Be wary of locks
When MySQL updates a table, it performs what is called a "lock" - it ensures that no other program can read or write from the table while it updates. As soon as MySQL has finished writing, it allows access again. So, it is best not to try to perform complicated SELECT statements (reads) on tables which are updated regularly, because your SELECT statement will stall.
Next chapter: Consider switching OS >>
Consider splitting off variable-length fields
Performance Optimising your code Write your code sensibly Use your tools wisely Avoid functions for maximum performance Use the Zend Optimizer Use a PHP code cache Read the manual carefully Get your loops right first Pre-increment where possible Don't think that using references will lower your RAM usage Be wary of garbage collection, part 1 Be wary of garbage collection, part 2 Listen to all errors, big and small Keep up to date Cache array data Compress your output Don't use CGI Don't use dl() Debug your code Cache your pages Use persistent connections Take advantage of new features Compile right Ignore the myths Avoid mod_access if you can Don't fret about Apache 2.0 Optimising your SQL Prioritise your data Optimise your tables Select as little data as possible Use shorter queries where possible Use the EXPLAIN statement Change your hardware Choose your data types carefully Size vs. Speed Declare fields NOT NULL Load data intelligently Consider splitting off variable-length fields Be wary of locks Consider switching OS Spot slow queries Perform joins carefully Index your data Make sure your indexes are being used Normalisation Upgrade MySQL Increase your buffers Keep tabs on your server Lock your tables when appropriate Don't rely on automatic type conversion Benchmark, benchmark, benchmark! Know MySQL's strengths Server PHP Accelerators Caching PHP PHP as a CGI or a module? Size of scripts Summary Exercises Further reading Next chapter
Table of Contents
Follow us on
Identi.ca or Twitter