Actually, I’m feeling pretty good right now. Once you understand BGP, you understand how the internet ties together, and that is pretty cool. BGP is one of those topics that you read about, that people make you feel is super-complex, but that just isn’t as bad as people say. The rules are fairly straight-forward, you just have to know them. In the real world a BGP configuration can presumably get pretty complex, but for the BSCI, I don’t think what you’re expected to know is difficult at all.  The key is probably having a solid understanding of route-maps, since they are used to manipulate BGP attributes and will probably show their face in an exam Sim question, if you get one on BGP that is.

At this point I am no BGP expert, but for the BSCI, I think I have a good enough grasp of it, and probably a bit more than I need (route reflectors and confederations seem to be outside the scope of CCNP). I’ll just have to make sure to memorize a couple tables over the next couple of weeks. Beyond that, I look forward to Reading Sam Halabi’s BGP bible, but I will save that for when I am done with the CCNP and am preparing for the CCIE written.

Tomorrow I’ll spend the day labbing BGP. I’ll also integrate IGPs and redistribution into my lab scenarios. I’ll probably make one big lab with 9 or 10 routers running different protocols in multiple AS with redistribution, route filtering, etc. This will serve as a good review for everything i’ve learned to date.

Doyle doesn’t need recommendations, but I will give him one just the same. Volume 2 of Routing TCP/IP is just as good as Volume 1. It is well-written, concise, and to the point. It’s a pleasure to read. Then again, I’m a bit of a nerd.

This weekend I got back on track and covered IS-IS, which I really didn’t like very much. Well, I’m almost done with it. I read the chapters on it in both Doyle and the Exam Cert Guide, so I still have the labs in the Lab Portfolio to do. I will probably cover a big chunk of Redistribution today as well, so i’m getting back to my desired pace once again.

Lastly, I think I have said this before, but while the CCNA Cxam Certification Guide by Wendell Odom provided you with everything you needed to pass, the CCNP Exam Certification Library does not. The information in it is very terse, and often unclear unless you’ve already read something that explained the subject in much more detail. Cisco recommends the authorized self-study guide, but I am using Doyle’s Routing TCP/IP for everything. Volume 1 covers the interior routing protocols and some other stuff and Volume 2 covers bgp, multicasting, nat, etc. I had initially only ordered volume 1, but I just ordered Volume 2. I’ve just gone ahead and created a listmania on amazon for those that are interested. It’s specifically designed for people that intend on doing a CCNP and going straight to the CCIE. It distinguishes between what I think is required and what is just recommended, based on my experience with the material so far.

My Book List

The requirements stated that a particular route should not be advertisted out of a particular router. At first I was thinking that I could do something with access lists and network commands, excluding the particular network from routing updates, but I got to thinking how convoluted that would be in a real network, and that there had to be a better way.

So, I went searching the cisco docs and found part of the solution. I had to use something called a distribute-list. The docs referenced using it with an access-list. I thought there had to be a better way still, so I went to my IE Workbook Volume 1 and searched for distribute-list. Sure enough, I found it used in conjuction with something called a prefix list, which wasn’t mentioned in the cisco docs as an option (but was in the inline help).

So here’s how it actually works. The prefix-list basically tells the distribute list which networks can and can’t be advertised. It uses matching in the same way that access lists do. Each entry has a sequence number, and it proceeds from 1->X in order. Once a match is made, it no longer tries to match against any other entries in your prefix list. When used in conjuction with a distribute-list, anything that isn’t matched, is denied.

Here’s the scenario: R2 receives EIGRP updates from R3. R3 has a series of loopbacks configured, and we want R2 to prevent the network on R3′s loopback 3 from being advertised out.

R2 config:

router eigrp 1
network 172.16.0.0
no auto-summary
neighbor 172.16.12.1 Serial0/0

R3 config:

interface Loopback0
ip address 172.16.3.1 255.255.255.224
!
interface Loopback1
ip address 172.16.3.33 255.255.255.224
!
interface Loopback2
ip address 172.16.3.65 255.255.255.224
!
interface Loopback3
ip address 172.16.3.97 255.255.255.224

router eigrp 1
network 172.16.0.0
no auto-summary

Now what we want to do is Filter the 172.16.3.64/27 network from being advertised out of R2. Ok, no sweat. On R2:

ip prefix-list permitOut seq 5 deny 172.16.3.64/27

ip prefix-list permitOut seq 10 permit 172.16.0.0/16

router eigrp 1

distribute-list prefix permitOut out

On R1 (which is connected to R2):

Show 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

172.16.0.0/24 is subnetted, 2 subnets
C       172.16.12.0 is directly connected, Serial0/0
C       172.16.1.0 is directly connected, Loopback0

So what happened? I just wanted to block 172.16.3.64/27 from being advertised, not all EIGRP routes…

There are two things you have to understand about the behavior of distribute lists. The first is that it’s not enough to deny a particular route and expect the rest of your routes to be advertised. You need a permit statement to follow your deny that will explicitly allow all the other routes. The problem with my permit statement is that the ip prefix-list command doesn’t match in the same way that routing protocol network commands do. Matching against 172.16.0.0/16 will ONLY match against a network with the /16 prefix. Consequently, when the router checks 172.16.3.96/27 or 172.16.4.0/25 against 172.16.0.0/16, it will not find a match, and those routes won’t be advertised because there is an implicit deny at the end of the prefix-list. Instead you have to permit all other prefixes. You could do this in a couple of ways:

1. ip prefix-list permitOut seq 10 permit 172.16.0.0/16 ge 17
2. ip prefix-list permitOut seq 10 permit 0.0.0.0/0 le 32

Either of these will work for our requirements in this lab. #2 is probably better in general because it would help avoid the situation where networks that get added in the future, other than 172.16.0.0/16 wouldn’t get advertised.

Now if I do show ip route on R1 again:

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, 9 subnets, 3 masks
D       172.16.34.0/24 [90/41536000] via 172.16.12.2, 00:00:05, Serial0/0
D       172.16.3.32/27 [90/41152000] via 172.16.12.2, 00:00:05, Serial0/0
D       172.16.23.0/24 [90/41024000] via 172.16.12.2, 00:00:05, Serial0/0
C       172.16.12.0/24 is directly connected, Serial0/0
D       172.16.4.0/25 [90/41664000] via 172.16.12.2, 00:00:05, Serial0/0
C       172.16.1.0/24 is directly connected, Loopback0
D       172.16.2.0/24 [90/20640000] via 172.16.12.2, 00:00:05, Serial0/0
D       172.16.3.0/27 [90/41152000] via 172.16.12.2, 00:00:05, Serial0/0
D       172.16.3.96/27 [90/41152000] via 172.16.12.2, 00:00:05, Serial0/0

The 172.16.3.64/27 route is no longer present.

So the final config on R2 is:

ip prefix-list permitOut seq 5 deny 172.16.3.64/27
ip prefix-list permitOut seq 10 permit 0.0.0.0/0 le 32

router eigrp 1
network 172.16.0.0
distribute-list prefix permitOut out
no auto-summary
neighbor 172.16.12.1 Serial0/0

http://www.internetworkexpert.com/resources/doccd.htm

Anyway, I thought this ended yesterday, but they extended it. The workbooks+class cost me $750. I now have most of the materials I need to take me through to the end of my CCIE studies. I’m still missing a couple of books on the reading list though.

Today I was trying to configure a 4 router lab, with 1 router acting as a frame relay switch. I set it up in one instance, but a serial interface on my Frame relay switch kept flapping. I then tried splitting the switch off into its own instance. Same problem. Then I split the devices evenly between two instances and restarted my computer… Same problem. I was starting to get a little pissed, as i’d wasted 30 minutse on this. You should know i was trying to do this with 2611XM routers and a 12.4T AdvSec ios. Well, it seems that was the problem. By changing the FR switch to use a 7200 router and a 12.4T AdvEnt Ios, it stopped flapping and all of a sudden my eigrp routes showed up like they were supposed to. FWIW I’m still using the 2611XM for R1->3. I’m just using the 7200 for the frame relay switch. An hour later I can finally get on with the lab….

There’s also an application called MindBerry which can import mind maps from Freemind, so that you can have your mind map with you wherever you are, making it an invaluable study tool. I hate wasting time, and this will allow me to make better use of the time I’m on the subway commuting to work. I don’t have this program yet, but when I upgrade to a new blackberry in the near future, I will definitely be getting it.

copyrighted by Cisco Systems, Inc. All rights are reserved.

Now, maybe I am a little slow, but it took me a some time to wrap my head around this one. First of all, you need to look at both the Figure 7-7 and Table 7-2 together (the figure in this post is mislabeled because it is from the First Edition). When you first look at the diagram, all you see is a series of circles and ellipses with arrows pointing around and between them. You see references to input events. You see circles with text in them. It’s all very confusing. To make sense of it, you have to understand what the diagram is trying to represent.

When a route goes in to an active state, it can either get stuck in active, or eventually return back to a passive state. The diagram is illustrating the various input events that can cause a route to go from active to passive, vice versa, or to stay in its current state, and it shows how DUAL keeps track of the state with various flags.

Looking at the diagram, there is only 1 passive state. Passive is good. When the route is passive, r=0 and O=1. Everything is working. However, there are 4 different types of active states where: (r=1, O=0), (r=1, O=1),(r=1,O=2),(r=1,O=3).

To understand, look at the bottom left circle and the ellipse below it. What the ellipse with the text IE4 and the arrow around it is really saying is that “if we receive Input Event 4 (any input event other than last reply or a query from the successor), we are going to stay in an active state where r=1 and O=0. As long as we keep receiving IE4, we will stay in this state. However, if we receive IE11 (Last reply received; FC met with current FD), then we will return to a passive state. If we receive IE9, we will become active( r=1,O=1). If we receive IE10, we will become active (r=1,O=2).

It’s that simple. If you have any corrections, don’t be shy. I’m just learning this stuff too. FWIW this level of detail is probably overkill for the BSCI exam, however I’ve decided to study in such a way that I’m almost ready for the CCIE written by the time I finish my CCNP. That means i’ll be spending as much time in Doyle’s book as the BSCI Cert Guide. The extra detail and complexity will hopefully give me ideas for useful posts, so you don’t have to spend too much time reading about what kind of cheese I like.

Anyway, I started studying for the BSCI today. I read chapters 1-4 in the official BSCI cert guide. For the CCNA, the Cert Guide by Wendell Odom covered the material very well, with a good description of troublehshooting, the various show and debug commands, etc. I find the equivalent for the BSCI to be lacking, however. Fortunately, I already have Routing TCP/IP by Doyle and Carroll, so my routine will be to read that, and then read the Cert Guide. Doyle is much more comprehensive, but then it should be, because it is a CCIE level book. I’ve just started reading the section on EIGRP, but I flipped through it and it seems to talk about things that were completely avoided by the BSCI Cert Guide. The Cert Guide may or may not be good enough for the exams, but I figure there’s no harm in some extra knowledge right now, as my goal is to do the CCIE once i’m done with CCNP. I would love ot do the CCIE written by the end of the year. I will probably do CCDP as well, since it is only a single extra exam.

Anyway, I am trying to pick up the pace again. Hopefully work won’t interfere too much.

I have to say that that night I was kind of freaking out… I just didn’t have time to review. I had originally rescheduled from Tuesday to get an extra couple days of review, but it didn’t work out like that. I made the decision yesterday morning not to reschedule again. I got home late and finished reviewing the book at around 1am , and then did a practice test. It was the first that i was doing for ICND2, and I got 2 frame relay sims.

I have to say that Boson sims are really lame. I can’t imagine that people actually use these things for training purposes (real equipment or dynamips, or a combination of the two is the only way to go imo). Anyway, I read the question and…. nothing. It just didn’t make any sense to me. It wanted me to make a full-mesh frame relay network work so that all the hosts could ping each other, and it specifically said not to change any ip addresses. The scope of the question limited what you were supposed to change in the configuration, but the configuration was practically blank. None of the IPs were set on any router interfaces. There was no routing protocol configured on any of them, etc. Anyway, after scratching my head I clicked show answer and based on that it seemed like the configurations of the routers for the sim were totally wrong, because it listed only a couple ospf commands to get everything working, when in fact you would have had to configure everything from scratch. I clicked next to continue with the exam and got another frame relay question with exactly the same problem. Boson needs to fix these questions. They somehow expect you to make things work without IP addresses…

I finished the test and didn’t do so hot, probably because i was tired, so I ended up taking the day off of work so that I could review before the real test and go in with a clear mind. Some people would say that if you’re still reviewing the day of the test, you’re probably screwed. I feel I’ve learned things well, but with cisco exams there are always little bits of trivia that can come and bite you in the ass. I ran through the boson practice exam again right before the real test. I skipped anything with subnetting though. I don’t have problems doing the math and they tend to tire me out if i’m trying to do them fast.

Anyway, I’m on to BSCI now. I like the routing stuff so I think that should be a fun course. Maybe the blog will get a bit more intersting then. I just felt like everything was so basic at the CCNA level that it wasn’t worth writing much about. For anyone that’s wondering, there’s nothing really surprising on the CCNA exams. Read the books, make up your own labs with either real gear or dynamips, practice subnetting math if you’re bad at it, and don’t worry about it too much. The exams really aren’t very hard.