Routing Issues after a Link Flap - UDP

[jjendreas1]

I have a reocurring issue that I was looking to get a bit of guidance on.  Here's the issue - I have two pieces of hardware that use a proprietary protocol for a heartbeat between devices (usinga specific port).  The packets are a simple UDP packet, and is used as a sort of ping.   These appliances are on different networks, usually with an MPLS or a VPN.  The issue occurs when that link goes down and comes back up.

All traffic looks to be routing good, with the exception of the UDP ping packets between the appliances.  This causes them to think the far-side is offline.  Our resolution has been to restart the core routers on the network.  We also have had success with other L3 equipment (mainly firewalls) by flushing any stateful connection tracking between the two enpoints through the device (or in the case of a specific firewall vendor, we create a blackhole route as a backup route if the link fails).

We are looking for a solution that would not require rebooting the core router/L3 switch.   My thought would be to disable fast-switching and see if that has an effect, or if there is away to clear any cached routes. 

Thanks

JJ

[Mark Malone · Accepted solution]

The only thing I could think is to use an ip sla udp script with an EEM script monitoring the addresses by port number and when it fails set the next action to clear conn udp or clear the router table clear ip route for those specific addresses

just an option might work , you could disable fast switching for that interface as a test but I wouldn't disable it on the switch globally

http://www.cisco.com/c/en/us/td/docs/ios-xml/ios/ipsla/configuration/15-mt/sla-15-mt-book/sla_udp_echo.html

http://mccltd.net/blog/?p=1741