[dayglored]

No offense to FReepers but this is a crap report. There is no such thing as a "Master Code". It's actually a very simple, but huge, error.

Something screwed up in a BGP router update. BGP is how the internet finds the Facebook servers (including their DNS nameservers).

If you lose or misconfigure BGP, it's like the whole service/site has disappeared.

But once it's corrected, it all comes back. So just be patient.

[CTyank]

Exactly. Wiping out petabytes of storage would almost certainly be an inside job and would take a lot of time by someone with high level admin authority. Data loss if it occurred would be detected
and stopped. Mirrored servers are off line - only storage is mirrored, and that storage is likely to be backed up as well. At worst there may be 24 hours of some data that would be lost.

Its an infrastructure issue, a nasty one, but fixable.

[Paperpusher]

But once it’s corrected, it all comes back. So just be patient.

Darn!

[no-s]

hmm, I wonder if there are exploits around ASNs possible when something like this happens? Looks like they’re back in business now so it was not a long outage....