provided your tie-breaker (let it be a lock disk or a quorum server) is configured and working properly, single node can hold the cluster up if the other one fails for a reason or gets deliberately shutdown. As a matter of fact it is one of the advantages of having a cluster, as in, you do maintenance on the inactive node and reboot if necessary, while the application runs on the other node.
Having said that, when was the last time you performed a failover test ? Was there anything major happening to your data center, like a storage array change, networking reconfiguration etc (major event) ?
Make sure to check the cluster/package log and syslog for any recent cluster related messages before committing such an act.
Normally, nothing should go wrong, but living in the trenches for the past 15+ years, I know that there is nothing normal about any data center.
________________________________
UNIX because I majored in cryptology...