Hello all -

I've been having a lot of trouble with a RoS installation at one of my sites. It's running on an Axiomtek NA-510. I've had issues with it rebooting itself spontaneously sometimes, other times going unresponsive and having to be rebooted. Interestingly, I've already replaced the unit with a backup unit of the same model and the backup is behaving as badly as the first one was. Also interestingly, I'm running this same model under nearly identical circumstances in a sister facility nearby, and I've had no issues!

Today I got smart and added some more topics to the log, and also had the watchdog ping the router's default gateway to reboot it it it lost contact. It did, and it rebooted itself about an hour ago.

Looking at the log, about two hours before it rebooted and about five minutes after I added the 'route' topic to the error log, I started seeing a bunch of route topic entries - I was not logged into the router during any of this:

Aug/03/2012 12:52:06 route,debug,calc Begin calculation
Aug/03/2012 12:52:06 route,debug,calc End calculation
Aug/03/2012 12:52:06 route,debug Begin redistribution
Aug/03/2012 12:52:06 route,debug Accept withdraw x.x.x.x/xx
Aug/03/2012 12:52:06 route,debug Accept withdraw x.x.x.x/xx
Aug/03/2012 12:52:06 route,debug Accept withdraw x.x.x.x/xx
Aug/03/2012 12:52:06 route,debug Accept withdraw x.x.x.x/xx

...and some of these:

Aug/03/2012 12:52:06 route,debug Commit prefix x.x.x.x/xx
Aug/03/2012 12:52:06 route,debug Commit prefix x.x.x.x/xx
Aug/03/2012 12:52:06 route,debug Commit prefix x.x.x.x/xx

...and loads of these:

Aug/03/2012 13:01:46 route,debug,calc Select route
Aug/03/2012 13:01:46 route,debug,calc dst-address=x.x.x.x/xx
Aug/03/2012 13:01:46 route,debug,calc attributes
Aug/03/2012 13:01:46 route,debug,calc protocol=STATIC
Aug/03/2012 13:01:46 route,debug,calc distance=1
Aug/03/2012 13:01:46 route,debug,calc scope=30
Aug/03/2012 13:01:46 route,debug,calc target-scope=10
Aug/03/2012 13:01:46 route,debug,calc next-hop= address=x.x.x.x
Aug/03/2012 13:01:46 route,debug,calc need-gateway-check=2
Aug/03/2012 13:01:46 route,debug,calc comment=
Aug/03/2012 13:01:46 route,debug,calc origin-type=STATIC

...and then it starts all over again:

Aug/03/2012 13:01:46 route,debug End redistribution
Aug/03/2012 13:04:47 route,debug,calc Begin calculation
Aug/03/2012 13:04:47 route,debug,calc End calculation
Aug/03/2012 13:04:47 route,debug Begin redistribution
...etc..

This goes on for quite a while, then the watchdog reboots the system:

Aug/03/2012 15:31:59 watchdog,error,critical watchdog cannot ping address x.x.x.x, rebooting

After boot and all the stuff in initial config the logs start up again. All the nets being acted on in the above log are on a GRE tunnel that I maintain for a client. and the cycle of calculation and redistribution seems random, sometimes two minutes apart, sometimes 15 -20 minutes apart.

I've sent this and a supout to support, but I'm open to any and all ideas - could this be causing my router to lose connectivity, or is it unrelated?

Update on this: the Watchdog reset was unrelated, that was a goof on my part, and here's a tip for others: if you are going to use the ping feature of watchdog, make sure you are not constraining pings with firewall filters to the point that you start getting timeouts. This can cause the watchdog to think you've gone offline and reboot the router.

So I'll revise this post: has anyone had any stability issues with GRE tunnels, possibly causing connectivity loss or a locked up router?