High Availability Drupal: A Clustering Case Study

SESSION OVERVIEW
In this session, we illustrate how to make Drupal highly available using only open-source software and commodity hardware, using a production cluster as our reference example. Making your Drupal site available at all times, regardless of hardware failure, scheduled maintenance, or capacity increase requires building a cluster with redundant components. Drupal 5.x has no built-in awareness of replication issues, and so your high-availability solution must appear to Drupal to be one giant machine.

TRACK
site building

AGENDA
* Hard problems: high-availability DB, high-availability /files
* High Availability and High performance: Using the redundancy required for high availability to increase performance
* Failure mode testing
* Future directions and unsolved problems

GOALS
At the end of this program you will know how to build a Drupal cluster with no single point of failure.

RESOURCES
A reasonable knowledge of linux, mysql, and apache and basic networking knowledge is assumed. Reading Scaling Drupal will give you some idea of the approach; reading some of the LVS documentation will help you understand the issues behind load balancing.