RE: keeping configuration lines minimal

Hey Jack Router!

1. I agree here completely; for IPv4 and IPv6 Frame Mappings I recommend the broadcast keyword once per DLCI as a default configuration. You can always reverse this should you come across a subtask that requires you to. Also note that you could duplicate the keyword on multiple Frame Maps in IPv6 with no ill effect, but for consistency I do once per Frame Map there as well.
2. If I am given the freedom on the L2-L3 name resolution approach; I ALWAYS go with static mappings. ALWAYS.
3. I personally would not bother with this one as a defaultb&it is so obvious to me when I need to worry about this.
4. I did not bother with this one either. I learned to become very fast and efficient with show ipv6 int brief, the Windows clipboard, and notepad during IPv6 configurations. All of this ended up being faster than manually coding Link Local addresses, and I feel a bit more confident in not having changed an address when I was not explicitly told I could.

In summary, while I would not personally do your items 3 and 4, I will say that I think you will be just fine with the grading script, and I do not believe that if the human proctor gets involved, they would think that you were trying to bbrute forceb the lab. They would think you were a smarty-pants.

I hope these posts have helped you J.R., and I love your bscreen nameb.

From: Jack Router [mailto:pan.router_at_gmail.com]
Sent: Thursday, October 21, 2010 5:16 PM
To: Anthony Sequeira; 'Sergey Matashuk'; 'Cisco certification'
Subject: RE: keeping configuration lines minimal

Thanks for the answer. I have however some specific commands that Ibd like to clarify. Assume that nothing explicitly prevents or obliges the use of those commands:

1. bFrame map ip ... broadcastb. I know what for broadcast is for and that it can be not required. But only to save time Ibd like to put this every time.

2. bNo frame inverse-arpb: in some situations it will do anything but can save time later.

3. bip ospf mtu ignoreb : it wonbt hurt and can save time later.

4. bipv6 address FE00::X link-localb, where X is router# : will save time when testing connectivity.

I understand that use of some commands may leave a perception that candidate does not understand the technology, but certainly it does not prove that the candidate is ignorant either. It is only about saving time. I believe that labs are graded with automated scripts. Do they catch those buselessb commands ?

From: nobody_at_groupstudy.com [mailto:nobody_at_groupstudy.com] On Behalf Of Anthony Sequeira
Sent: 20-Oct-10 18:54
To: Sergey Matashuk; Cisco certification
Subject: RE: keeping configuration lines minimal

CCIE instructors see the question time and time again - are we penalized for bover-configurationb in the CCIE lab exam? The answer - bnot typicallyb. Let us walk through some examples to see exactly what we are talking about here.

First of all, I encourage students to ask two questions when they are about to bover-configureb something. Question 1 - can this additional configuration I am about to make actually gain me points (might Cisco be grading for it)? Question 2 - can this additional configuration I am about to make actually hurt me (cause point loss)? If the answers are a resounding YES and NO, then it is definitely a configuration you should consider making.

A simple example would be setting a Layer 2 switch port for a VLAN with:

switchport access vlan 100

Versus:

switchport mode access
switchport access vlan 100

Might Cisco be grading for the specific configuration of DTP mode OFF on the port, perhaps. So the answer to the first question is YES. Notice, on the other hand, this configuration should not cost us points in any way, so the answer to the second question is NO. We see that these questions lead us to the conclusion...if it can only help us and not hurt us - GO FOR IT!

While many times we are not penalized for over-configuration, remember that we are always looking for the simple, time-saving, straightforward solution to the task at hand. I have seen ridiculous amounts of silly over-configuration from students that do not understand this principle. One example that comes to mind is the student that is asked to iBGP peer between R1, R2, and R3 using AS 100. The student then takes it upon himself to configure peer groups, loopback peerings, and router-IDs. All of this is for bgood measureb and absolutely none of it was required and gained the student any points! In fact, when asking the second question about the over-configuration causing point-loss, the answer here might be...byes, it can cause point loss because I am wasting so much time!b

Let us also remember that the key to solving the CCIE lab exam comes down to reading very carefully and following explicit instructions versus implicit instructions that exist in the task. Often times we discover additional configuration steps that we should take due to implied requirements.

I discuss this issue in greater detail in the following blog post:

http://blog.ine.com/2008/11/12/the-lab-made-me-do-it-%E2%80%93-implicit-versus-explicit/b(

________________________________________
From: nobody_at_groupstudy.com [nobody_at_groupstudy.com] On Behalf Of Sergey Matashuk [matashuk_at_gmail.com]
Sent: Wednesday, October 20, 2010 5:26 PM
To: Cisco certification
Subject: keeping configuration lines minimal

What guys are you thinking about keeping configuration lines minimal
on a real CCIE lab exam?

I`ve did INE`s mock lab exam some time ago, and i`ve noticed I`ve loos
the points for one of the tasks due to one more route-map entry. The
route-map were used for redistribution tagging, and my solution was to
add explicit permit clause. All the task requirement were solved, just
the route-map MAY had one entry less.

So, I`m wandering will such approach ends loosing points in real lab
exam? Following keeping conf lines minimum principle we can end up for
example "summarizing" network statements at routing protocol config
mode (to permit two or more interfaces by noe network statement), but
I never heard about loosing points for explicitly permiting every
interface running routing protocol with network statement. And I
personaly prefering to add interface to routing protocol with 0.0.0.0
wildcard, just to have more granular control. We can have interfaces
added (eg tunnels) on later tasks, and I don`t want end up
troubleshhoting routing at the end of exam.

Where is a reasonable border betweeing keeping configuration as small
as possible and having some "freedom" to configure things? Should we
been paranoid on configuration lines? Let`s discuss all of your
thoughts.

Blogs and organic groups at http://www.ccie.net
Received on Thu Oct 21 2010 - 18:52:21 ART