I am seeing the same behavior as yourself, on code as late as 5.1R4.3.
I mistakenly replied off-list to the contrary. Thanks for the clarification.
I searched the JTAC problem report, and came up empty.
Ben Bird
> -----Original Message-----
> From: Richard A Steenbergen [mailto:ras@e-gerbil.net]
> Sent: Friday, July 05, 2002 2:44 PM
> To: Kent Yu
> Cc: juniper-nsp@puck.nether.net
> Subject: Re: protocol bgp group/neighbor local-preference
>
>
> On Fri, Jul 05, 2002 at 10:19:59AM -0400, Kent Yu wrote:
> > It works fine on our routers.
>
> Ok let's pick a random route, I've got a policy-statement
> setting local
> pref to 75, and local-preference statement on the eBGP neighbor it is
> learned from at 50, and a local-preference on the group it is
> in at 90.
>
> With the policy-statement:
>
> 12.164.224.0/21 *[BGP/170] 21:53:47, MED 0, localpref 75
>
> Without the policy-statement:
>
> 12.164.224.0/21 *[BGP/170] 21:56:38, MED 0, localpref 100
>
> Thus indicating the local-preference statement has no impact
> on learned
> routes from eBGP neighbors. As someone mentioned offlist, you
> can use a
> single prefix-list and match on the neighbor the route was
> received from,
> but this is just plain nasty. I'm at a loss as to why someone
> would make a
> neighbor statement for setting localpref on routes going out
> to iBGP peers
> (which is almost never done), without providing a statement to set the
> localpref when learned from eBGP peers (which is done quite
> often based
> on the status of the peer (transit, backup transit, peer,
> depref'd peer,
> etc), usually at the group level, but sometimes at the
> neighbor level).
>
> This is 5.3R1.2 and all I have available to test with, so if there is
> something I'm missing bug-wise please feel free to fill me in. :)
>
> --
> Richard A Steenbergen <ras@e-gerbil.net>
> http://www.e-gerbil.net/ras
> PGP Key ID: 0x138EA177 (67 29 D7
> BC E8 18 3E DA B2 46 B3 D8 14 36 FE B6)
>
This archive was generated by hypermail 2b29 : Mon Aug 05 2002 - 10:42:36 EDT