Sorry about the last few hours of downtime. The site was down due to a combination of technical factors that just happened at precisely the wrong time. Practically, the datacenter was running automated reboots for diagnostic purposes just when the kernel was being upgraded on the server. The upgrade process didn't get a chance to finish updating the bootloader configuration, which either corrupted it or resulted in a misconfiguration (still looking into this). Due to this, the server got kicked into Linux Single-User mode on reboot, and was just stuck like that. It should be OK now. In case you were wondering, this was not in any way related to the incident yesterday with Blake's account being compromised.