Refer to the exhibits.

The correct answer is C . In a link-trap-based wired deployment, the switch sends a linkUp or linkDown SNMP trap to FortiNAC-F, but that trap does not contain the endpoint MAC address. After receiving the link trap, FortiNAC-F must contact the switch and perform a Layer 2 poll to read the forwarding table and determine which MAC address was added or removed on the port. The FortiNAC-F study guide states that link traps trigger FortiNAC-F to perform a Layer 2 poll to update its awareness of devices connected to the edge device, and the wired link-trap workflow specifically shows FortiNAC-F performing a Layer 2 poll before locating the host record and provisioning access.

The symptoms in the exhibit are classic stale Layer 2 visibility: FortiNAC-F still shows a rogue host on a port where no host is connected, while also failing to show hosts on ports where endpoints are actually connected. That means FortiNAC-F is not successfully refreshing the switch MAC table information. Since link traps depend on FortiNAC-F being able to poll the switch after the trap, a contact failure with the modeled switch is the most likely cause.

Option A is wrong because logical network settings affect access enforcement, not whether FortiNAC-F can see current MAC-to-port mappings. Option B is wrong because the FortiNAC-F agent is not required for basic switch-port visibility; Layer 2 visibility comes from switch polling, MAC notification traps, or RADIUS. Option D is tempting, but the broader failure shown here is not merely a policy or endpoint-side issue—it is that FortiNAC-F cannot obtain current Layer 2 data from the switch. In practice, you would still verify SNMP/CLI credentials while troubleshooting, but the best answer to the symptom pattern is that FortiNAC-F cannot contact/query the switch successfully.