How to sleep without the server-crash-fear (High Availability and High Performance Hosting for Drupal)
Hosting Drupal for a small and medium site is not very hard. But what if the Drupal Site is business criticality for the customer? Or the servers should handle thousands of Users at the same time? With just duplicating the Webserver it isn't done, the problems will just begin. Of course you could be awake 24/7 hours to keep the server alive, but is this really what a system admin want to do? Well not me.
But building a High Performance Site isn't only setting up the right server software, it is also a lot of brainwork in the beginning. Like using Node Teaser Displays is really cool for the themer, but can be deathly for your servers.
I will show how we at Amazee Labs have setup our Drupal Cluster with multiple Webservers, Varnish, Memcache, MySQL Replication, GlusterFS and more. You will get some insights about the common problems like big sites (more than 10000 Nodes) or a lot of authenticated users bring with, how we handle these problems, and which tools we are using.
I will also include some outlook to the future of High Performance with Drupal.