Dial Backup on EIGRP Adjacency Failure?

ayurkowski · ‎02-10-2005

Hi, Makarand.

I have a 1720 and a 3660 (with a multiport AM16 modem card option) both connected to a LAN with EIGRP configured. I would like to have an async /aux backup on the 1720 call the 3660 IF EIGRP adjacency goes down and there is a traffic requirement; however, it only dials if E0 is *physically* disconnected (or administratively shut down). Once it DOES dial, it DOES route traffic, so obviously the dialer configuration is correct.

I have a static (backup) route with an admin cost of 200 to the remote networks on the 3600, and assumed this would be used if traffic couldn't get thru the LAN; this doesn't seem to be the case. Suggestions?

Regards,

- Alan

sstudsdahl · ‎02-10-2005

Alan,

Can you provide your dial backup configuration? What does the route that needs to fall out of your routing table in order for the floating static to be installed in the routing table look like? If the E0 interface remains up, and the route you have the floating static configured for remains in the routing table, then the dial process will not work.

You might consider using the "dialer watch-list" command as an alternative. This will allow the dialer to watch a specific route in the routing table, and if it disappears, will cause the dial process to begin.

Take a look at the link below. It gives advantages/disadvantages of both methods.

http://www.cisco.com/warp/public/123/backup-main.html

Steve

ayurkowski · ‎02-10-2005

I could have probably explained this a bit better (but I hate doing these piddly line drawings!)

Extracts of config follow:

(A) FA0 E0 192.168.22.x FA0/0 FA0/1 (B)

| | |

+--- REMOTER ----- SWITCH ------- MAINR -----+

| \ / |

Async5 Async34

(A) 192.168.25.0/24

REMOTER

chat-script dialbackup "" "AT" OK "ATDT\T" TIMEOUT 45 CONNECT \c

modemcap entry USR:MSC=&F1S0=1

!

interface Ethernet0

ip address 192.168.22.102 255.255.255.0

!

interface FastEthernet0

ip address 192.168.25.252 255.255.255.0

!

interface Async5

ip address 192.168.45.2 255.255.255.0

encapsulation ppp

carrier-delay 60

dialer in-band

dialer idle-timeout 60

dialer map ip 192.168.45.1 name MAINR broadcast xxx

dialer map ip 192.168.22.0 name MAINR broadcast xxx

dialer watch-group 1

dialer-group 1

async default routing

async mode dedicated

no peer default ip address

ppp authentication chap

!

router eigrp 1

passive-interface FastEthernet0

network 192.168.22.0

network 56.0.0.0

auto-summary

eigrp log-neighbor-changes

!

ip route 56.0.0.0 255.0.0.0 192.168.45.1 200

ip route 192.168.22.0 255.255.255.0 192.168.45.1 200

!

access-list 101 permit ip any any

access-list 101 deny ospf any any

dialer watch-list 1 ip 192.168.22.0 255.255.255.0

dialer-list 1 protocol ip list 101

!

line aux 0

exec-timeout 0 0

script dialer dialbackup

modem InOut

modem autoconfigure type USR

transport input all

autoselect ppp

speed 115200

flowcontrol hardware

MAINR

(since dialup works, I'll only show what's needed for addressing)

interface FastEthernet0/0

ip address 192.168.22.252 255.255.255.0

!

interface FastEthernet0/1

ip address 56.100.100.252 255.0.0.0

!

interface Async34

ip address 192.168.45.1 255.255.255.0

!

router eigrp 1

passive-interface FastEthernet0/1

network 192.168.22.0

network 192.168.25.0

auto-summary

eigrp log-neighbor-changes

!

ip route 192.168.25.0 255.255.255.0 192.168.45.2 200

(B) 56.0.0.0/8

The issue here is the access to the .22.0 nodes sitting on the switch.

Is that adequate FYI?

sstudsdahl · ‎02-10-2005

Based on the configurations you provided, it looks like your network statements for the two routers are actually from the opposite router. If this is the case, if you issue the command "show ip route", you likely only have the directly connected routes listed and you should not be able to communicate to opposite ends of your routers. I'm guessing this was a typo (I've done that myself) and base the info below on that assumption.

On the REMOTER router, you have your dialer watch-list setup to monitor the 192.168.22.0/24 network between the two routers. If you change this to "dialer watch-list 1 ip 56.0.0.0 255.0.0.0" the dial process should begin without having to do anything with the ethernet interface.

What happens here is that when the two routers lose communication with each other, the EIGRP learned route within REMOTER is removed from the routing table. As soon as the 56.0.0.0/8 network disappears from the routing table, the dial process will begin. This would be the same thing that happens when you administratively or physical shutdown the ethernet interface of REMOTER. The 192.168.22.0/24 network disappears from the routing table which causes the REMOTER router to dial since it has been told to watch the 192.168.22.0/24 network.

Steve

ayurkowski · ‎04-06-2005

Finally got back to this problem.

The previous comment regarding the reversed network statements was correct (e.g., a typo), and the backup configuration does, in fact work. However, I have one big concern: the dial backup comes up if the 56.x network is taken down for maintenance, even IF the link between MAINR and REMOTER is still up.

Is there possibly a way to advertise a loopback address on MAINR via EIGRP? I'm experimenting with this, but no luck thus far. I fundamentally need to have the backup link dial ONLY if it is unable to contact the main site.

- Alan