Imagine that you have a cluster with two machines side by side in an active/standby configuration. Let’s say you have your data replicated, and the systems are basically identical except for the IP address and hostname. You can use heartbeat to share an IP address such that if the primary fails, the secondary takes over. [...]

Linux magazine released an article today by Joe Brockmeier titled Rethinking Gmail: Reliability Matters. The article makes some good points, and makes an obvious statement that to some, email is a mission critical application. I don’t dispute the points. I’d like to discuss why these systems fail to begin with, and how as an end [...]