Redistribution.

cisco_trooper

I am finally getting back into the swing of things here. The IS-IS induced sabbatical has come to an end. This is the story of redistribution.

The first minute:
I have created an 8 router lab in which eigrp is being redistributed into ospf, and ospf into eigrp. Looking at the output of show ip ospf border-routers, my initial thought is that I have created a routing loop since the redistribution is occuring on two different routers. I will let you know what I come up with.

RTR_H#sho ip ospf border-routers

OSPF Process 10 internal Routing Table

Codes: i - Intra-area route, I - Inter-area route

I 10.100.100.2 [65] via 172.20.78.1, FastEthernet1/0, ASBR, Area 1, SPF 2
I 10.100.100.2 [129] via 172.20.68.1, Serial2/1, ASBR, Area 1, SPF 2
I 10.100.100.2 [129] via 172.20.78.1, FastEthernet1/0, ASBR, Area 1, SPF 2
I 10.100.100.0 [129] via 172.20.68.1, Serial2/1, ASBR, Area 1, SPF 2
I 10.100.100.0 [129] via 172.20.78.1, FastEthernet1/0, ASBR, Area 1, SPF 2
I 10.100.100.0 [65] via 172.20.78.1, FastEthernet1/0, ASBR, Area 1, SPF 2
i 10.100.100.4 [65] via 172.20.78.1, FastEthernet1/0, ABR, Area 1, SPF 2
i 10.100.100.4 [65] via 172.20.68.1, Serial2/1, ABR, Area 1, SPF 2
i 10.100.100.6 [1] via 172.20.78.1, FastEthernet1/0, ABR, Area 1, SPF 2

cisco_trooper

I have decided now, having looked at this again while being actually awake, that I was not thinking clearly in my initial assumption that I have created a routing loop. There is simply two paths to these networks. Who would have thought sleep deprivation could affect you so much?
Anyway...on with my day.

cisco_trooper

I'm looking to introduce problems into the network. It isn't working out so well. It seems redistribution works pretty well between OSPF and EIGRP since external EIGRP routes have an AD of 170, which is higher than OSPF and OSPF external routes are marked as external, with preference given to internal routes. If anyone has an idea about how I can make this scenario fail, please feel free to enlighten me, but so far this seems too good to be true. I'm certain this scenario has a weakness, but I have yet to find it.

cisco_trooper

I have succeeded in breaking this topology. I shutdown two of the edge routers to bring their attached networks down. One of the redistributing routers then entered an invalid route to this network into its routing table. Probably learned from the redistributing protocol. 11 minutes later that route is still in the routing table, failing to correct itself. Several other routers also have invalid routes to this network. At this point it is clear a problem exists. This is fantastic. I knew it was working too well to be true. The point of all this is, if you know how to break it, you know what not to do.

cisco_trooper

Redistribution is occurring over Frame Relay PVCs, and IP split horizon is disabled on these interfaces by default, which seems to be at the root of the invalid LSAs continuing to be propogated throughout the OSPF routing domain. Lesson learned. I don't know if this is something I used to know, but I certainly will not be forgetting it again anytime soon.

cisco_trooper

Apparently I'm the only one interested in what I think, and the routers tend to agree with this on this lonely night.

networker050184

You should implement distribute lists to keep the routes learned from a protocol being advertised back into that protocol which will help prevent loops.

Also OSPF doesnt use split horizon rules and it is not diabled by default for EIGRP.

cisco_trooper

Thank goodness. Thought I was the only one blowing a Saturday night making love to my Cisco equipment. LOL.

Yeah, I haven't got to the distribute lists and route maps yet. I had to figure out how to make the topology fail first. It wasn't really necessary but I just HAD to.

Mrock4

How dare you assume you're the only one wasting saturday night on cisco! I am too......well, dynamips..since it uses real IOS images, maybe it still counts, right?

I actually have been having trouble with a certain redistribution scenario myself lately. It seems almost to be a bug in dynamips. Just to be sure, I am going to start everything from scratch here in a few minutes, and spend my saturday night figuring it out.

I'll have to re-read through this post to double check myself..

Edit: Just noticed you're in Overland Park...I'm from Columbia, MO. Are you enjoying the lovely weather ?

cisco_trooper

Mrock4 wrote:

How dare you assume you're the only one wasting saturday night on cisco! I am too......well, dynamips..since it uses real IOS images, maybe it still counts, right?

I actually have been having trouble with a certain redistribution scenario myself lately. It seems almost to be a bug in dynamips. Just to be sure, I am going to start everything from scratch here in a few minutes, and spend my saturday night figuring it out.

I'll have to re-read through this post to double check myself..

Edit: Just noticed you're in Overland Park...I'm from Columbia, MO. Are you enjoying the lovely weather ?

Yeah man. My favorite part is the 2 degree wind chill.

Mrock4

Haha, I bet. Once my time is done I'll be heading back to live (most likely) in a suburb near Kansas City. I really like Kansas City, but I'm more fond of living in a less populated area. I can't wait to be honest with you! I miss the midwest.

Figured my issues out last night..I was pretty much being retarded with my redistribution scenario. I was almost disappointed that I figured it out so quickly. I went on to practice my EIGRP authentication.

cisco_trooper

Interesting observation I never noticed before. Redistributing OSPF into EIGRP, I noticed that when I summarized the routes coming from OSPF, they were no longer marked as external routes and had an AD of 90. That's all I have for now.

Mrock4

hmm..I think you just created my lab for tonight. I've gotta test that now.

dtlokee

Once you create the aggregate it's no longer an external route but now an EIGRP originated route, and it will supress the advertisment of all routes that match the aggegate (unless you specify a "leak-map"). Internesting that it showed up with an AD of 90 though, I would have expected 5.

cisco_trooper

Hmm. I wonder if I'm the only one working Doyles book the same time as the BSCI book....

Mrock4

I have doyle's book..don't know if it'd be fair to say I'm "working" it but......slowly..

jezg76

Mrock4 wrote:

I have doyle's book..don't know if it'd be fair to say I'm "working" it but......slowly..

I hear you on that. That book definitely lets you know where you stand.

cisco_trooper

Doyle's book is the shizzle. The level of detail this book offers brings to me a certain level of confidence I was not getting out of the BSCI book alone. While it is true that it does let you know where you stand, which can be frightening, I find it a great relief that this book fills in some major gaps of knowledge that I wasn't getting out of the BSCI book. Reading the BSCI book I always dream up a lab that can't be implemented with the scope of the topics covered. Doyle corrects this problem, BIG TIME. Redistribution is much more interesting than BSCI leads you to believe fellas.

cisco_trooper

BBBLLLAAAHHH!!!

Do you guys find that by the time you get near the end of a book you really start forgetting the crap from the beginning of the book? Maybe I should come up with better labs...

cisco_trooper

Does OSPF ignore subnet masks on loopback interfaces?

I have loopback interfaces with subnet mask 255.255.255.224 being advertised via OSPF. Other routers receiving these routes are only getting the interface IP address with a 32 bit mask. Never noticed this before....

jezg76

I always use the ip ospf network point-to-point interface config command to make it advertise the mask I want when using a loopback.

cisco_trooper

jezg76 wrote:

I always use the ip ospf network point-to-point interface config command to make it advertise the mask I want when using a loopback.

MUUHAHAHAHAAAAHHHHUUUAAAAA!!!

That works...saves me from creating a bunch of networks.

jezg76

cisco_trooper wrote:

BBBLLLAAAHHH!!!

Do you guys find that by the time you get near the end of a book you really start forgetting the crap from the beginning of the book? Maybe I should come up with better labs...

I do agree with this 100%. I'm trying to find the best way to remember old stuff as I dive into new stuff. I do know having a stack of ever-growing notecards to whip out whenever has helped a bunch with the memorization aspect of this test.

dtlokee

Yes, according to the RFC a loopback is advertised as a /32 regardless of the subnet mask in use. One way and the easiest is to use the ip ospf network point-to-point command as already mentioned. Other ways include redistributing the loopbacks into OSPF and putting the loopback into it's own area and using the "area x range" command to summarize the area into ospf. All sorts of fun.

cisco_trooper

dtlokee wrote:

Once you create the aggregate it's no longer an external route but now an EIGRP originated route, and it will supress the advertisment of all routes that match the aggegate (unless you specify a "leak-map"). Internesting that it showed up with an AD of 90 though, I would have expected 5.

You know, that is weird come to think of it. Check it out, this is the config from the redistributing router:

! Robinson config
interface FastEthernet1/1
ip address 192.168.3.129 255.255.255.224
ip summary-address eigrp 1 192.168.4.0 255.255.255.0 5 THIS IS AD OF 5
ip summary-address eigrp 1 192.168.3.0 255.255.255.128 5 THIS IS AD OF 5
ip summary-address eigrp 1 172.16.0.0 255.255.0.0 5 THIS IS AD OF 5
duplex auto
speed auto
!
router eigrp 1
redistribute ospf 1 metric 1000 100 1 255 1500
redistribute eigrp 2
passive-interface FastEthernet1/0
network 192.168.3.0
auto-summary
!
!
Robinson#sho ip route
Codes: C - connected, S - static, R - RIP, M - mobile, B - BGP
D - EIGRP, EX - EIGRP external, O - OSPF, IA - OSPF inter area
N1 - OSPF NSSA external type 1, N2 - OSPF NSSA external type 2
E1 - OSPF external type 1, E2 - OSPF external type 2
i - IS-IS, su - IS-IS summary, L1 - IS-IS level-1, L2 - IS-IS level-2
ia - IS-IS inter area, * - candidate default, U - per-user static route
o - ODR, P - periodic downloaded static route

Gateway of last resort is not set

172.16.0.0/16 is variably subnetted, 3 subnets, 3 masks
C 172.16.2.20/30 is directly connected, Serial2/0.103
D 172.16.0.0/16 is a summary, 00:12:21, Null0
D 172.16.1.0/24 [90/2297856] via 172.16.2.22, 00:47:08, Serial2/0.103
192.168.4.0/24 is variably subnetted, 3 subnets, 3 masks
D 192.168.4.72/29 [90/2297856] via 192.168.4.6, 00:46:40, Serial2/0.104
C 192.168.4.4/30 is directly connected, Serial2/0.104
D 192.168.4.0/24 is a summary, 00:11:56, Null0
192.254.0.0/32 is subnetted, 1 subnets
C 192.254.0.1 is directly connected, Loopback0
O 192.168.1.0/24 [110/2] via 192.168.3.34, 00:47:21, FastEthernet1/0
D 192.168.2.0/24 [90/156160] via 192.168.3.130, 00:46:57, FastEthernet1/1
192.168.3.0/24 is variably subnetted, 9 subnets, 3 masks
O 192.168.3.96/27 [110/2] via 192.168.3.34, 00:47:22, FastEthernet1/0
O 192.168.3.64/27 [110/2] via 192.168.3.34, 00:47:22, FastEthernet1/0
C 192.168.3.32/27 is directly connected, FastEthernet1/0
D 192.168.3.0/25 is a summary, 00:16:50, Null0
D 192.168.3.0/24 is a summary, 00:10:56, Null0
D 192.168.3.192/27
[90/156160] via 192.168.3.130, 00:46:58, FastEthernet1/1
D 192.168.3.160/27
[90/156160] via 192.168.3.130, 00:46:58, FastEthernet1/1
C 192.168.3.128/27 is directly connected, FastEthernet1/1
O 192.168.3.128/25 is a summary, 00:12:20, Null0
D 192.168.0.0/16 is a summary, 00:09:32, Null0

Podres#sho ip route
Codes: C - connected, S - static, R - RIP, M - mobile, B - BGP
D - EIGRP, EX - EIGRP external, O - OSPF, IA - OSPF inter area
N1 - OSPF NSSA external type 1, N2 - OSPF NSSA external type 2
E1 - OSPF external type 1, E2 - OSPF external type 2
i - IS-IS, su - IS-IS summary, L1 - IS-IS level-1, L2 - IS-IS level-2
ia - IS-IS inter area, * - candidate default, U - per-user static route
o - ODR, P - periodic downloaded static route

Gateway of last resort is not set

D 172.16.0.0/16 [90/2172416] via 192.168.3.129, 00:13:39, FastEthernet1/1
D 192.168.4.0/24 [90/2172416] via 192.168.3.129, 00:13:14, FastEthernet1/1
192.254.0.0/32 is subnetted, 1 subnets
C 192.254.0.5 is directly connected, Loopback0
D EX 192.168.1.0/24 [170/2588160] via 192.168.3.129, 00:48:16, FastEthernet1/1
C 192.168.2.0/24 is directly connected, Loopback3
192.168.3.0/24 is variably subnetted, 6 subnets, 3 masks
D 192.168.3.0/25 [90/30720] via 192.168.3.129, 00:18:08, FastEthernet1/1
D 192.168.3.0/24 is a summary, 00:12:13, Null0
C 192.168.3.192/27 is directly connected, Loopback2
C 192.168.3.160/27 is directly connected, Loopback1
C 192.168.3.128/27 is directly connected, FastEthernet1/1
D EX 192.168.3.128/25
[170/2588160] via 192.168.3.129, 00:48:16, FastEthernet1/1
D EX 192.168.0.0/16 [170/30720] via 192.168.3.129, 00:10:51, FastEthernet1/1


The Summarized Routes are in bold. According to the summarization taking place on Robinson's FA1/1 interface these SHOULD have an AD of 5. Now I don't have route maps and all that good stuff in place in this lab yet, but an AD of 5 should be the preferred path over any routes "accidentally" learned. I have to admit I'm kinda curious what is going on here. WHY AREN'T THESE ROUTES AD OF 5?!?!? Surely it isn't because of the lack of distribution filtering...

cisco_trooper

Oh, and for those curious, this is basically the network in Doyles book. Obviously the redistribution chapter. Chapter 11.

cisco_trooper

OK. I just referenced Doyle's Chapter on EIGRP and his example of route summarization has the summarized routes in the routing table with an AD of 90. This is beginning to piss me off. It is documented all over the place that EIGRP summary routes have an AD of 5, and this is even specified in the summarization command by default. W T F!?!?!?

networker050184

It will only have an AD of 5 on the local router. AD is asigned by the local router and not sent to neighbors.

cisco_trooper

networker050184 wrote:

It will only have an AD of 5 on the local router. AD is asigned by the local router and not sent to neighbors.

This doesn't appear to show up in the routing table, however when I do sho ip route X.X.X.X
for the summary route it does show a distance of 5. I guess we have to trust it is using this route. I was really expecting this to show up in the routing table.

! Robinson Config
interface FastEthernet1/1
ip address 192.168.3.129 255.255.255.224
ip summary-address eigrp 1 192.168.4.0 255.255.255.0 5
ip summary-address eigrp 1 192.168.3.0 255.255.255.128 5
ip summary-address eigrp 1 172.16.0.0 255.255.0.0 5
duplex auto
speed auto
!
router eigrp 1
redistribute ospf 1 metric 1000 100 1 255 1500
redistribute eigrp 2
passive-interface FastEthernet1/0
network 192.168.3.0
auto-summary
!
!
Robinson#sho ip route
Codes: C - connected, S - static, R - RIP, M - mobile, B - BGP
D - EIGRP, EX - EIGRP external, O - OSPF, IA - OSPF inter area
N1 - OSPF NSSA external type 1, N2 - OSPF NSSA external type 2
E1 - OSPF external type 1, E2 - OSPF external type 2
i - IS-IS, su - IS-IS summary, L1 - IS-IS level-1, L2 - IS-IS level-2
ia - IS-IS inter area, * - candidate default, U - per-user static route
o - ODR, P - periodic downloaded static route

Gateway of last resort is not set

172.16.0.0/16 is variably subnetted, 3 subnets, 3 masks
C 172.16.2.20/30 is directly connected, Serial2/0.103
D 172.16.0.0/16 is a summary, 00:12:21, Null0
D 172.16.1.0/24 [90/2297856] via 172.16.2.22, 00:47:08, Serial2/0.103
192.168.4.0/24 is variably subnetted, 3 subnets, 3 masks
D 192.168.4.72/29 [90/2297856] via 192.168.4.6, 00:46:40, Serial2/0.104
C 192.168.4.4/30 is directly connected, Serial2/0.104
D 192.168.4.0/24 is a summary, 00:11:56, Null0
192.254.0.0/32 is subnetted, 1 subnets
C 192.254.0.1 is directly connected, Loopback0
O 192.168.1.0/24 [110/2] via 192.168.3.34, 00:47:21, FastEthernet1/0
D 192.168.2.0/24 [90/156160] via 192.168.3.130, 00:46:57, FastEthernet1/1
192.168.3.0/24 is variably subnetted, 9 subnets, 3 masks
O 192.168.3.96/27 [110/2] via 192.168.3.34, 00:47:22, FastEthernet1/0
O 192.168.3.64/27 [110/2] via 192.168.3.34, 00:47:22, FastEthernet1/0
C 192.168.3.32/27 is directly connected, FastEthernet1/0
D 192.168.3.0/25 is a summary, 00:16:50, Null0
D 192.168.3.0/24 is a summary, 00:10:56, Null0
D 192.168.3.192/27
[90/156160] via 192.168.3.130, 00:46:58, FastEthernet1/1
D 192.168.3.160/27
[90/156160] via 192.168.3.130, 00:46:58, FastEthernet1/1
C 192.168.3.128/27 is directly connected, FastEthernet1/1
O 192.168.3.128/25 is a summary, 00:12:20, Null0
D 192.168.0.0/16 is a summary, 00:09:32, Null0
!
Robinson#sho ip route 192.168.4.0 255.255.255.0
Routing entry for 192.168.4.0/24
Known via "eigrp 1", distance 5, metric 2169856, type internal
Redistributing via eigrp 1, ospf 1, eigrp 2
Advertised by ospf 1 subnets
eigrp 2
Routing Descriptor Blocks:
* directly connected, via Null0
Route metric is 2169856, traffic share count is 1
Total delay is 20000 microseconds, minimum bandwidth is 1544 Kbit
Reliability 255/255, minimum MTU 1500 bytes
Loading 1/255, Hops 0
!

qplayed

doyle? can i get an isbn?

networker050184

Routing TCP/IP Volume 1 Second ISBN# 1587052024

Routing TCP/IP Volume 2 ISBN# 1578700892

Very good books from all the reviews I have heard. Just started volume one a few days ago and its great so far.