Comments on: Warning: upgrade MySQL without testing at your own risk http://www.xaprb.com/blog/2009/08/08/warning-upgrade-mysql-without-testing-at-your-own-risk/ Stay curious! Mon, 22 Mar 2010 09:10:39 +0000 http://wordpress.org/?v=2.9.2 hourly 1 By: Xaprb http://www.xaprb.com/blog/2009/08/08/warning-upgrade-mysql-without-testing-at-your-own-risk/#comment-16746 Xaprb Tue, 11 Aug 2009 01:38:24 +0000 http://www.xaprb.com/blog/?p=1202#comment-16746 What Sheeri said: you can collect all your queries with mk-query-digest :-) There's a code snippet in mk-upgrade's SYNOPSIS that illustrates. What Sheeri said: you can collect all your queries with mk-query-digest :-) There’s a code snippet in mk-upgrade’s SYNOPSIS that illustrates.

]]> By: Xaprb http://www.xaprb.com/blog/2009/08/08/warning-upgrade-mysql-without-testing-at-your-own-risk/#comment-16745 Xaprb Tue, 11 Aug 2009 01:37:36 +0000 http://www.xaprb.com/blog/?p=1202#comment-16745 We actually have a plan for non-deterministic queries! We'll rewrite them to be deterministic. I don't think fingerprinting explain will help with non-optimizer problems, no. But query timing will. That was the first thing we implemented. The 5.1 regressions I've seen in the real world so far are optimizer-related and will show up in EXPLAIN. We actually have a plan for non-deterministic queries! We’ll rewrite them to be deterministic.

I don’t think fingerprinting explain will help with non-optimizer problems, no. But query timing will. That was the first thing we implemented.

The 5.1 regressions I’ve seen in the real world so far are optimizer-related and will show up in EXPLAIN.

]]> By: Morgan Tocker http://www.xaprb.com/blog/2009/08/08/warning-upgrade-mysql-without-testing-at-your-own-risk/#comment-16744 Morgan Tocker Mon, 10 Aug 2009 23:36:49 +0000 http://www.xaprb.com/blog/?p=1202#comment-16744 Non deterministic queries have got to be hard to catch (order by rand() etc). Do you think fingerprinting explain commands will help catch non-optimizer plan regressions? I'm thinking of how bad the performance degradation of parser was in early 5.0 releases (it got better). Non deterministic queries have got to be hard to catch (order by rand() etc).

Do you think fingerprinting explain commands will help catch non-optimizer plan regressions? I’m thinking of how bad the performance degradation of parser was in early 5.0 releases (it got better).

]]> By: Sheeri K. Cabral http://www.xaprb.com/blog/2009/08/08/warning-upgrade-mysql-without-testing-at-your-own-risk/#comment-16740 Sheeri K. Cabral Mon, 10 Aug 2009 17:00:02 +0000 http://www.xaprb.com/blog/?p=1202#comment-16740 You forgot to mention that you can collect all your queries with mk-query-digest. :) You forgot to mention that you can collect all your queries with mk-query-digest. :)

]]>