A method for providing split-brain avoidance for sub-secondary High Availability (HA) systems involves identifying the network segments that are shared by two nodes participating in HA, running the periodic network-path and next hop monitoring protocol over all these network segments, and doing the protocol over-ride to clearly detect and distinguish network and interface faults from node and service faults in HA systems. This ability is used by the Standby node to provide the split-brain avoidance without requiring any hardware bus interconnect between HA nodes or presence of any external quorum server.