1. Home
  2. Cisco Systems
  3. Monitoring line status where interfaces won't go down

Monitoring line status where interfaces won't go down

Hi all

consider the following situation:

Customer edge Router (e.g. Cisco 28xx) with a primary link to the provider edge router via an Etherconnect line, i.e. a connection where the carrier provides a modem which is conected to the router. A backup link is established over a steady DSL connection. On both interfaces we have /30 transfer networks an run ospf with the provider edge routers on the other ends.

Now we want to monitor the primary connection from the provider backbone. Usually we do not use ospf and simply ping the customer edge IP adresses of the two transfer networks in order to test their reachability. Now in case of the running ospf this is not an option since both IP addresses would be always reachable as long as one connection is up.

An option would be to monitor the interface status via snmp. That's good as long as a failing connection implies an interface status change to down. Unfortunately this is not the case since the connection to the carrier's modem would always be up.

Well, I've read a lot about tracking objects and routing policies but I'm not sure whether or not (and if yes, how exactly) this would be of any help in this case.

Any suggestions?

Thanks a lot in advance,

Grischa

Reply to
Gr!scha S+egemann
Loading thread data ...

BFD?

Reply to
Paul Matthews

i think you are ruling out the best way to do this - routing protocols are there to check reachability (as well as other things) - so why not use it?

Nope - far end address across the link is not "up" from your perspective unless the link works and you have 2 way packet exchange.

Near end may stay up though - in which case you are pinging the wrong object.

BFD may work as another suggested - but you need modern devices / code at both ends since this is fairly recent.

cisco have a way of linking "reachability" into statics using SAA / RTR or whatever it is called now

formatting link

Reply to
Stephen

I think there's a misunderstanding.

Imagine following. R1 and R2 are on one LAN, R3 and R4 at remote LAN. Running OSPF R3 will be reachable and all of it's interfaces as long as at least one LAN link is working.

How to discover failing LAN link A when all interfaces are still up using SNMP? R1's and R3's interfaces will stay up because of the modems.

R1 --- Modem --------- LAN link A --------- Modem --- R3 R2 --- Modem --------- LAN link B --------- Modem --- R4

Reply to
Andre Wisniewski

I think you should be able to use IP SLA to monitor status of the links. You will have to put some simple ACL in place to prevent ICMP traffic between R1/R3 pair to travel over Link B and vice versa. That way when link will go down so does IP SLA monitor.

Andrey.

Reply to
Andrey Tarasov

ACK ;)

That was (in principle) my question.

Yes, I thought so. But so far I have not seen how to configure this exactly.

Brilliant! Thanks a lot. I must have been blind. Of course the easiest and most simple solution is to deny icmp echo traffic from one WAN interface to the other on the router in question.

Thank you :)

Grischa

Reply to
Gr!scha S+egemann

You cheat.

We pull the OSPF neighbour table back and use that to decide......

The other thing you could watch is the traffic on the inbound Ethernet port.

If it works the same as one i have been using, then the only background keepalive traffic comes from the 2nd router, and with OSPF using hello packets as keepalive, you have a fixed amount of incoming multicast.

Pick up the inbound packet count using SNMP that vanishes (ie no delta between samples at least a couple of sec more than the OSPF hello time, you have a layer 2 fault between the 2 routers.

Reply to
Stephen

Cabling-Design.com Forums website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.