Personal schedule for Padraig O'Sullivan
subscribe to Padraig O'Sullivan's
This tutorial focuses on the diagnostic tools you can turn to when confronted with an unexplained performance problem. Tools explained include iostat, vmstat, and mk-query-digest. The emphasis is on practical usage under pressure, and less on performance theory.
The Drizzle Project is a fork of the MySQL 6.0 server. One of the many goals of Drizzle is to enable a large plugin ecosystem by improving, simplifying, and modernizing the application programming interfaces between the kernel and the modules providing services for Drizzle. This tutorial serves to showcase the new APIs for Drizzle's replication through a series of in-depth examples.
MySQL's EXPLAIN output is rich with information, but can be difficult to understand. This session helps EXPLAIN make sense by showing how it reveals the server's estimated query execution plan.
Many people know Linux terminologies such as ext3, tmpfs, cfq io scheduler, OOM killer, etc. But many times it is not appropriately configured. In this session, the speaker will show Linux performance tuning and stabilization practices for MySQL, such as how to avoid OOM Killer, performance difference between I/O schedulers, tuning Linux kernel parameters, profiling with SystemTap, etc.
One of the key elements of database expertise is understanding how
queries are executed. This talk gives a detailed and consistent
overview of query execution in MariaDB/MySQL covering both algorithms,
code, internal and external representations of query plans. Deeper
understanding of this area should benefit both current and future
server and storage engine developers, and DBAs.
In this session, the speaker will explain some important but not well known indexing topics. Advanced topics such as indexes and lock contentions, how indexes impact MySQL optimizer will be covered. This session is the sequel to "Mastering the Art of Indexing" (http://www.mysqlconf.com/mysql2009/public/schedule/detail/6661), presented at the last MySQL Conference.
Facebook runs MySQL on commodity servers. Commodity servers have become extremely fast. Learn about the methods we use to identify performance bottlenecks in MySQL and the projects we have completed or started to keep pace with hardware advances.
This session shows you how to get rich historical data about your MySQL server's performance and usage with the open-source Cacti graphing tool. Best of all, it's easy, open-source, and free -- so you can stop flying blind.
Vapour is a server that speaks the memcached protocol and uses MySQL Cluster (NDB) as the storage backend.