Terminology for failover routes

[boyer]

I am looking for the cisco terms used to describe route failover between two routers.  I have a circuit through AT&T and the last time I upgraded the tech set it up so while my old router (cisco 7204VXR) was turned on and the line protocol was up my routes would point to the 7204VXR router however when I powered up the new unit (7206VXR) on the new circuit once the line protocol was up the routes would then be switched to that unit.  This was a change on the AT&T Indianapolis access router not the 7204/7206 listed above.  I tried to ask the tech when I was doing the switchover but I'm not sure he understood what I was asking.  I've thrown the question out to others at AT&T but never been able to get the answer for which I'm looking.  I am not using any sort of dynamic routing so there is no RIP(2), OSPF or EIGRP in the mix at least not on my side.  I'm getting ready to do another cut over between an old circuit and a new one and would like to be able to tell the tech "please use the Cisco Default Router Switchover Protocol and set circuit A as the primary and circuit B as the secondary".  I just don't know what the terminology for that protocol/machinery is called.  Thanks!  --Mike

[John Blakley]

It sounds like they just did a test and turn up on the circuit. If you're not using a routing protocol of any sort, but you have public addressing, the ATT tech may be using static routing on their end to prefer one path over the other.

HTH,
John

*** Please rate all useful posts ***

[boyer]

I agree, I'm just trying to nail down possible solutions to how this might be accomplished so when I talk with the tech I can be on the same page.  Looking through my cisco books I see in the static routing sections where specific routes are pointed at their gateway so traffic is routed correctly.  For example  ip route 192.168.1.0 255.255.255.0 1.1.1.1  not knowing I might guess two statments like below would operate where the first statement is used unless the gateway becomes unavailable then the second statement is used.  ip route 192.168.1.0 255.255.255.0 1.1.1.1 ip route 192.168.1.0 255.255.255.0 2.2.2.2  but I don't ever really do this sort of routing so I'm not really sure if this is how it works.  Reading further where it talks about static routes that track interfaces they talk about using the HRSP (Hot Router Standby Protocol) or VRRP (Virtual Router Redundancy Protocol).  Maybe these are common place and are what would be used or maybe the simple statements above work.  So is what I'm talking about a "failover route" or is there some other phrase that correctly describes what I'm trying to accomplish.  Thanks again.  --Mike

[John Blakley]

Failover routes that are static are called floating routes. The look like:

ip route 192.168.1.0 255.255.255.0 1.1.1.1

ip route 192.168.1.0 255.255.255.0 2.2.2.2 254

The 254 indicates that it's a lower preference than the first static route, which by default is 1. When the first route falls out of the table because 1.1.1.1 is no longer responding, the second route gets put into the routing table.

It's possible that ATT is doing that, especially if you're not running any routing protocols with them. I wouldn't consider HSRP/VRRP as a failover for their terms. They may use it, but I doubt it. These are considered next hop routing protocols (NHRP) and allow for virtual addresses to be assigned to routers. In the instance of two routers using hsrp, RouterA may have 192.168.50.2 and RouterB may have 192.168.50.3. The two routers share 192.168.50.1, and all of the clients would set their gateways to 192.168.50.1. When the active router, in this case RouterA, fails, the second router takes over as role of the primary.

I work with ATT quite a bit, and in all honesty, the techs generally are running automated applications that does everything on the back end. If they're moving around static routes to route for your network, they may have to do that stuff manually, but generally the person that you're on the phone with has about an 85% chance they won't be able to fix your issue if the automated script fails. I have had the pleasure of working with the actual engineering team in Alabama once. I brought up a circuit and I was only getting a single route, but I was expecting much more. I explained to the tech on the phone that they've put us in the incorrect vrf and it needed to be fixed. That person let me go, and the next day (which was a Saturday), ATT called and stated they were getting me in touch with engineering. The next person that called asked me what was going on, and I explained to him that I believed we were put in the wrong vrf because I was only getting a single route. Sure enough, he was the appropriate engineer to talk to because he was the one that actually managed the mpls network along with other engineers in that group. It was nice to actually talk to someone that knew what I was stating

HTH,
John

*** Please rate all useful posts ***

[boyer]

OK John, I went back through my books on floating static routes and that does sound like the correct or most appropriate solution.  I sadly know enough to be dangerous, but not what you would call good.  Its funny you say that about AT&T and messing up only getting one of your routes.  I have four routes and a couple migrations back they only got half of them migrated.  Sadly the level 1 or maybe 2 tech support people do fall into the same bracket I do, they know enough to be dangerous not good so putting two people (myself and them) together can sometimes produce mixed results.  I know during the last migration I had put on my checklist to run a traceroute to an IP on each subnet to make sure they were terminating at the new router like they were supposed to.  I want to thank you for going into the extra detail to set my mind at ease and help me narrow down my options.  --Mike

[Richard Burts]

Mike

From what I understand of your description I do not believe that either HSRP or VRRP are what you are looking for. I believe that floating static routes is a pretty good fit to your description.  In a floating static route AT&T would configure two routes, one with a next hop of your old router and the other with a next hop of your new router. On the static route pointing to your old router they would configure a distance (or weight depending on which terminology they understand) on the route to the old router that is higher than the default distance of 1 to your new router. As long as your new router is down then its static route will not be in the AT&T routing table. And when your new router comes up the static route to it will replace the static route to your old router.

This works quite well if the connections to you are some type of serial interface when your router being down will keep the line protocol down on their router. If the connection is some type of Ethernet then it may need some type of route tracking to correctly manipulate which route is used.

HTH

Rick

[boyer]

Hi Richard, I believe the floating static routes makes the most sense.  I do have serial interfaces so the routes will drop when the interface goes down so I think in my case that will work.  The O'reilly book made a point about stating if it was on an Ethernet interface it would not function the same way so its worth taking note.  My next migration will be from a DS3 to Metro Ethernet so at that point I'm guessing the above scenareo won't apply but this time it should.  Thanks again for the reply.  I can sure use all the advice I can get.

--Mike