Archive for October, 2012

Oh for the Wait Interface, part 3

October 31, 2012

To repeat myself (from “part 2”):

“a while back I wrote a post expressing a longing for wait-interface tuning”.

The original post, and “part 2” were both written in 2007, while I was working for a startup in Pasadena, as a DBA.  Now I work for Oracle, supporting MySQL, and by now, MySQL has incorporated a wait interface.  So, to do justice, I am writing this “part 3” and saying – it’s been done!

MySQL now has the performance schema, which instruments a vast number of events and waits.  The man page for the current GA version (5.5) is at http://dev.mysql.com/doc/refman/5.5/en/performance-schema.html .   It’s a big chapter, as the wait interface is highly configurable and versatile.

http://mysql.wisborg.dk/2012/09/03/what-is-the-mysql-performance-schema-and-why-is-it-needed/ gives a quick overview about the performance schema (“P_S”).  To get an idea of the wide variety of things instrumented, see, for example, http://www.markleith.co.uk/2012/07/24/a-mysql-replication-load-average-with-performance-schema/
and
https://blogs.oracle.com/svetasmirnova/entry/performance_schema_in_version_5 .

Because there is so much that can be configured with the performance schema (enabling the tracking of various events and waits), and because the queries to get the information out can be complicated, you might also want to check out Mark Leith’s collection of useful examples at http://www.markleith.co.uk/ps_helper/ .

So, there you go, now we have it!

(And now I have three posts on the subject, each with the heading capitalized a little differently, ’cause that’s the way I roll…)