A database drive failure causes at least a partial (and possibly a complete) Sync outage between 05:21 to 05:41 (PST) this morning.
On this hardware platform (we've only seen it on this single type of hardware), this certain style of drive failure causes the database to be unable to write to disk, which causes the web servers to hang as connections get "stuck" waiting on the database.
We have experienced this exact problem before and Services Operations has been discussing options for fixing this flaw in the architecture. The immediate goal is to limit impact to users on the failed database only. We will have something worked out before FF4 launch.
-jv