Thanks Acee for the details. My concern was also in regards that, in a BFD-less setup during helper-mode, OSPF hello packets from the neighbor are not honored at all and is not used as a heartbeat mechanism, and hence waits for the entire grace period before announcing the peer down. So any BFD implementation done on software (may be multi-hop peers if applicable), this may still cause BFD to cause remote inactivity notification, with data plane still intact.
Regards, Anil On Tue, Aug 25, 2015 at 8:23 PM, Acee Lindem (acee) <a...@cisco.com> wrote: > Hi Anil, > > OSPF and OSPFv3 graceful restart pre-dated BFD so this wasn’t explicitly > covered. However, given that the intension is that the data plane is > preserved during restart, an implementation could interpret this as a > topology change and terminate helper mode as documented in section 3.2 of > RFC 3623. > Hope this helps, > Acee > > > > From: OSPF <ospf-boun...@ietf.org> on behalf of Anil Raj < > anilra...@gmail.com> > Date: Tuesday, August 25, 2015 at 10:19 AM > To: OSPF WG List <ospf@ietf.org> > Subject: [OSPF] OSPF GR and BFD operability > > Hi, > > I need clarification on the behavior when BFD notifies remote inactivity > for a OSPF session, when OSPF neighbor is undergoing Graceful restart? OSPF > router in helper mode will ideally inactivate the restarting neighbor only > after grace period, and if BFD notifies neighbor down to OSPF, should the > adjacency be terminated? If not, will it cause a blackhole for the entire > grace period if the neighbor is actually down? > > Appreciate if you can help here. > > Regards, > Anil > >
_______________________________________________ OSPF mailing list OSPF@ietf.org https://www.ietf.org/mailman/listinfo/ospf
