to your cluster release that uses Galera and I think that if you had support for ploop snapshots, the introduction of a new master into the cluster wouldn’t affect the donor master as much. I haven’t used any of these tools yet so I could be wrong, but it seems to me that ploop support might help.

  • With xtrabackup, it only has to lock tables that aren’t InnoDB, it has the ability to take backups of InnoDB tables *without* taking locks and without any overhead of block device or filesystem snapshots.

  • sir , why there is only –compress method but no –uncompress for xtrabackup2.0 … , is there any plan in future to add uncompress method contained in xtrabackup binary ?

  • Sam,

    –compress uses the qpress (.qp) archive format to store compressed files. So you can use the qpress utility to uncompress backups. qpress is a fast, multi-threaded and open source compression utility, it is available from http://www.quicklz.com/

  • @rucypli

    Correct, you have to uncompress the backup with qpress before preparing it with –apply-log.

  • @Stewart Smith

    Hmm.. I got an error when trying to run an incremental after a compressed base backup. The error was that it was not able to find the xtrabackup_checkpoints file. The base dir had the xtrabackup_checkpoints.qp instead..


  • I ran into the issue with xtrabackup_checkpoints.qp as well. The only way for me to run an incremental on top of a compressed full backup (or a compressed incremental) is to extract the xtrabackup_checkpoints file first. Seems like this is a pretty easy fix, but the bug hasn’t been picked up by anyone and has been sitting there for 2 weeks.


    This combined with the inability to create streamed tar backups has made 2.0.0 a fairly painful experience for me.

Leave a Reply