DHCP Server Stops Working

I have 3 FVS318Ns at endpoints at various locations.  They are usually very solid, needing a reboot maybe twice a year.

Recently, one of them needed a reboot because the vpn tunnel would not come up.  No biggie I thought.  I rebooted like normal and expected another 6 months of uptime.  But it went down in a few hours.  Rebooted again--up for a few hours then down again.  Rebooted again, and that's when the problem started.

The unit acted bricked.  No wan led, no wan connection.  None of the systems were issued IPs from the units dhcp server.  I put in a replacement router to keep the site alive and left with the bad unit.

After two weeks, I powered up the unit and confirmed that no dhcp IPs were being issued.  I did not try the dmz port which I had assigned to a different vlan with a different dhcp server.  I did not try setting a static IP on my system and trying to access the router that way.

As a last resort, I defaulted the router.  The dhcp server came back to life with default settings.  I was catching a flight to the same site that had this unit in production, and planned to re-configure from scratch there.

As I was entering the dhcp reservations again, the dhcp server stopped working again.  This time, I tried every port until I got an IP on the dmz port set to vlan 2 from its own dhcp server.  I logged back into the router and examined the dhcp log.  

I found entries that mentioned 'duplicate entry' with an IP address matching a reservation I set as well as a message about downloading the newest version of the server and asking support questions on the list serv and to not contact the developers directly.  At the top of the log was the word 'exiting'.  This looked like a dump from the dhcp server so I examined the 'duplicate entry' that it was referring to and deleted that reservation.  Once I did this, the dhcp server started working again.  

As soon as I added back the entry, it would give the same error message.  I skipped the entry and added another one and everything was fine.  But when I entered another entry I'd get the same messages with that reservation's IP address in the error message.  There was no logical order to the error messages--they didn't occur on specific IP addresses or at particular points in time or entry 'slots'.  I had to delete any reservations that had error messages because the server would stop working until I deleted them.  Then I could continue adding others until the same problem occurred.

Eventually, I was able to add all the reservations back and everything was functioning normally again.  Then I had to reconfigure tunnels, wireless, and more.  Total time to reconfigure: 3.5hrs.

My question is what caused this unit to suddenly have problems like this?  And would this type of issue be hardware-related and covered under the warranty?

Feel free to ask questions for details, but keep in mind that I had to get the unit working so I don't have any copies of the logs or anything like that.