I found this sample config on CCO and have trouble understanding why interesting traffic is defined, since the backup interface command is used.
In my understanding, when the main link drops, the backup dialer interface (which has been kept in standby) goes up (the dialling occurs), then down - when the main link is back up, with consideration to
the timers configured in the backup delay command.
As I see for now - backup interface and defining interesting traffic + proper routing info are 2 separate ways of causing the backup interface to dial, and this configuration would work the same without the dialer-list + dialer-group parts
Sample config:
<deleted>
isdn switch-type basic-ni
!
interface Serial0
backup delay 10 30
backup interface Dialer1
ip address 192.168.10.2 255.255.255.252
encapsulation ppp
!
interface BRI0
no ip address
encapsulation ppp
dialer pool-member 10
isdn switch-type basic-ni
ppp multilink
!
interface Dialer1
ip address 172.20.10.2 255.255.255.0
encapsulation ppp
dialer pool 10
dialer idle-timeout 900
dialer string 5552222
dialer load-threshold 80 outbound
dialer-group 1
ppp multilink
!
dialer-list 1 protocol ip permit
<deleted>
Link to the full document:
http://cisco.com/en/US/tech/tk801/tk379/technologies_configuration_example09186a008010456b.shtml
-------
It's me again and I have found the explanation:
it seems that for the 3 dial-backup methods - backup interface, floating static routes and dialer-watch, in order for the call to be placed, the first 2 need the interesting traffic to be defined and the 3rd does not.