Support Forum
The Forums are a place to find answers on a range of Fortinet products from peers and product experts.
Cheeks194th
New Contributor

Looping Restart

Hi everyone. We utilize mainly Fortigate 50b units within our company. For the last couple of years we have been having a weird issue to which I can find no answer. Our units are plugged into an UPS to protect it against power outages. There are times though, due to unplanned building maintence, weather, that the UPS does not last long enough to keep the unit powered. If the router looses power, when it regains it, it gets stuck in an endless loop of rebooting. The only way I have found to fix it is to reformat and upload the configuration files again. Any insight into this would be appreciated. They are running 4.0.3,build0106,090616 If you need any other information let me know. And yes...I waited this long before posting to find my solution.
4 REPLIES 4
ede_pfau
SuperUser
SuperUser

Hi, sounds like the flash memory is corrupted during power outage. Maybe the unit draws (blind) current over the LAN cable shield as it' s not grounded anymore via mains (which terminates at the UPS not the wall socket). Just a wild guess but that would definitely cause trouble for the FGT and that might corrupt the flash. You could test this by setting up a FGT powered by a UPS but without any TP cables connected. Then disconnect the UPS' s mains and let it run on battery for a moment. Then switch off the UPS simulating Battery Depleted. Hopefully this time the FGT won' t crash. If it does then...the flash itself might be faulty. If out of warranty you can replace it yourself, it' s a socketed CF card. Keep us posted, please.

Ede

"Kernel panic: Aiee, killing interrupt handler!"
Ede"Kernel panic: Aiee, killing interrupt handler!"
Kenundrum
Contributor III

I have the exact same problem with 60Bs. It seems as if losing power at the wrong moment will cause flash corruption. Thankfully it' s almost always fixable through a reformat and TFTP of the firmware. Most of my units are not connected to a UPS. I haven' t been able to determine what set of circumstances cause the failure, so I pretty much go with the assumption that no matter what, whenever a fortigate is not gracefully rebooted, there is a small chance that the flash will corrupt and will result in a kernel panic endless reboot cycle. Thankfully it seems to happen very infrequently definitely less than 5% of the time. Over the past 3 years with 50 units, I' ve had it happen about 10 times- but my firewalls are forcefully rebooted way more often than I' d like due to the environments they are deployed in. I have had one unit actually get damaged flash due to this problem- so it' s not always fixable.

CISSP, NSE4

 

CISSP, NSE4
emnoc
Esteemed Contributor III

Don' t these FGT devices run from memory? I have probably over the couse of 5-6 years , never gracefully shutdown a fortigate, and never once had a corrupted CF memory. Maybe I' ve been lucky

PCNSE 

NSE 

StrongSwan  

PCNSE NSE StrongSwan
vanc
New Contributor II

Logging to the CF card definitely makes its life shorter. The flash memory cells have very limited write cycles. Better set up a syslog server (a Linux box, or a Windows box with simple syslog daemon). I am currently running 3 FGTs and don' t remember having corrupted flash.
Labels
Top Kudoed Authors