Backup interface with frame relay subinterfaces?

[dtecco]

Hello,

Can you utilize the backup interface feature for subinterfaces that are on the same physical interface?

Example:

serial 0.1 ---> PVC #1

serial 0.2 ---> PVC #2 (backup interface for serial0.1)

It is my understanding that the physical interface itself needs to go down in order to trigger the backup interface. How is that possible with frame relay subinterfaces?

[gpulos]

if this were one of two FR links to a specific site, you could utilize the 2nd FR link as the backup. (perhaps they can both be configured via routingProtocol or floatingStaticRoute for redundancy if not already done so)

one FR subinterface can go down, LMI loss, and the 2nd, if configured so, can act as the redundant link to that site.

see this link for some examples:

http://www.cisco.com/en/US/tech/tk713/tk237/tech_configuration_examples_list.html

[dtecco]

I apologize, perhaps I should elaborate further....

The two FR links are going to DIFFERENT sites.

PVC #1 is going back to the main hub in a "hub and spoke" environment.

PVC #2 is going back to a disaster recovery center that is at a completely different location.

Our provider has given us one circuit, with two DLCIs. Each DLCI is pointing to a different location. All routing is static.

This is what I am basically hoping to do (example)...

interface serial0.1 point-to-point

description FR Link A to main hub router

frame-relay interface-dlci A

backup delay 10 30

backup interface serial0.2

!

interface serial0.2 point-to-point

description FR Link B to remote backup router

frame-relay interface-dlci B

Will the backup link kick in if serial0.1 goes down? Keep in mind that both FR links are going to completely different routers. If that does work, then I take it I would have to implement a floating static route also?

[Richard Burts]

Dave

Knowing a bit more about your environment is helpful. I am not convinced that the backup interface will work for you based on what you have told us. If we knew a bit more about your requirements we might be able to find better solutions. Knowing that you have a primary site and a backup site is good, but we do not know whether the destination addressing is the same at the backup site as the primary site or is it different addressing?

I believe that some kind of routing solution is likely to be more effective than the backup interface solution. It might be floating static routes, it might be a dynamic routing protocol, or it might be Policy Based Routing. If we knew a bit more about your requirements we could determine what is a better solution.

HTH

Rick

[dtecco]

Appreciated Rick,

First off, the use of dynamic routing protocols is not an option. All devices are using static routes, and there is no way around that.

Secondly, the destination addressing at the backup site will be different from the primary site. In other words, each destination site is on its own network.

To elaborate further on the destination of the backup FR link...

The location is a disaster recovery center - which will only be used in case of, well, a disaster. That is the reason why the "backup interface" feature looked so appealing, because you can actually set the delay of the backup activation. Ideally we would like the backup delay to be set around 6 hours. This isnt a true fail over scenario. Like I said, this backup link is only in case of events such as natural disasters.

We are hoping to use the backup interface solution, along with static floating routes. I wanted to confirm this solution would work in our scenario before I move forward.

[ilya.varlashkin · Accepted solution]

If your PVC is configured on point-to-point sub-interface on both sides, then you can rely on PVC status provided by your service provider via LMI. You can check this by forcing sub-interface on one of the sides down and see 'sh frame-relay pvc ' on the other side. If they don't provide end-to-end status or if that is unreliable, you can configure your own end-to-end frame-relay keepalive like this:

!

map-class frame-relay MyFR

frame-relay end-to-end keepalive mode bidirectional

!

interface serial0.1 point-to-point

frame-relay interface-dlci A

frame-relay class MyFR

!

You must configure both sides for keepalive.

Now you can setup floating static routes like:

ip route 0.0.0.0 0.0.0.0

ip route 0.0.0.0 0.0.0.0 250

Similar applied on the hub site(s).

Although you could use 'backup' interface, there's very little if any value in it with 'always-on' circuits.

Beware that if your PVC is configured as point-to-point sub-intf only on the remote site, while hub is configured with PVC either on point-to-multipoint sub-interface or on the main interface, then you cannot use this method (even floating routes won't help) and have to configure dynamic routing protocol.

[dtecco]

Thank you ilya (and rburts). That is exactly what I was looking for.