RE: Strange issues with routing

From: Brian Dennis (bdennis@internetworkexpert.com)
Date: Wed Jan 24 2007 - 20:35:06 ART


Basic route redistribution on a single router is a two step process.
The first step is to bring in the routes from the routing table for the
protocol being redistributed (show ip route <protocol>). The second
step is to bring in the connected interfaces that the protocol is
enabled on (i.e. network statements).

When you manually redistribute connected with a route-map you are
breaking the second step that the router is doing for you automatically.
This second step was done when you redistributed, in your case, OSPF
into EIGRP. You need to match the interfaces that the router was
bringing in for you automatically in your route-map.

If you remember these two steps it will save you a lot of headaches
going forward ;-)

Brian Dennis, CCIE4 #2210 (R&S/ISP-Dial/Security/SP)
bdennis@internetworkexpert.com
 
Internetwork Expert, Inc.
http://www.InternetworkExpert.com
Toll Free: 877-224-8987
Direct: 775-745-6404 (Outside the US and Canada)

 

-----Original Message-----
From: nobody@groupstudy.com [mailto:nobody@groupstudy.com] On Behalf Of
Bliz Man
Sent: Wednesday, January 24, 2007 1:46 PM
To: ccielab@groupstudy.com
Subject: Strange issues with routing

Hello group!
I've got following simple topology:

Sw1 -----vlan10---- f0/0 R3 s0/0
 (svi) |
                 |---------f0/0 R4 s0/0

EIGRP: f0/0@R3,R4, SVI on Sw1, OSPF - on s0/0@R3,R4

I'm doing mutual redistribution on R3 and R4 :

CCIE-R4#sh run | be router
router eigrp 1
 redistribute ospf 1 metric 1000 100 255 1 1500
 network 1.1.4.4 0.0.0.0
 network 1.1.10.4 0.0.0.0
 no auto-summary
!
router ospf 1
 router-id 1.1.4.4
 log-adjacency-changes
 redistribute eigrp 1 subnets
 network 1.1.4.4 0.0.0.0 area 0
 network 1.1.11.4 0.0.0.0 area 0

CCIE-R3#sh run | be router
router eigrp 1
 redistribute ospf 1 metric 1000 100 255 1 1500
 network 1.1.3.3 0.0.0.0
 network 1.1.10.3 0.0.0.0
 no auto-summary
!
router ospf 1
 router-id 1.1.3.3
 log-adjacency-changes
 redistribute eigrp 1 subnets
 network 1.1.3.3 0.0.0.0 area 0
 network 1.1.11.3 0.0.0.0 area 0

And on Sw1 I have 2 routes to 1.1.11.0/24. - simple

CCIE-SW1#sir 1.1.11.1
Routing entry for 1.1.11.0/29
  Known via "eigrp 1", distance 170, metric 2585856, type external
  Redistributing via eigrp 1
  Last update from 1.1.10.4 on Vlan13, 00:31:07 ago
  Routing Descriptor Blocks:
    1.1.10.4, from 1.1.10.4, 00:31:07 ago, via Vlan13
      Route metric is 2585856, traffic share count is 1
      Total delay is 1010 microseconds, minimum bandwidth is 1000 Kbit
      Reliability 255/255, minimum MTU 1500 bytes
      Loading 1/255, Hops 1
  * 1.1.10.3, from 1.1.10.3, 00:31:07 ago, via Vlan13
      Route metric is 2585856, traffic share count is 1
      Total delay is 1010 microseconds, minimum bandwidth is 1000 Kbit
      Reliability 255/255, minimum MTU 1500 bytes
      Loading 1/255, Hops 1

BUT...
When I add some loopback on R4: lo5 on R4

interface Loopback5
 ip address 5.5.5.5 255.255.255.255
end

And try to redistribute it to EIGRP through simple route-map as
connected:

CCIE-R4#sh route-map CONN
route-map CONN, permit, sequence 10
  Match clauses:
    interface Loopback5
  Set clauses:
  Policy routing matches: 0 packets, 0 bytes

And add this string to R3 config under EIGRP: redistribute connected
metric
1000 100 255 1 1500 route-map CONN
on Sw1 one route to 1.1.11.0/24 removed from RIB

CCIE-SW1#sir 1.1.11.1
Routing entry for 1.1.11.0/29
  Known via "eigrp 1", distance 170, metric 2585856, type external
  Redistributing via eigrp 1
  Last update from 1.1.10.3 on Vlan13, 00:02:15 ago
  Routing Descriptor Blocks:
  * 1.1.10.3, from 1.1.10.3, 00:02:15 ago, via Vlan13
      Route metric is 2585856, traffic share count is 1
      Total delay is 1010 microseconds, minimum bandwidth is 1000 Kbit
      Reliability 255/255, minimum MTU 1500 bytes
      Loading 1/255, Hops 1

At that moment I have "debug eigrp fsm" on Sw1:

CCIE-SW1#debug eigrp fsm
EIGRP FSM Events/Actions debugging is on
CCIE-SW1#
1w2d: DUAL: rcvquery: 1.1.11.0/29 via 1.1.10.4 metric
4294967295/4294967295,
RD is 2585856
1w2d: DUAL: Find FS for dest 1.1.11.0/29. FD is 2585856, RD is 2585856
1w2d: DUAL: 1.1.10.4 metric 4294967295/4294967295
1w2d: DUAL: 1.1.10.3 metric 2585856/2585600 found Dmin is 2585856
1w2d: DUAL: send REPLY(r1/n1) about 1.1.11.0/29 to 1.1.10.4
1w2d: DUAL: RT installed 1.1.11.0/29 via 1.1.10.3
1w2d: DUAL: dest(5.5.5.5/32) not active
1w2d: DUAL: rcvupdate: 5.5.5.5/32 via 1.1.10.4 metric 2585856/2585600
1
CCIE-SW1#w2d: DUAL: Find FS for dest 5.5.5.5/32. FD is 4294967295, RD is
4294967295 found
1w2d: DUAL: RT installed 5.5.5.5/32 via 1.1.10.4
1w2d: DUAL: Send update about 5.5.5.5/32. Reason: metric chg
1w2d: DUAL: Send update about 5.5.5.5/32. Reason: new if
1w2d: DUAL: Removing dest 1.1.11.0/29, nexthop 1.1.10.4

And when I remove "redistr connected" on R4 :

1w2d: DUAL: rcvquery: 5.5.5.5/32 via 1.1.10.4 metric
4294967295/4294967295,
RD is 2585856
1w2d: DUAL: Find FS for dest 5.5.5.5/32. FD is 2585856, RD is 2585856
1w2d: DUAL: 1.1.10.4 metric 4294967295/4294967295 not found Dmin is
4294967295
1w2d: DUAL: Peer total/stub 3/0 template/full-stub 3/0
1w2d: DUAL: Dest 5.5.5.5/32 entering active state.
1w2d: DUAL: Set reply-status table. Count is 1.
1w2d: DUAL: Doing split horizon on Vlan13
1w2d: DUAL: Going from state 1 to state 3
1w2d: DUAL: dest(1.1.11
CCIE-SW1#.0/29) not active
1w2d: DUAL: rcvupdate: 1.1.11.0/29 via 1.1.10.4 metric 2585856/2585600
1w2d: DUAL: Find FS for dest 1.1.11.0/29. FD is 2585856, RD is 2585856
1w2d: DUAL: 1.1.10.3 metric 2585856/2585600
1w2d: DUAL: 1.1.10.4 metric 2585856/2585600 found Dmin is 2585856
1w2d: DUAL: RT installed 1.1.11.0/29 via 1.1.10.3
1w2d: DUAL: RT installed 1.1.11.0/29 via 1.1.10.4
1w2d: DUAL: dest(5.5.5.5/32) active
1w2d: DUAL: rcvreply: 5.5.5.5/32 via 1.1.9.6 metric
4294967295/4294967295
1w2d: DUAL: reply co
CCIE-SW1#unt is 1
1w2d: DUAL: Clearing handle 0, count now 0
1w2d: DUAL: Freeing reply status table
1w2d: DUAL: Find FS for dest 5.5.5.5/32. FD is 4294967295, RD is
4294967295
found
1w2d: DUAL: send REPLY(r1/n1) about 5.5.5.5/32 to 1.1.10.4
1w2d: DUAL: Removing dest 5.5.5.5/32, nexthop 1.1.9.6
1w2d: DUAL: Going from state 3 to state 1
1w2d: DUAL: Removing dest 5.5.5.5/32, nexthop 1.1.10.4
1w2d: DUAL: No routes. Flushing dest 5.5.5.5/32

And RIB for route 1.1.11.0/24 follows:

CCIE-SW1#sir 1.1.11.1
Routing entry for 1.1.11.0/29
  Known via "eigrp 1", distance 170, metric 2585856, type external
  Redistributing via eigrp 1
  Last update from 1.1.10.4 on Vlan13, 00:00:36 ago
  Routing Descriptor Blocks:
    1.1.10.4, from 1.1.10.4, 00:00:36 ago, via Vlan13
      Route metric is 2585856, traffic share count is 1
      Total delay is 1010 microseconds, minimum bandwidth is 1000 Kbit
      Reliability 255/255, minimum MTU 1500 bytes
      Loading 1/255, Hops 1
  * 1.1.10.3, from 1.1.10.3, 00:00:36 ago, via Vlan13
      Route metric is 2585856, traffic share count is 1
      Total delay is 1010 microseconds, minimum bandwidth is 1000 Kbit
      Reliability 255/255, minimum MTU 1500 bytes
      Loading 1/255, Hops 1

What's the trick????!!!!



This archive was generated by hypermail 2.1.4 : Thu Feb 08 2007 - 23:46:57 ART