Loopback IP Doesn't Respond Due to RPF, But Reboot Fixes It

[KrisK]

I'm getting inconsistent test results for a batch of Fortigate devices.

After a few days the Fortigate loopback IP stops responding to pings. The flow shows "reverse path check fail, drop". 

No configuration changes are made and the Fortigate is rebooted. 

The loopback IP starts responding again. The routing tables for devices that do respond look the same as devices that do not respond.

Has anyone seen this or know why a reboot would fix the issue?

[jintrah_FTNT]

Hello,

It is possibly that return route to source is available or became valid post restart. The article Technical Note: Details about FortiOS RPF (Reverse... - Fortinet Community may be of help.

Best regards,

Jin