Using a backup interface with static routes?

Answered Question
Mar 4th, 2008
User Badges:

Hello all,


The primary connections to our remote routers are point-to-point frame relay circuits. We have ISDN connections that will act as a backup, in case the primary fails.


My question is, what is the best backup method when there is no dynamic routing between the routers?


I would like to use the backup interface command, but I dont think that will work with static routing.


The floating static routes is a possibility, but it has its pros and cons that I dont like. I was hoping to use the backup interface utility, but I do not think that is possible with static routing.


Thanks,


Correct Answer by Richard Burts about 9 years 2 months ago

Dave


I agree with most of what Victor says here. I would like to clarify one point: you commented and Victor sort of agreed that it would work only if there were a dynamic routing protocol. This is not a correct assumption. It can also work quite well when there are only static routes. The key concept is the floating static route has an administrative distance higher than the default of 1. With its administrative distance configured the floating static route works fine with dynamic routing protocols and with static routes also. Whatever the primary route is (whether from dynamic or from static) its administrative distance is lower than the floating static and is placed into the routing table. When the primary route fails and is withdrawn from the routing table then the floating static is placed into the routing table. This would allow it to send traffic over your ISDN connection.


HTH


Rick

  • 1
  • 2
  • 3
  • 4
  • 5
Overall Rating: 4.3 (3 ratings)
Loading.
lamav Tue, 03/04/2008 - 14:25
User Badges:
  • Blue, 1500 points or more

"The floating static routes is a possibility, but it has its pros and cons that I dont like. I was hoping to use the backup interface utility, but I do not think that is possible with static routing."


D:


You're correct. The backup interface command is used in lieu of the static route. You don't use them together. The purpose of each is to inform the router of the interface to use to send and receive traffic in the event that the primary link fails.


I have used floating statics before and they worked like a charm. The only problem I have seen with them is when they are misconfigured.


Remember that simply adding a static route with an AD of, say, 200 is not enough. There must be a comparable route in the routing table for the router to consider the route with an AD of 200 as a floating static.


For example, if you configure a static route like...


ip route 10.10.80.0 255.255.254.0 bri0 200


...there better be a dynamic route to 10.10.80.0 255.255.254.0 in the routing table. When the comparable dynamic route is lost, it will result in the floating staic being placed in the routing table.


HTH


If so, please rate all helpful posts.


Victor

Edison Ortiz Tue, 03/04/2008 - 14:42
User Badges:
  • Super Bronze, 10000 points or more
  • Hall of Fame,

    Founding Member

The problem with using backup interface in a Frame-Relay environment is the interface may be up/up and the frame-relay cloud may be down.


You can get around this dilemma by configuring EEK between the 2 routers

http://www.cisco.com/en/US/docs/ios/12_0t/12_0t5/feature/guide/FRKeep.html


But I highly suggest you do some extensive testing before deploying into production.


The requirement for the backup interface to work is that the frame-relay interface is down/down.


HTH,


__


Edison.

dtecco Wed, 03/05/2008 - 06:45
User Badges:

Ok, I still need clarification on this.


How do I use my ISDN connection as a backup when I am only doing static routing?


It seems to me that backup interfaces, dialer watch, and floating static routes only work when you have some sort of dynamic routing occuring on your router.


What is the ISDN backup solution people use when they are ONLY doing static routing?

lamav Wed, 03/05/2008 - 08:54
User Badges:
  • Blue, 1500 points or more

D:


As I pointed out above (you may want to go back and read it closely), the static route that points to the ISDN interface informs the router to use the ISDN interface to route traffic for the network specificed in the static route.


You are sort of correct in saying that the static route will only work when dynamic routing is configured in that the static route will be a BACKUP for a COMPARABLE dynamic route, which will fail once the primary link is down.


So, if you're running EIGRP on a remote router, and in that router's routing table you have a route to the 10.10.0.0 255.255.0.0 network, which is your core network, you can configure a floating static route on that remote router to 10.10.0.0 255.255.0.0 with an AD of 200.


The static route will NOT be placed in the routing table because a COMPARABLE dynamic route, with a lower AD, already exists in the routing table. However, if the primary link fails, you will lose your EIGRP neighbor and the dynamic route will time out. It is only then that the COMPARABLE static route will be placed in the routing table. So, the router will switch ALL traffic destined for 10.10.0.0 to the ISDN interface, since that is the ONLY route it will have.


HTH


Please rate this post if it has helped you


Victor

Correct Answer
Richard Burts Wed, 03/05/2008 - 09:11
User Badges:
  • Super Silver, 17500 points or more
  • Hall of Fame,

    Founding Member

  • Cisco Designated VIP,

    2017 LAN, WAN

Dave


I agree with most of what Victor says here. I would like to clarify one point: you commented and Victor sort of agreed that it would work only if there were a dynamic routing protocol. This is not a correct assumption. It can also work quite well when there are only static routes. The key concept is the floating static route has an administrative distance higher than the default of 1. With its administrative distance configured the floating static route works fine with dynamic routing protocols and with static routes also. Whatever the primary route is (whether from dynamic or from static) its administrative distance is lower than the floating static and is placed into the routing table. When the primary route fails and is withdrawn from the routing table then the floating static is placed into the routing table. This would allow it to send traffic over your ISDN connection.


HTH


Rick

lamav Wed, 03/05/2008 - 09:35
User Badges:
  • Blue, 1500 points or more

Rick:


I agree with your description of AD and the selection of a route. In fact, I went into it extensively and in a very detailed manner above for Dave to understand it. I gave him examples, too.


As far Dave's thought that the static route will only work when dynamic routing is also configured, I understood what Dave was getting at, which is why I said he was "sort of" correct about that. I then went on to elaborate on how route selection occurs. The main point regarding route selection is the AD, not so much whether dynamic routing is configured or not, which is why I went into the AD concept of route selection in so much detail.


HTH


Victor



dtecco Thu, 03/06/2008 - 06:17
User Badges:

Victor,


I understand what you are saying. I perfectly understand the concept of dynamic routing and floating static routes. I clearly understand how to use AD. I do not think you fully understand what I am asking here.


Please throw dynamic routing out of this discussion. What I am asking has nothing to do with dynamic routing at all. I realize that you may have brought up dynamic routing to use as a detailed example, but I perfectly understand how dynamic routing works in an environment with a floating static route. I do not wish to discuss dynamic routing protocols in this discussion anymore.


The routers I have in this scenario only have static routes. There is no dynamic routing whatsoever.


Rick, you had the answer I was looking for...but I was not expecting to hear that. Are you sure you can have a static route for the primary link - and then have a floating static route for the secondary link on top of that? I did not think you can back up a static route with a floating static route. I tested it out in a lab environment, and I could not get it to work...


ip route x.x.x.x 255.255.255.0 serial0.1

ip route x.x.x.x 255.255.255.0 dialer1 200


I must have fat fingered the config or something, because I could not get the floating static route to kick in when I simulated the primary link going down. I am going to try it again today and see if it works.


Thanks,

lamav Thu, 03/06/2008 - 06:49
User Badges:
  • Blue, 1500 points or more

"Are you sure you can have a static route for the primary link - and then have a floating static route for the secondary link on top of that?"


Of course you can.


As we mentioned above, the primary route that will forward traffic out the primary link will have an AD LOWER than the floating static. That primary route may be dynamic or static. It doesn't matter.


The backup/floating static route will forward traffic out the backup interface and have an AD HIGHER than the primary route. The higher AD is why we call it floating.


Lastly, those routes must be COMPARABLE -- same prefix length. 10.0.0.0/8 and 10.0.0.0/8 are comparable. 10.0.0.0/8 and 10.10.10.0/24 are NOT comparable.


If your setup still doesn't work, come back and post the config. You may also want to make sure that you have configured your interesting traffic correctly. Interesting traffic is what will force the dialer to dial out the ISDN interface.


[edit] To REALLY drive this point home, let me reiterate that the reason I say that both routes should be comparable is so that the router doesn't place your backup static route in the routing table without the primary link going down, which will cause your ISDN link to kick in for nothing, thereby incurring huge costs from TELCO.


In other words, if you configure a static route with an AD of 200, but there is no other competing route in the routing table, it WILL be placed in the routing table. The router will not reject it. This is fundamental to the route selection process.


HTH


Victor

dtecco Thu, 03/06/2008 - 10:39
User Badges:

Thanks Victor,


I tested out this theory in my lab, and it did not work. The remote router (the one with the backup interface configured) actually added the floating static route in its routing table when I disconnected the primary circuit and the secondary kicked in. That part worked as expected.


However, the local router was still using the primary static route (which is the return route) in its routing table. It would not cut over to its floating static route. In other words, the return route does not work. The local router is still trying to send traffic out the primary interface.



Here is what I have on the local router...


ip route 192.168.203.0 255.255.255.0 s0.1

ip route 192.168.203.0 255.255.255.0 dialer3 200



If I manually remove the first static route entry from the routing table, then communication will work.


How do I get the local router (the router that is receiving the ISDN backup call) to route traffic through the dialer interface without manually removing the serial interface route out of the table? Do I have to use the backup interface command on BOTH ends of the connection or something? I would find that hard to believe.


I attached an illustration. The remote router is the router that has the backup interface configured, and is initiating the ISDN call. Like I said, the remote router works fine. It is the local router that will not change the entry in its routing table.



Mohamed Sobair Thu, 03/06/2008 - 11:14
User Badges:
  • Gold, 750 points or more

Hi,


The (Backup interface) command would influence the (LINE PROTOCOL DOWN) on the secondary interface.


So, if you are configuring it, it wont work with static route simply because their is no connection for the second link . Its like when you unplugg a cable from a device.


you will still need static route to route traffic to a specific destination. (IF static route is the only option). because still if the interface is UP/UP , the router wont know how to route traffic since there is no entry in the routing table for it.


HTH

Mohamed

lamav Thu, 03/06/2008 - 11:38
User Badges:
  • Blue, 1500 points or more

Dave:


I asked you to post the configs if it didnt work. You havent.


Anyway, how are you failing the primary link?


Does the s0.1 interface on the local router go "down"? If not, it will not remove the primary static route from the routing table because it will think that the WAN connection is still OK.


HTH


Victor

dtecco Fri, 03/07/2008 - 05:45
User Badges:

Ok, I got it to work now.


When I simulated the outage of the primary link, my local router's serial interface was still reporting up/up. That is why it kept the primary route in the routing table.


I didnt post the configs because it wasnt a configuration error - it was human error. My outage simulation never truly brought both sides of the primary link down (only the remote side went down) - duh me.


I appreciate everyone's help, thanks for steering me in the right direction.

lamav Fri, 03/07/2008 - 14:06
User Badges:
  • Blue, 1500 points or more

Dave:


Im glad you finally got it to work.


Checking the state of the interface is what I suggested you do in my last message...


"Anyway, how are you failing the primary link?


Does the s0.1 interface on the local router go "down"? If not, it will not remove the primary static route from the routing table because it will think that the WAN connection is still OK."


Victor

Actions

This Discussion