Sun on Apr 13, 2010 Note: this article was originally published on http://blogs.innodb.com on April 13, 2010 by Inaam Rana. One of the well known and much written about complaint regarding InnoDB recovery is that it does not scale well on high-end systems. Well, not any more. In InnoDB plugin 1.0.7 (which is GA) and plugin 1.1 (which is part of MySQL 5.5.4) this issue has been addressed. Two major improvements, apart from some other minor tweaks, have been made to the recovery code. In this post I’ll explain these issues and the our solution for these. So at time of crash we had: Modified db pages 1007907 Redo bytes: 3050455773 And the recovery times were: Plugin 1.0.7 (also Plugin 1.1): 1m52s scan, 12m04s apply, total 13m56s Plugin 1.0.6: 31m39s scan, 7h06m21s apply, total 7h38m 1.0.7 (and Plugin 1.1) is better 16.95x on scan, 35.33x on apply, 32.87x overall https://blogs.oracle.com/MySQLinnodb/entry/innodb_recovery_is_now_faster