Firewall Support Center

FortiGate Reverse Path Check Fail: Find and Fix the Routing Problem

Troubleshoot FortiGate reverse path check failures with source-route checks, a filtered debug flow, and a repeatable validation plan.

Branch client10.20.10.25 enters through Branch-VPN
FortiGateCheck the route back to 10.20.10.25
MismatchReturn route points to a different interface
Illustrative troubleshooting view · Firewall Support Center

The symptom and what it tells you

A firewall rule can look correct while a packet is discarded during source validation. When a FortiGate trace reports a reverse path check failure, investigate how the firewall reaches the packet’s source. Begin with the receiving interface and source address shown in the trace. The destination route alone cannot explain this result.

Understand the source-route check

FortiGate uses Reverse Path Forwarding (RPF) to evaluate whether a source is reachable through the interface receiving its packet. Feasible-path mode and strict mode apply different route criteria. Confirm the configured mode in the affected VDOM before interpreting a multihomed design. Routing decisions in another VDOM do not establish reachability for this one.

Record one failing connection

For an example branch flow, record client 10.20.10.25, server 10.30.20.50, TCP port 443, the Branch-VPN interface, and a precise test time. Confirm that 10.20.10.25 is the address actually received. An upstream NAT device can change it. Reproduce a single new connection so that the trace and packet capture refer to the same attempt.

Inspect the route and configuration

get router info routing-table details 10.20.10.25
show full-configuration system settings
show full-configuration system interface

Interpret the result

Run these commands in the relevant VDOM. Compare the source lookup with the observed ingress interface and review strict-src-check and interface src-check settings. A branch subnet learned over a different tunnel, an overlapping network, or a route withdrawn during failover can explain the mismatch. Examine the active routing table and the intended routing design together.

Capture a bounded flow trace

diagnose debug reset
diagnose debug flow filter clear
diagnose debug flow filter saddr 10.20.10.25
diagnose debug flow filter daddr 10.30.20.50
diagnose debug flow filter proto 6
diagnose debug flow filter dport 443
diagnose debug flow show function-name enable
diagnose debug flow trace start 20
diagnose debug enable

Stop and preserve the evidence

diagnose debug disable
diagnose debug flow trace stop
diagnose debug flow filter clear
diagnose debug reset

Correct the route and retest

Use the evidence to correct an incorrect route, tunnel advertisement, subnet definition, or unexpected source translation. Treat a security-check change as a separate design decision with its own review. A route repair should preserve the intended source validation. Record the original configuration and a rollback before making the selected change.

Validation and escalation

Repeat the same client-to-server test and verify the observed ingress, source lookup, policy match, and response. Check both primary and failover paths when redundancy is involved. Preserve the software build, VDOM, filtered trace, source-route lookup, and timestamp if escalation is needed. Hardware offload can limit debug visibility, so missing trace output alone does not prove that no traffic exists.

Official references

Fortinet routing concepts

Fortinet debug flow