News and Commentary
LiveJournal Power-loss post-mortem
23 Jan 2005
LiveJournal has a great write-up of what went wrong when their power failed last week. The article discusses exactly why it took so long to recover from the power problem and is a fantastic look at disaster recovery of a large-scale web application.
It’s not often you can find this sort of detailed account of the technology behind a web site of this scale. I’m reading it carefully to learn from their experience.