From: James D. Holley, Jr. (jholley@xxxxxxxxxxx)
Date: Fri Feb 18 2000 - 21:46:42 GMT-3
I found something interesting about the command "summary address
x.x.x.x 255.255.255.0" when trying to redistribute VLSM subnet into
FLSM. See diagram:
[R1]S0-------------S0[R2]S1-------------S0[R3]S1------------
RIP[-----------------] [OSPF AREA 0] [OSPF AREA 1]
R1 is running RIP S0- 172.16.1.x/24
R2 is running RIP S0- 172.16.1.x/24 and S1- OSPF 172.16.2.4/30 Area 0
Mutually Redistributing OSPF and RIP at R2
R3 is running OSPF . S0- 172.16.2.4/30 and S1 172.16.3.0/24
R1
!
S0
ip add 172.16.1.1 255.255.255.0
!
router rip
net 172.16.0.0
R2
!
S0
ip add 172.16.1.2 255.255.255.0
!
S1
ip add 172.16.2.5 255.255.255.252
!
router rip
net 172.16.0.0
redist ospf 1 met 3
pass s1
distribute-list 1 out ospf 1
!
router ospf 1
net 172.16.2.4 0.0.0.3 area 0
net 172.16.3.0 0.0.0.255 area 1
summary-address 172.16.2.0 255.255.255.0
redistr rip sub met 1000
distr list 2 rip out
access-list 1 perm 172.16.2.0
access-list 1 perm 172.16.3.0
!
access-list 2 permit 172.16.1.0
The above config, if typed in correctly from memory (not a machine
example), will summarize the VLSM 172.16.2.4/30 into a /24 for RIP to
receive, but the "distribute-list 1 out ospf 1" will block the net
from being advertised.
CASE:
Before the distribute list is applied, when the summary-address
command is issued a type-5 LSA is generated in the OSPF Database on
the ASBR (R2). You can see it by typing in "sh ip ospf data".
Redistribution works perfectly.
Now, if you apply the distribute-list 1, above and "debug ip ospf LSA"
you can see that the LSA is becoming MAXAGED, and flushed. The
distribute list, access-list 1, above is blocking the 172.16.2.4 net
when the redistribution is done- (it doesn't match criteria). If
you alter the access-list to explicitly let it in, or deny some routes
and permit any, it will work. The following is an example of explicit
access-list:
Access-list 1 perm 172.16.2.0 0.0.1.255
This may be why sometimes the command works and sometimes it seems it
doesn't.
(If you got this far, thank you. ) Thoughts?
Special Thanks to Colby, Ed, and Scott at BroadWing's CCIE Practice
Lab in Reston, VA (Look for my review of their LAB next week.)
James D. Holley, Jr.
Sprint ENS
This archive was generated by hypermail 2.1.4 : Thu Jun 13 2002 - 08:22:53 GMT-3