Re: Potential Forwarding Loop
"Don Provan" <[email protected]> Wed, 20 Aug 2008 16:48:41 -0700
| Newsgroups | gmane.ietf.vrrp |
|---|---|
| Message-ID | <[email protected]> |
By "accept", 6.4.3 is speaking of receiving the packets into its IP layer for IP receive handling. The point in this section is that the master should not pretend to actually *be* the node with the VR's IP address other than to answer ARPs for it. (Unless, of course, it *is* the VR IP address's owner.) This has nothing to do with accepting the packets into the system to begin with and considering them for forwarding, although I suppose I can see why you didn't find that clear. But you can read these two sections as saying, "do not accept packets addressed to VR addresses you do not own either for local reception or for forwarding." The thought processes that led to the two observations don't really intersect, so I think that's why they two issues are raised separately and without any relation even though, as you're pointing out, they're really two parts of the same concept: ignore packets addressed to a VR IP address you don't own. I hope that clears things up, because I'm going to ignore your question about the forwarding loop: it's a somewhat obscure point, anyway, and it is sufficient to understand that forwarding the VR IP address owner's packets is a bad idea. Is there some reason you want to do it, anyway, or are your just interested in the clarification? -don > -----Original Message----- > From: [email protected] [mailto:[email protected]]On Behalf Of > Chiu Hsia Peng > Sent: Wednesday, August 20, 2008 5:33 AM > To: [email protected] > Subject: [VRRP] Potential Forwarding Loop > > > Hi, > The following is stated in the RFC 3768. > > 8.4. Potential Forwarding Loop > > A VRRP router SHOULD not forward packets addressed to the IP > Address(es) it becomes Master for if it is not the owner. Forwarding > these packets would result in unnecessary traffic. Also in the case > of LANs that receive packets they transmit (e.g., token ring) this > can result in a forwarding loop that is only terminated when the IP > TTL expires. > > ... > I am confused about the above description. Since 6.4.3 already > describe: A master, .. MUST NOT accept packets addressed to the IP > address(es) associated with the virtual router if it is not the IP > address owner. I think if a VRRP router doesn't accept a packet, it > can be implied that the router doesn't forward it either. > > What's the purpose of this section? > Could someone give me a concrete example about the potential > forwarding loop? > > Thanks. > _______________________________________________ > vrrp mailing list > [email protected] > https://www.ietf.org/mailman/listinfo/vrrp _______________________________________________ vrrp mailing list [email protected] https://www.ietf.org/mailman/listinfo/vrrp