Sometimes I suspect we get lost in forest, and all we can see are the trees. Let's look at this from the perspective of how data is moved from here to there. Comments below:
""Kim Graham"" wrote in message [EMAIL PROTECTED]">news:[EMAIL PROTECTED]... > This brings up a question. I understand that after the initial "hi I will > be handling your requests please use me as your destination mac address". > (Router talking to client). > > But what happens when the initial router fails and HSRP kicks in? After an > unreachable, would ClientA send out an arp or would RouterB initiate the > arping CL: The ARP process is used by any host ( router or PC or other workstation ) when it has data for a particular host at a particular IP address. The host knows through the XOR process that the destination host is on th same subnet. Since devices on the same subnet are operating at the L2 layer, a MAC is required. The host says, essentially "I have data for network address. What MAC should I use?" and the appropriate host replies "use this one - I'm that IP address, here is my MAC address" CL: So in the case you state, there is no reason for Router B to do anything. It does not have data to transmit to host A. >to re-establish connections to any client that was using RouterA > after it noticed that RouterA was not responding? > > Scenario: > > > ClientA ----- RouterA/B(HSRP) ------ ClientB > > ClientA sends a packet to ClientB > ClientA talks to the Virtual RouterA/B -- RouterA/B sends to ClientB CL: Not exactly. The router that is the HSRP primary does all the talking to host A. > RouterA/B tells ClientA -- RouterA will be handling your requests. CL: not exactly. The HSRP primary device, using the virtual IP/MAC, does all the communication at this point. there is no provision for a process as you describe. Well, maybe proxy ARP falls into this kind of category, but that's different. > RouterA/B tells ClientB -- RouterA will be handling your requests to ClientA > ClientA then sends more packets to ClientB via RouterA. CL: sure, in practical terms. But host A is still sending packets to the virtual IP/ virtual MAC address, not to physical addresses. > ClientB responds to ClientA via RouterA. > > Janitor comes in and accidentally unplugs RouterA's power cord. > > ClientA now has to re-establish a connection with ClientB. CL: well, in theory, host A never knows that a failover has occured. So far as host A is concerned, it is still communicating with the physical device whose IP and MAC are those that it learned at the beginning of tis process. that is, the virtual IP/MAC > > I have seen the above scenario happen in a failover test when implementing a > new core but did not have a bug in my ear to watch the MAC addresses. It > has my curiosity perked. > > In theory I beleive RouterB would re-establish communication after a failed > "hi are you there packet to RouterA". I will have to wait until a lab is > set up to play out the scenario. CL: what you should find is that from the host perspective, nothing changes. I don't have sniffer experience, but I would hazzard the guess that your sniffer traces will see no changes to source and destination IP's, and no change to source and destination MACs. I base this upon my understanding of the process of how a host sends packets. A more detailed look at the theory may be found in Comer's Internetworking with TCP/IP volume 1. CL: My point being that the rules of host to host communication do not split off into a zillion different special cases every time some fix or other is introduced. HSRP is based on the router side, and is designed specifically to keep things simple and consistent as far as the hosts on the particular segment are concerned. Packets move from host to host using the same rules and processes every time. These rules don't change just because there is an HSRP router pair on the segment. they do not change just because there is an OSPF virtual link somewhere along the line. They do not change just because you are on dial backup, rather than the primary WAN link. It becomes far easier to understand when you start from the fundamental principal, and move outwards, than if you get lost in the maze of looking at everything as a special case. CL: sorry for the soap box. over the past few days there have been several threads which have indicated to me that certain fundamentals are not understood. > > Kim Message Posted at: http://www.groupstudy.com/form/read.php?f=7&i=47248&t=47177 -------------------------------------------------- FAQ, list archives, and subscription info: http://www.groupstudy.com/list/cisco.html Report misconduct and Nondisclosure violations to [EMAIL PROTECTED]