As well as waking up feeling even worse, I heard a quick blip from the stack of computers next to me that sounds an awful lot like a click of death. ReiserFS was reporting inode read issues, so I checked the volume group, already knowing the answer. Luckily this volume group was just a backup stripe, but it still sucks. On the upside, this is my first disk failure in a while, so that’s good. Even better? This is my last IBM drive (and yes, it’s from the infamous Death Star production run).
Also, linux takes a really long time to shut down when hardware is failing. I need to figure out from an experienced admin if there’s a way to halt under disk failure conditions that’s faster than “trust in the filesystem journal and pull the plug.”