From: Manny Gonzalez (manny@xxxxxxx)
Date: Sun Nov 04 2001 - 04:11:08 GMT-3
How about trying this for laughs:
Put IGRP on A ---- B
Do the redist only on A and see what C gets from B...
Alternatively... before doing above, try putting a STATIC floater of the
same ADMIN DISTANCE as IGRP (100 right?) to the same net (the quad-zero)
and see what happens. I have seen many strange things w/ IOS but IGRP
and EIGRP seem to always have some trouble or two. Also, try a IOS
upgrade/downgrade just for G.P. I know that when all else fails, it may
be a IOS "feature" (bug) Sort of like the 12.0.7T always showing ISDN
SPIDS are NOT VALID when they are valid... :-)
John Neiberger wrote:
>
> Yes. I only did that because it was suggested a few times in the list
> archives. It seems to me that it ought to be irrelevant because that
> loopback network address should just be used as a default network when
> advertised to C, right? 30.0.0.0/8 is a dummy network that doesn't exist
> elsewhere in the network.
>
> I can see that it is a candidate GOLR on B as well but since I have a
> 0.0.0.0/0 route available on B it ought to be taken. As you can see from
> the output, that default route is taken when I try to ping 20.1.1.1. After
> I add the default-network command the router will no longer use the route
> available to it in the routing table. This just doesn't make sense to me.
>
> If the route is there--especialy since ip classless is configured--that
> route should be taken.
>
> Regards,
> John
>
> On Sun, 04 Nov 2001 01:41:42 -0500, Manny Gonzalez wrote:
>
> | Have you tried NOT using a default-network that points to a LOOPBACK?
> |
> | John Neiberger wrote:
> | >
> | > I'll start with the original scenario that worked so I can show you
> where I
> | > began before I show you what I'm trying to accomplish now. There are
> three
> | > relevant routers here:
> | >
> | > A----(ospf)----B----(rip)-----C
> | >
> | > A originates a default route to B and I use default-information
> originate in
> | > the RIP config to pass 0.0.0.0/0 to C. This works well. Then I took
> RIP
> | > away and tried this with IGRP and ip default-network.
> | >
> | > This took some tweaking before I could get B to originate default
> route to
> | > C with IGRP. Is it just me or did Cisco seem to make this very
> | > user-unfriendly?? Unbelievable. This is *so* easy with other
> protocols.
> | > Anyway...
> | >
> | > In the first scenario, B has a single gateway of last resort:
> 0.0.0.0/0 via
> | > router A. Beautiful. In the second scenario I end up with two
> candidate
> | > GOLRs but neither is picked and routing breaks!
> | >
> | > This makes *zero* sense to me. If ip classless is configured and
> still
> | > have 0.0.0.0/0 in my routing table then B should route all packets with
> | > unknown destinations to A, right?? Well, it's not working and I can
> | > consistently recreate it.
> | >
> | > If I remove the ip default-network statement routing works but then C
> has no
> | > default route.
> | >
> | > What could be wrong here? For grins, I'll paste in some command output
> to
> | > show you what I mean. R4 is "Router B" in the above scenario.
> | >
> | > Gateway of last resort is 152.1.3.2 to network 0.0.0.0
> | >
> | > 152.1.0.0/16 is variably subnetted, 2 subnets, 2 masks
> | > O IA 152.1.1.0/25 [110/74] via 152.1.3.2, 05:19:53, Serial0
> | > C 152.1.3.0/30 is directly connected, Serial0
> | > 130.1.0.0/16 is variably subnetted, 9 subnets, 2 masks
> | > I 130.1.3.0/24 [100/1188] via 130.1.4.2, 00:00:28, TokenRing0
> | > I 130.1.2.0/24 [100/1188] via 130.1.4.2, 00:00:28, TokenRing0
> | > I 130.1.1.0/24 [100/1188] via 130.1.4.2, 00:00:28, TokenRing0
> | > O 130.1.0.0/22 is a summary, 05:19:54, Null0
> | > I 130.1.7.0/24 [100/1188] via 130.1.4.2, 00:00:29, TokenRing0
> | > I 130.1.6.0/24 [100/1188] via 130.1.4.2, 00:00:29, TokenRing0
> | > I 130.1.5.0/24 [100/1188] via 130.1.4.2, 00:00:29, TokenRing0
> | > O 130.1.4.0/22 is a summary, 05:19:54, Null0
> | > C 130.1.4.0/24 is directly connected, TokenRing0
> | > C 30.0.0.0/8 is directly connected, Loopback1
> | > O*N2 0.0.0.0/0 [110/1] via 152.1.3.2, 05:19:56, Serial0
> | > R4#ping 20.1.1.1
> | >
> | > Type escape sequence to abort.
> | > Sending 5, 100-byte ICMP Echos to 20.1.1.1, timeout is 2 seconds:
> | > !!!!!
> | > Success rate is 100 percent (5/5), round-trip min/avg/max = 24/28/40 ms
> | > R4#
> | >
> | > After I add ip default-network 30.0.0.0:
> | >
> | > Gateway of last resort is not set
> | >
> | > 152.1.0.0/16 is variably subnetted, 2 subnets, 2 masks
> | > O IA 152.1.1.0/25 [110/74] via 152.1.3.2, 05:21:19, Serial0
> | > C 152.1.3.0/30 is directly connected, Serial0
> | > 130.1.0.0/16 is variably subnetted, 9 subnets, 2 masks
> | > I 130.1.3.0/24 [100/1188] via 130.1.4.2, 00:00:32, TokenRing0
> | > I 130.1.2.0/24 [100/1188] via 130.1.4.2, 00:00:32, TokenRing0
> | > I 130.1.1.0/24 [100/1188] via 130.1.4.2, 00:00:32, TokenRing0
> | > O 130.1.0.0/22 is a summary, 05:21:19, Null0
> | > I 130.1.7.0/24 [100/1188] via 130.1.4.2, 00:00:34, TokenRing0
> | > I 130.1.6.0/24 [100/1188] via 130.1.4.2, 00:00:34, TokenRing0
> | > I 130.1.5.0/24 [100/1188] via 130.1.4.2, 00:00:34, TokenRing0
> | > O 130.1.4.0/22 is a summary, 05:21:20, Null0
> | > C 130.1.4.0/24 is directly connected, TokenRing0
> | > C* 30.0.0.0/8 is directly connected, Loopback1
> | > O*N2 0.0.0.0/0 [110/1] via 152.1.3.2, 05:21:22, Serial0
> | > R4#
> | > R4#ping 20.1.1.1
> | >
> | > Type escape sequence to abort.
> | > Sending 5, 100-byte ICMP Echos to 20.1.1.1, timeout is 2 seconds:
> | > .....
> | > Success rate is 0 percent (0/5)
> | > R4#
> | >
> | > Any help would be appreciated. I'm about to swear off using IGRP and
> EIGRP
> | > for the rest of my life just on principle. :-)
> | >
> | > Thanks,
> | > John
> | >
This archive was generated by hypermail 2.1.4 : Fri Jun 21 2002 - 06:45:03 GMT-3