by Jonathan Clarke and Nicolas Charles
After a brief presentation of configuration management (CM) basics, we'll start with an ill-fated tale from the recent past about disaster recovery (also known as a case study, if you must): how our CM saved us, how it didn't, and what could have been done better. This could lead to a discussion about best practices around beers after the session.
We use Cfengine 3, and will introduce the software, overview the main differences with other open source CM tools before explaining why we like this choice. But Cfengine is not all: what enables us to manage our configuration completely are the practices and tools we've built around it. We will describe these in detail, probably offering a live demo (if Murphy permits).