[c-nsp] Monitoring failed state of DSL / Cable (Ethernet)interfaces

[Ted Mittelstaedt]

> -----Original Message-----
> From: cisco-nsp-bounces at puck.nether.net
> [mailto:cisco-nsp-bounces at puck.nether.net]On Behalf Of Dan Martin
> Sent: Tuesday, March 01, 2005 12:33 PM
> To: Brian Feeny; cisco-nsp
> Subject: RE: [c-nsp] Monitoring failed state of DSL / Cable
> (Ethernet)interfaces
>
>
> You may try "rping" on the router to ping some router upstream of the
> dsl modem.  When the poll fails, you take the message and use that to
> admin down the Ethernet interface on the router, thus triggering some
> switch over.
>
> Once its switched over, you will need to bring the interface back up so
> you can keep pinging so you will know when the dsl line comes back to
> life.
>

That won't work because as soon as the line is switched over the pings
will start working again - since they now will be routed around through
the backup link.

ALL of the commercial solutions I've seen that use DSL for a primary
assume PPP-mode DSL.  If the DSL line goes down the PPP packets don't
come in anymore and it is easy for the router to understand that the
DSL line went away.

If it's bridged DSL then one way is to set your script up to ping the
default gateway on the DSL side then check for the existence of the
MAC address in the arp table for that IP number.  If it isn't there then
you know the DSL line is down and you can rewrite the default route.
And when the DSL line comes back you will know because the ARP will
reappear.
Unfortunately this is beyond the capabilities of Cisco routers.

Running OSPF out both interfaces would be a better way to do it.  That
would require the provider to support it, though.

Ted