> From the RFC excerpt you gave, I would infer that on DAD failure of a global > address, the router doesn't install the address, and forwards all traffic > to/from that network based on the contents of its forwarding table. It does > no special filtering of packets that are sourced from or destined to that > network. > > Do you agree? > > Thanks, > Peter
This sounds to me like a recipe for a hard-to-debug problem. I assume that when this DAD failure happens you'd have to stop sending RAs and responding to RSs (or ...?). So the only way the router should be seeing packets destined to this interface would be either some misconfiguration OR a cached RA/RS response where the max/valid lifetimes had not yet expired AND a neighbor re-discovery had not yet been performed. I haven't computed how long this could in theory persist. But again, regardless of any technical correctness, it sounds like the making of a weird debugging problem. -------------------------------------------------------------------- IETF IPv6 working group mailing list ipv6@ietf.org Administrative Requests: https://www.ietf.org/mailman/listinfo/ipv6 --------------------------------------------------------------------