ALTER TABLE tbl_name CONVERT TO CHARACTER SET utf8 COLLATE utf8_general_ci;

So I decided to take the bull by the horns and find out what exactly was causing the overload and how I could increase MySQL performance.

my.cnf

A good starting point is the MySQL config file on your server. Mine is located at /etc/my.cnf. It seems the default set up and example files are for 1970s server with 32mb of RAM!

For a modern server with 1gb of RAM the my-large.cnf is probably a good starting point (you can find these in /usr/share/doc/mysl-server-[versio]) .
From here there are several tweaks but perhaps the biggest improvement will be by making sure caching is turned on.

Under the [mysqld] config settings add:

query_cache_type=1

query_cache_limit=1M

query_cache_size=32M

Log Slow Queries

The best way to find out what is hogging your MySQL resources is to log slow queries. This is done by adding something like the following lines of code to your my.cnf:

log-slow-queries = slow.log
long_query_time = 2

This will write any query taking more than 2 seconds to the slow query log. I used a relative path here as absolute ones didn’t seem to work on my server. The slow log can be found in the MySQL data folder at /var/lib/mysql on my box

NOTE: you’ll need to restart MySQL for any of the above changes to take effect:

/etc/init.d/mysqld restart

Optimising Slow Queries

Armed with the information from the slow query log you can set about tweaking the queries that are slowing things up. The slow log gives us several useful bits of info besides the query itself. This includes time taken and number of rows examined – if this number is particularly high then it probably means your table’s indexes need looking at (it usually is)

The EXPLAIN Command

MySQL’s EXPLAIN command is perhaps your most useful tool in working out what indexing strategy to use. I’m only going to touch on this as it is a pretty complicated topic that I barely understand myself!

The useful info that can be gleaned from the EXPLAIN command are whether the indexes you have set up are being used for a particular query. To use the EXPLAIN command just place ‘EXPLAIN’ in front of the query you are analysing.

More to come…

What I wanted to do here was find a way of speeding up a query that was run on MySQL database on a different server. Also I’m a big fan of caching generally, especially when the data is only updated once or twice a day.

When I was running the queries directly off the remote server it was noticably slow but with this solution in place things are flying!

Whilst there are ways of caching MySQL queries using MySQL this would be pretty pointless for a remote query, the only solution was to cache the results locally.

The solution:

class_db.php by Troy Wolf

What this class does is provide an layer between your query and database server (it supports several DBs) which runs the query and then writes the results to a file in a serialized format. E.g.

a:10:{i:0;a:8:{s:13:”Dunster House”;s:10:”short_desc”;s:157:”A warm welcome to Dunster House. We are a friendly family run Guest House situated within walking distance of Torquay’s main attractions. Beds from £20″;s:4:”town”;s:7:”Torquay”;s:3:”lat”;N;s:3:”lon”;N;}….

In the class you need to hard code the details of your DB connection(s) and set up your cache directory. Then it’s as straightforward as:

include("class_db.php");

$d = new db(0); //use first connection info from class_db

$sql = "SELECT `id`
 FROM accommodation
 WHERE `type` = ".$catNo;
$q_name = 'num_'.md5($catNo;);
$rows = $d->fetch($sql, 3600, $q_name);//run query, cache for 1hr, save cahce file as $q_name

foreach($rows as $row){
...
}

Easy as that. You have now run the query on the remote server and created a cached local result which will be called until it expires.

For an example of the script in action: http://www.devon-stay.com/Ilfracombe/camping.htm