Techniques for managing disaster recovery sites are disclosed. In one particular embodiment, the techniques may be realized as a method for managing disaster recovery sites comprising generating a heartbeat at a first node, transmitting the heartbeat from the first node to a second node, determining whether a network connection between the first node and the second node has failed, determining whether the second node has received an additional heartbeat from the first node, and changing a state of the secondary node based on the determination of whether the second node has received the additional heartbeat.