Magento Database Maintenance

Magento does many things well, but maintaining an efficient database is not one of them. Having a lot of products is a good reason to have a large database. Sadly, this isn't the only way your database can become big and sluggish. Maintaining the database through log cleaning can result in a dramatic improvement in site performance and latency. This guide explains how to clean up a Magento database that has grown too large for its own good.

• NOTE: Before performing any operations on the database, please make a backup.

Magento maintains several tables for logging. These tables log things such as customer accesses and which products have been compared. Magento has a mechanism for cleaning these logs regularly, but unfortunately this feature is disabled by default and most customers do not turn it on. There are three ways to clean out these tables: via Log Cleaning in the Magento Admin, via log.php in the ../shell directory, and manually via phpMyAdmin or mysql client.

The following tables are managed by Magento's Log Cleaning function:

log_customer
log_visitor
log_visitor_info
log_url
log_url_info
log_quote
report_viewed_product_index
report_compared_product_index
report_event
catalog_compare_item

1. In the Magento Admin, go to System > Configuration.
2. In the left menu under Advanced click on System.
3. Under "Log Cleaning", change "Enable Log Cleaning" to YES and configure the Save Log for 15 days:
   
4. Click "Save Config"

The shell utility log.php can be configured as a cron job or run manually to clean on-the-fly.

1. From the Magento root directory, type:
   

   php -f shell/log.php clean

• the –days switch can be used to specify how many days back to save.

This is the most efficient way to clean the logs for those more comfortable working with databases. It is faster than the built-in Magento tools, and it allows you to clean a couple of other tables not included in those tools.

1. Open the database in PhpMyAdmin via the Siteworx Control Panel.
2. In the right (main) frame, select the check box for the following tables:
   

   dataflow_batch_export
   dataflow_batch_import
   log_customer
   log_quote
   log_summary
   log_summary_type
   log_url
   log_url_info
   log_visitor
   log_visitor_info
   log_visitor_online
   report_viewed_product_index
   report_compared_product_index
   report_event

   
   

3. At the bottom of the page, click the drop-down box that says "With Selected:" and select "Empty".
4. A confirmation screen will appear. Click "Yes". This will truncate all of the selected tables.

We've seen 2GB+ databases drop to a quarter of their size after cleaning the logs, so it is very important that this sort of maintenance is performed regularly. Particularly if your time-to-first-byte latency starts growing larger, and you've implemented the other performance tweaks.