The postmortem is pretty weak and us, geeks, would have loved to see more detail. Tumblr would have gotten some good karma with a detailed explanation. Well, unless this is all made up just to look impressive and the real reason was something else like human error that wiped the production DB. But even in that case honesty would have paid off - remember GitHub's recent DB wipe and their excellent explanation?
On a side note - anyone know what DB they are using? The cynic in me is thinking "Hey, another MongoDB + FourSquare 'success' story of webscale awesomeness."
Agreed. I'd like to see more detail. C.f. facebook's recent post-mortem (http://www.facebook.com/notes/facebook-engineering/more-deta...). Lots of detail. When a downtime blog post is so vague like Tumblr's was, my first instinct is that there's something they are deliberately not telling about what went wrong. Or maybe they haven't figured out exactly why it went down yet. I hope a more detailed postmortem is coming.
On a side note - anyone know what DB they are using? The cynic in me is thinking "Hey, another MongoDB + FourSquare 'success' story of webscale awesomeness."