[Xorp-users] XORP RIP enabled. Why are static routes required?

[Mark Handley]

>1. I have done a tcpdump on the interface connecting the two routers. With 
>only RIP enabled I cannot see the RIP traffic from the peer router. 
>However, with static route I can see the peer RIP traffic with the correct 
>routes.

It sounds like the static route is being used for forwarding, and RIP
still isn't working, at least from the debugging output.


>From your config and debug output, it looks like your addressing is:


172.20.1/23                       172.20.2/24                     172.20.3/24
   ----------- Router 1 ------------------------------- Router 2 ----------
      172.20.1.1     172.20.2.1                 172.20.2.2    172.20.3.1
            eth4     eth3                            eth??    eth??


And I believe both the config and the debug output you sent were from
router 1, not from router 2?

Can you confirm that this is the way it's configured, and that it was
router 1's config and debug output you sent, not router 2's?  It's not
completely impossible that you've switched two addresses, or something
similar (it's hard to tell for sure, because you're using
default-system-config, and you didn't send us the results of running
"ip addr" or "ifconfig").

If the diagram above is correct for router 1, then I can't see
anything wrong with the router 1 config.



There is something a little strange in the debug output though:

[ 18291  +413 xrl_redist_manager.cc ] got redist add_route for "connected" 
172.20.1.0/24 172.20.1.1
[ 18291  +413 xrl_redist_manager.cc ] got redist add_route for "static" 
172.20.2.0/24 172.20.2.2
[ 18291  +413 xrl_redist_manager.cc ] got redist add_route for "connected" 
172.20.2.0/24 172.20.2.1
[ 18291  +413 xrl_redist_manager.cc ] got redist add_route for "static" 
172.20.3.0/24 172.20.2.2

The static route for 172.20.2.0/24 nexthop 172.20.2.2 is curious,
given that 172.20.0.0/24 is a directly connected interface.  This
route should not be used, because it would be overridden in the RIB by
the connected route.  If this router's address were 172.20.2.2 instead
of 172.20.2.1, then this would make sense, but it can't be or we
wouldn't see the connected route.  Anyway my expectation is that this
route is a red herring, and that removing this would make no
difference.  If removing it *does* cause the ping to stop working,
then this is a useful clue as to what is wrong.


Of course it's also possible that the problem is not on router 1 at
all, but on router 2, and that this is causing the RIP adjacency to
fail to be set up.

Cheers,
	Mark