Comments on: Progress in percents: 0 1 2 3 … http://mituzas.lt/2008/10/26/innodb-crash-recovery/ where ideas come and die Fri, 26 Feb 2010 17:15:26 +0000 hourly 1 http://wordpress.org/?v=3.0-alpha By: OurDelta - Builds for MySQL » This week in OurDelta - Vol 2 http://mituzas.lt/2008/10/26/innodb-crash-recovery/comment-page-1/#comment-188527 OurDelta - Builds for MySQL » This week in OurDelta - Vol 2 Thu, 06 Nov 2008 00:34:47 +0000 http://dammit.lt/?p=227#comment-188527 [...] the InnoDB recovery code. Surely this can be made faster without a major rewrite. Sure there can be rollbacks also and [...] [...] the InnoDB recovery code. Surely this can be made faster without a major rewrite. Sure there can be rollbacks also and [...]

]]> By: Peter Zaitsev http://mituzas.lt/2008/10/26/innodb-crash-recovery/comment-page-1/#comment-188429 Peter Zaitsev Mon, 27 Oct 2008 07:01:53 +0000 http://dammit.lt/?p=227#comment-188429 Domas, I wrote about it over a year ago http://www.mysqlperformanceblog.com/2007/07/17/innodb-recovery-is-large-buffer-pool-always-better/ And there is MySQL bug filed on it http://bugs.mysql.com/bug.php?id=29847 which is of course classified as Feature request. The workaround one can often use in this case is to REDUCE innodb_buffer_pool_size and remove innodb_flush_method=O_DIRECT so OS cache can be used for caching. This can speed up recovery 10x in some cases. Domas,

I wrote about it over a year ago
http://www.mysqlperformanceblog.com/2007/07/17/innodb-recovery-is-large-buffer-pool-always-better/

And there is MySQL bug filed on it http://bugs.mysql.com/bug.php?id=29847 which is of course classified as Feature request.

The workaround one can often use in this case is to REDUCE innodb_buffer_pool_size and remove innodb_flush_method=O_DIRECT so OS cache can be used for caching. This can speed up recovery 10x in some cases.

]]> By: Kevin Burton http://mituzas.lt/2008/10/26/innodb-crash-recovery/comment-page-1/#comment-188428 Kevin Burton Mon, 27 Oct 2008 06:20:23 +0000 http://dammit.lt/?p=227#comment-188428 Wow... that is REALLY REALLY stupid and NEEDS to be fixed. It's also not scalable across cores (obviously) which is why my InnoDB recoveries always use one core. Dumb dumb dumb. Wow… that is REALLY REALLY stupid and NEEDS to be fixed.

It’s also not scalable across cores (obviously) which is why my InnoDB recoveries always use one core.

Dumb dumb dumb.

]]>