[c-nsp] Weird Frame-relay / L3 connectivity problem

Peter Rathlev peter at rathlev.dk
Mon Apr 12 13:09:13 EDT 2010


On Sat, 2010-04-10 at 21:49 -0700, Kenny Sallee wrote:
> Here's the problem - over the last couple days - at random - we loose L3
> connectivity ONLY - can't ping, telnet, BGP drops etc...
[...]
> Here's the weird part - the MPLS provider fixes it by removing the interface
> / vrf config on their side and adding it back!  This has happened 3-4 times
> over the last couple of days.  The provider side router is a Cisco router -
> not sure on model or version of code.
> 
> Has anyone seen ANYTHING like this?  Any ideas on what to look for?  So far
> our customer has been pretty pissed off so priority has been on bringing the
> interface back up (which means only a little time to troubleshoot).

I've seen something a bit like this, though that was with Catalyst 3550s
as ethernet CPE devices. When running 12.2(25)SEC and 12.2(25)SEE EMI
they would sometimes stop forwarding traffic on one or more VRF-lite
interfaces (all SVIs). There were no error messages, traffic just
stopped flowing. Only resolution was to delete all configuration related
to the VRF and reapply.

It only happened when making configuration changes; adding or deleting
other SVIs or VRFs could trigger the problem.

What's the providers take on this? The problem initially appears to be
in their end according to what you write.

-- 
Peter




More information about the cisco-nsp mailing list