I don't think it's that they "left it running" like you would leave a backup app running... they literally left the entire disaster-recovery site up and running and live. Cermak (referred to in the article as the "backup") is an entire datacenter, hosting a running copy of the exchange to be used in a failover scenario.

You'd have to have more than 1 person involved to forget that DR is still active when completing these failover exercises and tests off-hours.