The server was completely swapped for a different server, then data was restored from the old server. There was ~1.5 hours of down time. For the afternoon, user and app streaming (notifications) were choked up by a previously-existing bug, and were not restored until the evening.
Times in CDT:
# 10:22 in, readying power
# 10:25 backup
# 10:29 shut down services
# 10:30 backup
# 10:33 swap servers
# 10:48 server in place
# 11:11 restart test
# 11:18 services live