«

We currently have 1751-Vs in the field and we have WIC-2A/S modules in
them and one of the asynch ports on the card go to an external modem.
The routers WAN connection is 64k Frame and we use EIGRP on the WAN to
maintain routes. We have the router set up to do DDR using the modem
when a route to a key destination drops out of the routing table using
a static route (if the Frame goes down, EIGRP loses this neighbor, it
drops out of the table and uses the static route to initiate dial
backup).

Well...we have been informed by our telco that for some reason some of
our locations Frame circuit will not support the EIGRP and have
suggested we static route all of the locations traffic out. We have
the statics set up and think we have all of out bases covered, but now
our dial backup doesn't work because the static route does not drop
out when the Frame goes down.

Is there a DDR solution for the static routing scenario? I have seen
references to the "backup interface" command, but not much clear info.
from Cisco regarding how to use it.

Any ideas?

Thanks
-Robert

On 2004-11-08, Robert B. Phillips II <> wrote:
> We currently have 1751-Vs in the field and we have WIC-2A/S modules in
> them and one of the asynch ports on the card go to an external modem.
> The routers WAN connection is 64k Frame and we use EIGRP on the WAN to
> maintain routes. We have the router set up to do DDR using the modem
> when a route to a key destination drops out of the routing table using
> a static route (if the Frame goes down, EIGRP loses this neighbor, it
> drops out of the table and uses the static route to initiate dial
> backup).
>
> Well...we have been informed by our telco that for some reason some of
> our locations Frame circuit will not support the EIGRP and have
> suggested we static route all of the locations traffic out. We have
> the statics set up and think we have all of out bases covered, but now
> our dial backup doesn't work because the static route does not drop
> out when the Frame goes down.
>
> Is there a DDR solution for the static routing scenario? I have seen
> references to the "backup interface" command, but not much clear info.
> from Cisco regarding how to use it.
>


Well the problem with FR is that that in most cases the interface will
not go down when the FR dies so that kind of backup won't work. You can
try setting up a GRE tunnel between your locations and use keepalives.
If FR goes down this should take down the tunnel also (this should be
true for any recent IOS version). Then you can use floats and let the
DDR kick in.

--
Andrej Brkic
Fakultet Prometnih Znanosti, Zagreb, Croatia
E-mail:

In article <> ,
says...
> Well...we have been informed by our telco that for some reason some of
> our locations Frame circuit will not support the EIGRP and have
> suggested we static route all of the locations traffic out. We have
> the statics set up and think we have all of out bases covered, but now
> our dial backup doesn't work because the static route does not drop
> out when the Frame goes down.
>
> Is there a DDR solution for the static routing scenario? I have seen
> references to the "backup interface" command, but not much clear info.
> from Cisco regarding how to use it.
>
> Any ideas?
>

It is possible (with newer IOS'es) to use SAA to probe if you can reach
wanted endpoint and then to use that info for policy based routing. Of
course, you can always run BGP between endpoints to get the same
functionality.

HTH,

--
-Ivan.

*** Use Rot13 to see my eMail address ***

In article <>,
Andrej Brkic <> wrote:
>On 2004-11-08, Robert B. Phillips II <> wrote:
>> We currently have 1751-Vs in the field and we have WIC-2A/S modules in
>> them and one of the asynch ports on the card go to an external modem.
>> The routers WAN connection is 64k Frame and we use EIGRP on the WAN to
>> maintain routes. We have the router set up to do DDR using the modem
>> when a route to a key destination drops out of the routing table using
>> a static route (if the Frame goes down, EIGRP loses this neighbor, it
>> drops out of the table and uses the static route to initiate dial
>> backup).
>>
>> Well...we have been informed by our telco that for some reason some of
>> our locations Frame circuit will not support the EIGRP and have
>> suggested we static route all of the locations traffic out. We have
>> the statics set up and think we have all of out bases covered, but now
>> our dial backup doesn't work because the static route does not drop
>> out when the Frame goes down.
>>
>> Is there a DDR solution for the static routing scenario? I have seen
>> references to the "backup interface" command, but not much clear info.
>> from Cisco regarding how to use it.
>>
>
>
>Well the problem with FR is that that in most cases the interface will
>not go down when the FR dies so that kind of backup won't work. You can
>try setting up a GRE tunnel between your locations and use keepalives.
>If FR goes down this should take down the tunnel also (this should be
>true for any recent IOS version). Then you can use floats and let the
>DDR kick in.
>
>--
>Andrej Brkic
>Fakultet Prometnih Znanosti, Zagreb, Croatia
>E-mail:

It sounds like you are using your frame relay in multipoint mode, which
is a pain with EIGRP. The easy solution is to change your frame relay
configuration on every router with more than one PVC to use a
point-to-point subinterface for each PVC, hiding the point-to-multipoint
nature of frame relay (which your telco says they don't support) from
the router and treating each PVC as an independent point-to-point link.

If you are already running point-to-point subinterfaces on all your
PVCs, I would have serious doubts about the competence of the telco and
press them for the technical details of the reason their network can
even distinguish EIGRP packets from other traffic. If they
actually have a viable reason, I'd love to hear it! But if they do, you
could switch from EIGRP to BGP for your routing (not supported on all
Cisco hardware, IOS release, feature set combinations), which uses point
to point TCP connections for communications between peers. Another
option is to use Cisco proprietary "frame relay end-to-end keepalive" to
detect when a PVC is no longer usable for traffic.

As Andrej points out, with frame relay, you MUST use some mechanism
other than LMI updates to detect when a frame relay PVC goes down
without signaling link down on the local interface.

Good luck and have fun!

--
Vincent C Jones, Consultant Expert advice and a helping hand
Networking Unlimited, Inc. for those who want to manage and
Tenafly, NJ Phone: 201 568-7810 control their networking destiny
http://www.networkingunlimited.com

Vincent,
We are indeed implementing point-to-point subinterfaces for the
different PVCs on the Frame. At a few sites we were having issues with
connectivity on one PVC being really poor (i.e. low ping response and
slow ping response, certain apps that use the WAN timing out), so we
called the telco and they "rebuilt the PVC", no joy. They told us
after the rebuild that the PVC was clean, but we still were
experiencing our issues. Tthe telco finally told us it was because of
EIGRP. When we changed the router config from EIGRP to static routes
the problems went away. Which leads me to believe it is a cirucit
problem, since was have identical configs in different locations with
supposedly the same circuit provisioning and EIGRP does not hold
anything up.

I am pushing back on the telco to elaborate and help us, but that
is so far an uphill battle. Anyway I have received some good feedback
from the group here ... looks like it is time to hit the lab and start
tinkering.

Thanks
-Robert

On Tue, 09 Nov 2004 00:49:19 GMT,
(Vincent C Jones) wrote:

>In article <>,
>Andrej Brkic <> wrote:
>>On 2004-11-08, Robert B. Phillips II <> wrote:
>>> We currently have 1751-Vs in the field and we have WIC-2A/S modules in
>>> them and one of the asynch ports on the card go to an external modem.
>>> The routers WAN connection is 64k Frame and we use EIGRP on the WAN to
>>> maintain routes. We have the router set up to do DDR using the modem
>>> when a route to a key destination drops out of the routing table using
>>> a static route (if the Frame goes down, EIGRP loses this neighbor, it
>>> drops out of the table and uses the static route to initiate dial
>>> backup).
>>>
>>> Well...we have been informed by our telco that for some reason some of
>>> our locations Frame circuit will not support the EIGRP and have
>>> suggested we static route all of the locations traffic out. We have
>>> the statics set up and think we have all of out bases covered, but now
>>> our dial backup doesn't work because the static route does not drop
>>> out when the Frame goes down.
>>>
>>> Is there a DDR solution for the static routing scenario? I have seen
>>> references to the "backup interface" command, but not much clear info.
>>> from Cisco regarding how to use it.
>>>
>>
>>
>>Well the problem with FR is that that in most cases the interface will
>>not go down when the FR dies so that kind of backup won't work. You can
>>try setting up a GRE tunnel between your locations and use keepalives.
>>If FR goes down this should take down the tunnel also (this should be
>>true for any recent IOS version). Then you can use floats and let the
>>DDR kick in.
>>
>>--
>>Andrej Brkic
>>Fakultet Prometnih Znanosti, Zagreb, Croatia
>>E-mail:
>
>It sounds like you are using your frame relay in multipoint mode, which
>is a pain with EIGRP. The easy solution is to change your frame relay
>configuration on every router with more than one PVC to use a
>point-to-point subinterface for each PVC, hiding the point-to-multipoint
>nature of frame relay (which your telco says they don't support) from
>the router and treating each PVC as an independent point-to-point link.
>
>If you are already running point-to-point subinterfaces on all your
>PVCs, I would have serious doubts about the competence of the telco and
>press them for the technical details of the reason their network can
>even distinguish EIGRP packets from other traffic. If they
>actually have a viable reason, I'd love to hear it! But if they do, you
>could switch from EIGRP to BGP for your routing (not supported on all
>Cisco hardware, IOS release, feature set combinations), which uses point
>to point TCP connections for communications between peers. Another
>option is to use Cisco proprietary "frame relay end-to-end keepalive" to
>detect when a PVC is no longer usable for traffic.
>
>As Andrej points out, with frame relay, you MUST use some mechanism
>other than LMI updates to detect when a frame relay PVC goes down
>without signaling link down on the local interface.
>
>Good luck and have fun!

Thanks Ivan, I will look into this.

On Mon, 8 Nov 2004 22:04:54 +0100, Ivan Ostreš
<> wrote:

>In article <> ,
> says...
>> Well...we have been informed by our telco that for some reason some of
>> our locations Frame circuit will not support the EIGRP and have
>> suggested we static route all of the locations traffic out. We have
>> the statics set up and think we have all of out bases covered, but now
>> our dial backup doesn't work because the static route does not drop
>> out when the Frame goes down.
>>
>> Is there a DDR solution for the static routing scenario? I have seen
>> references to the "backup interface" command, but not much clear info.
>> from Cisco regarding how to use it.
>>
>> Any ideas?
>>
>
>It is possible (with newer IOS'es) to use SAA to probe if you can reach
>wanted endpoint and then to use that info for policy based routing. Of
>course, you can always run BGP between endpoints to get the same
>functionality.
>
>HTH,

Andrej,
Thanks for the feedback, I will investigate this further.

-Robert

On Mon, 8 Nov 2004 20:28:59 +0000 (UTC), Andrej Brkic <>
wrote:

>On 2004-11-08, Robert B. Phillips II <> wrote:
>> We currently have 1751-Vs in the field and we have WIC-2A/S modules in
>> them and one of the asynch ports on the card go to an external modem.
>> The routers WAN connection is 64k Frame and we use EIGRP on the WAN to
>> maintain routes. We have the router set up to do DDR using the modem
>> when a route to a key destination drops out of the routing table using
>> a static route (if the Frame goes down, EIGRP loses this neighbor, it
>> drops out of the table and uses the static route to initiate dial
>> backup).
>>
>> Well...we have been informed by our telco that for some reason some of
>> our locations Frame circuit will not support the EIGRP and have
>> suggested we static route all of the locations traffic out. We have
>> the statics set up and think we have all of out bases covered, but now
>> our dial backup doesn't work because the static route does not drop
>> out when the Frame goes down.
>>
>> Is there a DDR solution for the static routing scenario? I have seen
>> references to the "backup interface" command, but not much clear info.
>> from Cisco regarding how to use it.
>>
>
>
>Well the problem with FR is that that in most cases the interface will
>not go down when the FR dies so that kind of backup won't work. You can
>try setting up a GRE tunnel between your locations and use keepalives.
>If FR goes down this should take down the tunnel also (this should be
>true for any recent IOS version). Then you can use floats and let the
>DDR kick in.