Adam,
the problem with dialer 0 is, that it will not go down, so the default route pointing to it will not disappear.
In the ISDN/BRI/Modem case, that is fine, since outgoing traffic gets routed to the interface which is "up (spoofing)" and
triggers (thru the dailer-group/list pair) the dail-out process.
For allways-online xdsl "dial-"connections with "dialer persistent" and "time-flat" charging, this brings those backup problems.
One solution is, iff the remote side never changes and has allways the same IP adress
(on my dail-infrastructure this is true, even over several dial-in devices.
On the backup interface, i have "no peer neighbor route".
You may set the "dsl" default route to the IP of the ISP's router (that is the peer-ip-address for which the router installs noramlly a host route to the dialer interface). During successful login, you see this ip address using "show users" .
For the backup dialer, the default-route may then go to the dialer interface with higher "cost" .
An other solution would be a track monitoring the state of the atm0 interface
an discarding the default route to your dialer0
track 11 interface ATM0 line-protocol
ip route 0.0.0.0 0.0.0.0 dailer0 track 11
The dialer watch is not nessassary if you want to dail the backup-device on demand.
The dialer watch watches the route and calls the backup line when the route disappers.
Discussion of the differnet backup-strategies in
http://www.cisco.com/en/US/tech/tk801/tk133/technologies_tech_note09186a008009457d.shtml
if the route you watch is the host-ip of the remote-side,
you should use this as the next hop fore the normal/dsl default-route and remove the dialer watch commands.
Hope this help's,
Juergen.
here, you have a floating staic route as the default route which gets recursively resolved when the peer neighbor (host)-route
gets installed during PPP negotiation of the main interface.
If dsl goes down, or the ppp data flow stoppes (even when dsl is in sync!) the interface will go down ofter
5 ppp lcp keepalive packets got lost. Default (cisco) is every 10 seconds to send a keepalive,
most xDSL Carrier equipment and home-routers use every 30 seconds.
Tune with
keepalive 1
to every one second, may not work depending on Carrier and ISPside.
After that, the peer neighbor route diappers.
The default route will be dropped when the process checking them runs next,
you may tune that with
ip route static adjust-time 10
to every 10 seconds (works on 87x without too much cpu load).
gate#sh users
...
Interface User Mode Idle Peer Address
Vi1 PPPoE 00:00:00 aaa.bbb.ccc.ddd
gate# sh run
! ...
interface Dialer0
description DSL Dialer
dialer pool 1
dialer idle-timeout 0
dialer persistent
dialer-group 1
no cdp enable
!...
!
interface Dialer1
description ISDN/PSTN Dialer Backup
dialer pool 2
dialer-group 2
no peer neighbor-route
no keepalive
no cdp enable
! ...
!
ip route 0.0.0.0 0.0.0.0 aaa.bbb.ccc.ddd
ip route 0.0.0.0 0.0.0.0 dialer1 100
! ...
