[c-nsp] L3 Policy-Map hits switched traffic

[Bruce Pinsky]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Bernhard Schmidt wrote:
> Hi,
> 
> we have an interesting issue here with four Sup720A
> 
> 
> User1 --  R1 --------- R2 --------- R3 -------- R4 -- User2
>        (18)SXD7    (18)SXD7     (33)SRA      (18)SXD7
>                        |
>                        |
> 		    outside
> 
> All links between those routers are 10GE (WS-X6704-10GE) 802.1q trunks,
> all routers do OSPF in one common backbone VLAN. When User1 tries to
> reach User2 we can see on a NAM installed in R1 that the packet is sent
> on the backbone VLAN with the correct destination MAC of R4. However,
> the traffic is processed by 
> 
> interface Vlan998
>  ip policy route-map BLA
> 
> on R2 which sets the next-hop for this packet to outside. 
> 
> Is this a (known) bug or a feature? Since R2 only switches this traffic
> it should not hit the route-map on the SVI, or am I wrong here?
> 

I'm not an expert on the 6500 switching path, but I suppose that adding PBR
 to that VLAN might end up causing the hardware to be punted to the MSFC
for forwarding to insure that the packet doesn't need to be forwarded
according to something other than destination address.  Tim Stevenson or
Ian Cox could probably comment a bit more on this as well as these references:


http://www.cisco.com/univercd/cc/td/doc/product/lan/cat6000/122sx/swcg/cef.htm
http://www.cisco.com/univercd/cc/td/doc/product/lan/cat6000/122sx/swcg/layer3.htm#wp1033565

- --
=========
bep

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.4 (MingW32)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org

iD8DBQFFAHxNE1XcgMgrtyYRAlYaAKCnFAPBp6r+X4YQ01KOFdfyHmn+7wCg+hPD
a7VFXNkQ0yFKTtdIV4uGCHc=
=ZbZ9
-----END PGP SIGNATURE-----