Since my first post on MySQL transaction descriptors optimization introduced in Percona Server 5.5.30-30.2 and a followup by Dimitri Kravchuk, we have received a large number of questions on why the benchmark results in both posts look rather different. We were curious as well, so we tried to answer that question by retrying benchmarks on [...]
Adventures in archiving
February 11, 2013 By 16 Comments
One of our Remote DBA service clients recently had an issue with size on disk for a particular table; in short this table was some 25 million rows of application audit data with an on disk size of 345GB recorded solely for the purposes of debugging which may or may not occur. Faced with the task of [...]
Multiple purge threads in Percona Server 5.1.56 and MySQL 5.6.2
May 3, 2011 By 13 Comments
Part of the InnoDB duties, being an MVCC-implementing storage engine, is to get rid of–purge–the old versions of the records as they become obsolete. In MySQL 5.1 this is done by the master InnoDB thread. Since then, InnoDB has been moving towards the parallelized purge: in MySQL 5.5 there is an option to have a [...]
Blog Redesign 2011
January 16, 2011 By 9 Comments
After more than four years of using a theme that can charitably be described as rustic, we’ve updated our look. We’re using the News theme from StudioPress, which is a bargain; their themes and theme framework are an engineering marvel. We’re still using the amazing Sphinx search plugin from Ivinco for fast and relevant searches. [...]
Different flavors of InnoDB flushing
January 13, 2011 By 9 Comments
In my recent benchmarks, such as this one about the Virident TachIon card, I used different values for innodb_buffer_pool_size, like 13GB, 52GB, and 144GB, for testing the tpcc-mysql database with size 100G. This was needed in order to test different memory/dataset size ratios. But why is it important, and how does it affect how InnoDB works [...]
MySQL 5.5.8 – in search of stability
January 3, 2011 By 11 Comments
A couple of days ago, Dimitri published a blog post, Analyzing Percona’s TPCC-like Workload on MySQL 5.5, which was a response to my post, MySQL 5.5.8 and Percona Server: being adaptive. I will refer to Dimitri’s article as article [1]. As always, Dimitri has provided a very detailed and thoughtful article, and I strongly recommend reading if [...]
MySQL Partitioning – can save you or kill you
December 11, 2010 By 44 Comments
I wanted for a while to write about using MySQL Partitioning for Performance Optimization and I just got a relevant customer case to illustrate it. First you need to understand how partitions work internally. Partitions are on the low level are separate table. This means when you’re doing lookup by partitioned key you will look [...]
How (not) to find unused indexes
October 16, 2009 By 2 Comments
I’ve seen a few people link to an INFORMATION_SCHEMA query to be able to find any indexes that have low cardinality, in an effort to find out what indexes should be removed. This method is flawed – here’s the first reason why:
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 | CREATE TABLE `sales` ( `id` int(11) NOT NULL AUTO_INCREMENT, `customer_id` int(11) DEFAULT NULL, `status` enum('archived','active') DEFAULT NULL, PRIMARY KEY (`id`), KEY `status` (`status`) ) ENGINE=MyISAM AUTO_INCREMENT=65691 DEFAULT CHARSET=latin1; mysql> SELECT count(*), status FROM sales GROUP by status; +----------+---------+ | count(*) | status | +----------+---------+ |   65536 | archived | |     154 | active | +----------+---------+ 2 rows in set (0.17 sec) mysql> EXPLAIN SELECT * FROM sales WHERE status='active'; # query 1 +----+-------------+-------+------+---------------+--------+---------+-------+------+-------------+ | id | select_type | table | type | possible_keys | key   | key_len | ref  | rows | Extra      | +----+-------------+-------+------+---------------+--------+---------+-------+------+-------------+ | 1 | SIMPLE     | sales | ref | status       | status | 2      | const | 196 | Using where | +----+-------------+-------+------+---------------+--------+---------+-------+------+-------------+ 1 row in set (0.06 sec) mysql> EXPLAIN SELECT * FROM sales WHERE status='archived'; # query 2 +----+-------------+-------+------+---------------+------+---------+------+-------+-------------+ | id | select_type | table | type | possible_keys | key | key_len | ref | rows | Extra      | +----+-------------+-------+------+---------------+------+---------+------+-------+-------------+ | 1 | SIMPLE     | sales | ALL | status       | NULL | NULL   | NULL | 65690 | Using where | +----+-------------+-------+------+---------------+------+---------+------+-------+-------------+ 1 row in set (0.01 sec) |
The cardinality of status index is woeful, but provided that the application [...]
Should you move from MyISAM to Innodb ?
January 12, 2009 By 59 Comments
There is significant portion of customers which are still using MyISAM when they come to us, so one of the big questions is when it is feasible to move to Innodb and when staying on MyISAM is preferred ? I generally prefer to see Innodb as the main storage engine because it makes life much [...]
High-Performance Click Analysis with MySQL
December 22, 2008 By 13 Comments
We have a lot of customers who do click analysis, site analytics, search engine marketing, online advertising, user behavior analysis, and many similar types of work. The first thing these have in common is that they’re generally some kind of loggable event. The next characteristic of a lot of these systems (real or planned) is [...]
