post/mid mortem on 7+ hours of beeminder downtime

Site went down at 1:35am pacific last night due to massive hardware
failure. Some frustrating miscommunication with Linode support dragged
it out for a long time. Finally decided to redeploy to a fresh Linode
server. Looked dicey but now seems all data’s totally intact. Worst
case for a while there looked like a restore from 24-hour-old backup,
with possible painstaking reconstruction from logs.

As broadcast on all the social medias:
Blow-by-blow of server woes – and much gnashing of mandibles at the
beehive – ongoing at twitter.com/beemstat

On the bright side, check out the adorable “poppy bee” at
beeminder.com right now.


http://dreev.es – search://"Daniel Reeves"
Goal tracking + Commitment contracts == http://beeminder.com