[c-nsp] Route-Map "AND" clause
Oliver Boehmer (oboehmer)
oboehmer at cisco.com
Wed Jun 14 07:45:01 EDT 2006
hjan at libero.it <mailto:hjan at libero.it> wrote on Wednesday, June 14, 2006
12:34 PM:
>> unfortunately not, but you could obviously create a single ACL
>> covering both ACE's.
>
> I have done a test with an access-list 10 that permit 0.0.0.0 and
> 10.10.10.0 but the behaviour is the same...
right, and thinking about this more, it doesn't make too much sense to
return TRUE if a prefix matches two different ACLs, would it?
>
>
>> not sure I understand. what do you mean by "only if eigrp process 1
>> is down"? can you elaborate?
>
>
> Ok, i try to explain the problem.
> Some "guru" in my company sell to a customer a backup line in another
> city but the customer use eigrp with our POP.
> Imagine this topology
>
> CITY-1-POP<----------->CITY-2-POP
>> |
>> |
> CPE-1 CPE-2
>
> The customer network is a L2 network, CPE-1 is the primary link,
> CPE-2 will be backup link. CPE-X do eigrp_1 with CITY-X-POP. CPE-1
> and CPE-2 do hsrp with CPE-1 primary.
> CITY-1-POP is a man with its adress as CITY-2-POP.
>
> Now the problem...
> How can i control the input point ? Obviosly other city 2 client
> enter from city-2-pop over the backup link.
> The only thing i have thougth is to add an eigrp process within the
> two cpe and redistribute the client lan from CPE-2 only if this new
> eigrp goes down, i.e. CPE-1 fail.
> I guess things are pretty hard and the only way i see to solve the
> problem is use route-map.
What's your routing protocol in your core? You could address this
problem by increasing the administrative distance of the eigrp 1 process
on CITY-2-POP so it is higher than your core IGP process. This way,
CITY-2-POP will always prefer the prefixes it learns from the core,
until CPE-1 dies and CITY-1-POP no longer advertises the customer
prefixes.
Would this help?
oli
More information about the cisco-nsp
mailing list