From: topgun topgun (topgunrs1@yahoo.com.sg)
Date: Fri Jan 13 2006 - 22:02:03 GMT-3
Hi Scott,
I am not sure if the SP will preserve the parameter, but I doubt so.
But I am quite sure they will preserve the AS path which they propagate.
Both R1 and R2 are different AS.
So R1 can filter R2 AS bgp route coming in via MPLS, and so does R2.
Will lab the active standby design to ensure ok.
thanks
Scott Morris <swm@emanon.com> wrote:
Great question, and only he can answer that part! :)
Even if it is SP controlled though, you could always do an as-path prepend
to give yourself something to filter with on the other side. The community
part would be up to the SP passing that along! I just like to pretend they
cooperate and play nicely with others! (grin)
But what I'm thinking of is that the CE talks to the PE via BGP. The PE's
pass that info along. If the provider is doing things like an AS-override,
then all bets are off! :) Otherwise, the only other thing that pops to
mind is setting the MED. But I highly doubt the SP would care to preseve
that.
I was just trying to think of something that could be set on one side, seen
on the other side and used in a route-map to filter out as needed. Although
the more I look at it, I'm thinking that setting bgp weights is probably a
safer idea (AD weight, not the "weight" neighbor part) so that EIGRP routes
would be preferred over eBGP-learned routes.
"distance bgp 200 200 200" was my thought on this. That would handle making
the internal EIGRP routes AND external ones (90 and 170) preferred so that
anything available via the L3 switch would work better. That would be the
"yes or no" version of failover. He may want to get more particular with
some things prefer one way over another which gets back to the filtering or
weighting on BGP side.
Scott
_____
From: Peter McCreesh [mailto:petermccreesh@gmail.com]
Sent: Friday, January 13, 2006 1:44 PM
To: swm@emanon.com
Cc: topgun topgun; ccielab@groupstudy.com
Subject: Re: Redistribution BGP-EIGRP
Hi Scott,
Does this solution assume that you have control of the BGP router(s) in the
MPLS cloud.
I was working on this as if it was the SP who owned the MPLS routers and as
a result, weren't allowed to be changed.
I may be misunderstanding however. (I know you have no more info than I do
but what i'm getting at is the solution(s) you mentioned. They would be
implemented on the MPLS BGP router(s)??)
Thanks a lot,
Pete
On 1/13/06, Scott Morris wrote:
You could also consider things inside BGP like adding a bogus private AS to
the path, or setting a community and then filtering that on the other
side....
Scott
-----Original Message-----
From: nobody@groupstudy.com [mailto:nobody@groupstudy.com] On Behalf Of
topgun topgun
Sent: Friday, January 13, 2006 8:20 AM
To: topgun topgun; Peter McCreesh
Cc: ccielab@groupstudy.com
Subject: Re: Redistribution BGP-EIGRP
Hi all,
Just having this thought that BGP distance might not be a good solution,
as I have about 300-400 EIGRP routes inside the L3.
And also if new routes are added or coming in from the L3, does it mean I
have to add the new subnet into the access-list.
Is there a better way to resolve this issue?
Thanks in advnace.
Top
topgun topgun wrote:
Hi Peter,
The problem is when the R1 link to L3 switch is down, the R1 will still need
to reach 9.9.9.9 and 99.99.99.99. I can't filter the routes coming in from
MPLS (which is sent by R2). At this time, there is 9.9.9.9 and 99.99.99.99
BGP routes
When I bring up the R1 link to L3, the eigrp is formed. But my R1 still
prefer BGP routes. This is not desirable.
I tried using distance command for EIGRP routes in BGP protocol, and seem
ok. Thanks to Gustavo's suggestion... :)
e.g
router bgp 65007
distance 171 remote MPLS interface 0.0.0.0 9
access-list 9 permit 9.9.9.9
access-list 9 permit 99.99.99.99
Any comment in using this distance command in BGP?
Thanks
Top
Peter McCreesh
wrote:
Hi Top,
would it be an option to filter your own nets ( 9.9.9.9 and 99.99.99.99) out
of the BGP updates coming in from the MPLS cloud. You could then do a soft
reset on your BGP neighbor so the peering will stay up but your networks
should then "disappear" from the BGP updates coming in to you. Probably best
to do this on both R1 and R2
I hope i'm understanding your problem OK
..Pete
On 1/13/06, topgun topgun wrote:
>
> Hi all,
>
> I am having some design issue and would like to seek some advise.
>
> I have 2 routers, and their WAN link are connected to MPLS running BGP.
> Their LAN are connected to different L3 switch running EIGRP. The 2 L3
> switch are connected and run EIGRP too.
>
> MPLS cloud -----------R1 -----------L3 switch I I
> I-------------------------R2 ----------- L3 switch
>
>
> R1 and R2 are two redistribution point; redistribute EIGRP to BGP and BGP
> to EIGRP.
>
> My EIGRP cloud (L3 switches) has internal and external network, say
> 9.9.9.9 (internal) and 99.99.99.99 (external)
>
> R1 will prefer both 9.9.9.9 and 99.99.99.99 via L3 switch, and as such, R2
> will be via MPLS cloud (AD BGP 20 as compared to EIGRP 90 and 170).
> I want R2 to use L3 as the gateway like R1.
>
> Distance cannot be used as it can't work for external eigrp routes.
> bgp backdoor cannot work as I have to clear the ip bgp first.
>
> Appreciate your help.
>
> Thanks
>
> Top
>
>
> ---------------------------------
> Do you Yahoo!?
> New and Improved Yahoo! Mail - 1GB free storage!
>
> _______________________________________________________________________
> Subscription information may be found at:
> http://www.groupstudy.com/list/CCIELab.html
This archive was generated by hypermail 2.1.4 : Wed Feb 01 2006 - 07:45:49 GMT-3