It looks like a daemon on one of my servers got a little excited at 4am and slammed too much data into PostgreSQL, causing it to crash. There was a HUGE spike in disk I/O , then Postgres dropped into recovery mode , and seemed stuck there (by the logs, for several hours). I tried restarting Postgres to see if that would help. No luck.
I only had 4GB of free space on the machine ( and 3GB of databases ) so I had to reboot it to get a larger drive in there. I rebooted, did a filecopy of the main pg data directory, and got ready to investigate the issue...
... and then i noticed that Postgres was up and running and humming along as if nothing happened.
Should I be worried about this? This just seems weird.