08-16-2012 06:48 PM - edited 03-04-2019 05:17 PM
I'm attaching a representation, simplified, of my scenario.
I also attached the representative configuration of the two routers in question.
Site-A is our major site (the are actually almost two sites. Different Buildings, but interconnected with Dark-Fiber).
Site-X,Y,Z are the border sites of a big mesh with almost 30 sites and more the 40 Point-To-Point Links.
Everything works fine thank to EIGRP.
(Oh god, how this is better the any other alternate routing protocol)
On Site-A, by redundancy reasons, we have two 3560-E doing Core Routing.
(I was dreaming with VSS, but our money was a little short)
I Use HSRP to provide First Hope Redundancy,
and EIGRP to solve the routing to the radio mesh.
In HSRP I use a track, based on the status of a stub LoopBack, to decrement the priority of HSRP.
With this(shuting and no-shuting this loopback on each router) I can define which one is the Active box.
It works fine...
My problem becomes when I set the secondary box as Active.
(Task that I need to do eventually, to maintenance and things like that)
The problem is that I used the command "offset" on EIGRP to define which of those 3560 would be the prefered by the routers on sites X, Y, and Z.
And when the 3560-3 is the active, traffic goes from him(3560-3) to mesh, but remains coming from mesh to standby box(3560-2).
I was looking for some alternative using a route-map or something like that.
But I couldn't find anything.
My only alternative until now is to set a pair of EMM based on syslog message of track 1 state changing.
Track 1 goes up -> Change the "off-set" to 25
Track 1 goes down -> Change the "off-set" to 50
But I really don't like this kind of solution.
I prefer something already inside of the configuration.
Does anybody has a suggestion?
Edited by: Douglas Fernando Fischer -> First time -> Adjusting translations mistakes.
Edited by: Douglas Fernando Fischer -> Second time -> Attaching Config files with the implemented solution.
08-22-2012 06:15 PM
Well, after several tentatives... I surrendered myself to EMM.
It works well!
I attached the implemented solution to the original question with suffix "_final".
But my afraid is related to processing capacity on my real scenario
(remembering that this scenario is representative by safety reasons)
In my real scenario I have 10 Costumers, each one in a VRF with at least 5 Interface Vlans each, and several Eigrp Neighbors in total.
Discover and save your favorite ideas. Come back to expert answers, step-by-step guides, recent topics, and more.
New here? Get started with these tips. How to use Community New member guide