Broken link detection

[aojschonek]

Hi,

Please see the PNG file I attached. What you see is a redundant connection between the A and the B site. There is no routing protocol running between the A and the B site. The links between the A and B site are running over a set of switches. The problem with this setup is that for example when the link between switch S2 and site B (R4) goes down the link between S2 and site A (R2) will still be up and the link will not be removed from the routing table of R2, the network will not converge.

TCL scripting is not possible on a 3550/3560 switch.

It is not possible to run a routing protocol between the A and B site, static routing is used.

UDLD make no sense in this situation.

Is there a way to solve this problem?

thx.

[Collin Clark]

Take a look at Embedded Event Manager. It can query the other side and if it fails it can remove the route from the table. I'm not an expert on this, but if you post in Network Management, the experts there can probably help.

Hope that helps.

[Giuseppe Larosa]

Hello Jan,

possible tools are:

static routes with object tracking: you can track the state of the next-hop if not available a floating static route (higher AD) pointing to the colocated router can be used

http://www.cisco.com/en/US/docs/ios/12_3/12_3x/12_3xe/feature/guide/dbackupx.html

the feature should be used on both the routers connected with the primary link

these routers should be the Active in a HSRP group so that traffic goes to them first in normal scenario.

normal HSRP interface tracking cannot solve the problem for the switch in the middle but you can use object tracking also to track next hop reachability in the same way (this is possible with HSRP not with VRRP and GLBP but I may be wrong)

see

http://www.cisco.com/en/US/docs/switches/blades/3040/software/release/12.2_44_se/configuration/guide/swhsrp.html#wp1083653

Probably this second solution is to be preferred if reachability of next-hop can be defined as an object to be tracked

Hope to help

Giuseppe