http://www.pythian.com/blogs/501/mysql-set-global-innodb-fasttrue News and views from Pythian DBAs Sun, 27 Jul 2008 01:30:03 +0000 http://wordpress.org/?v=2.3.2 http://www.pythian.com/blogs/501/mysql-set-global-innodb-fasttrue#comment-66019 Innodb commits « François Schiettecatte’s Blog Thu, 21 Jun 2007 20:26:00 +0000 http://www.pythian.com/blogs/501/mysql-set-global-innodb-fasttrue#comment-66019 [...] Raj Thukral suggests setting the innodb_log_flush_at_trx_commit parameter in my.cnf to 0, thereby causing [...] […] Raj Thukral suggests setting the innodb_log_flush_at_trx_commit parameter in my.cnf to 0, thereby causing […]

]]> http://www.pythian.com/blogs/501/mysql-set-global-innodb-fasttrue#comment-65945 Xaprb Thu, 21 Jun 2007 18:11:05 +0000 http://www.pythian.com/blogs/501/mysql-set-global-innodb-fasttrue#comment-65945 A better "workaround" might be to disable AutoCommit and group logically related statements together in transactions, rather than letting InnoDB arbitrarily decide which bits of work are committed together as a unit. Unless you really are doing a bunch of rapid-fire single-statement transactions, in which case yeah, you've hit the ceiling of your commit rate and that won't help. Another good solution is to use a RAID controller with a battery-backed write cache and let it group flushes together from the cache. A better “workaround” might be to disable AutoCommit and group logically related statements together in transactions, rather than letting InnoDB arbitrarily decide which bits of work are committed together as a unit. Unless you really are doing a bunch of rapid-fire single-statement transactions, in which case yeah, you’ve hit the ceiling of your commit rate and that won’t help. Another good solution is to use a RAID controller with a battery-backed write cache and let it group flushes together from the cache.

]]>