[c-nsp] multipath doesn't work on received default routes?

[Bruce Pinsky]

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

Dan Hollis wrote:
> I'm receiving default routes from 3 upstreams, but ios doesn't seem to 
> want to multipath the default route. Only one default route is ever used, 
> verified by looking at the outbound counters on the interfaces.
> 
> i do have 'maximum-paths 6' in the 'router bgp 13941' section of my config.
> i also have cef enabled on all interfaces. i have confirmed it is enabled 
> on the interfaces, and i have also manually cleared route caches (clear ip 
> ro *) etc.
> 
> sh ip bgp 0.0.0.0 verifies that the routes are valid, but sh ip ro 0.0.0.0 
> shows only one route ever being entered into the table.
> 
> #sh ip bgp
> BGP table version is 1730946, local router ID is 2xx.xxx.xxx.xx6
> Status codes: s suppressed, d damped, h history, * valid, > best, i - internal
> Origin codes: i - IGP, e - EGP, ? - incomplete
> 
>     Network          Next Hop            Metric LocPrf Weight Path
> *  0.0.0.0          6x.xxx.xxx.xx7           0    200      0 xx9 i
> *>                  1xx.xx.xxx.1                  200      0 xxxx1 i
> *                   1xx.xxx.xxx.2                 200      0 xxxx2 i
> 
> #sh ip bgp 0.0.0.0 0.0.0.0
> BGP routing table entry for 0.0.0.0/0, version 1212671
> Paths: (7 available, best #3, table Default-IP-Routing-Table)
>    Not advertised to any peer
>    xx9
>      6x.xxx.xxx.xx7 from 6x.xxx.xxx.xx7 (2xx.xxx.xxx.xx0)
>        Origin IGP, metric 0, localpref 200, valid, external
>    xx9, (received-only)
>      6x.xxx.xxx.xx7 from 6x.xxx.xxx.xx7 (2xx.xxx.xxx.xx0)
>        Origin IGP, metric 0, localpref 100, valid, external
>    xxxx1
>      1xx.xx.xxx.1 from 1xx.xx.xxx.1 (1xx.xx.xxx.1)
>        Origin IGP, localpref 200, valid, external, best
>    xxxx1, (received-only)
>      1xx.xx.xxx.1 from 1xx.xx.xxx.1 (1xx.xx.xxx.1)
>        Origin IGP, localpref 100, valid, external
>    xxxx2
>      1xx.xxx.xxx.2 from 1xx.xxx.xxx.2 (2xx.xxx.xx.x1)
>        Origin IGP, localpref 200, valid, external
>    xxxx2, (received-only)
>      1xx.xxx.xxx.2 from 1xx.xxx.xxx.2 (2xx.xxx.xx.x1)
>        Origin IGP, localpref 100, valid, external
>    Local, (received-only)
>      2xx.xxx.xxx.xx3 from 2xx.xxx.xxx.xx3 (2xx.x.xxx.x4)
>        Origin IGP, metric 0, localpref 100, valid, internal
> 
> #sh ip ro 0.0.0.0
> Routing entry for 0.0.0.0/0, supernet
>    Known via "bgp 13941", distance 20, metric 0, candidate default path
>    Tag xxxx1, type external
>    Last update from 1xx.xx.xxx.1 00:22:31 ago
>    Routing Descriptor Blocks:
>    * 1xx.xx.xxx.1, from 1xx.xx.xxx.1, 00:22:31 ago
>        Route metric is 0, traffic share count is 1
>        AS Hops 1
> 
> I can manually add static default routes and it load balances then, but 
> this is not what I want...
> 

It's not isolated to just default routes, multipath has a list of
requirements for any path to be considered equal.  In general:


	In order to be candidates for multipath, paths to the same 			destination
need to have these characteristics equal to the 			best-path characteristics:

	- Weight
	- Local preference
 	- AS-PATH length
	- Origin
	- MED
	- One of these:
		- Neighboring AS or sub-AS (before the addition of the 				eiBGP
Multipath feature)
		- AS-PATH (after the addition of the eiBGP Multipath 				feature)

	Additional requirements for eBGP multipath:
	
	- The path should be learned from an external or 			
	confederation-external neighbor (eBGP).
	- The IGP metric to the BGP next hop should be equal to the 			best-path
IGP metric.

In this case, neither the neighboring AS's nor the complete AS-PATH of
other paths are equivalent to the selected best path's.

See
http://www.cisco.com/en/US/tech/tk365/technologies_tech_note09186a0080094431.shtml#bgpmpath

for more info.

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

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.0 (MingW32)

iD8DBQFDfX2dE1XcgMgrtyYRAn6BAKDl0oyTYOEuANGyE5AE7kS7QsqYNACfTOvN
u4bIc7eNfHMbKdHmjcodINA=
=vk+e
-----END PGP SIGNATURE-----