Fortigate AWS crash: Infinite reboot loop

Infinite reboot loop from nowhere on Fortigate AWS on demand instance


Situation:

Two FortiGate AWS on demand units (5.0.9 and 5.0.13) located in AWS (one in Frankfurt and second in Dublin) working as VPN concentrators. They connect multiple AWS accounts and one office. Routes are progpagated using BGP. Architecture described here.


Behavior


Without any particular reason both devices started to reboot. Reboot accoured every 2 minutes.

Start log:

System is starting...
fortiguard.c:761 register attr (null) errro -8
FG login: 
The system is going down NOW !!
Please stand by while rebooting the system.
Restarting system.
System is starting...
fortiguard.c:761 register attr (null) errro -8
FG login: 
The system is going down NOW !!
Please stand by while rebooting the system.
Restarting system.

CPU usage from AWS monitoring (cloud watch):



Solution

Situation was stabilized by bringing down every vpn tunnel. Tunnels were bringing up again. Tunnels were established with success as well as BGP. Both units worked properly from now on. Situation didn't repeat.

Ticket was closed as can't reproduce.
Suggetion from Fortinet Support: Upgrade to 5.2.x.