Thoughts on pre-configuration analysis approach

If anyone is willing to share, I'd like to hear other people's thoughts on the best way, or good ways, to approach the pre-configuration analysis when doing a lab.

 

If anyone isn't familiar with what I'm talking about, there is lots of INE wisdom against just sitting down and hacking away at the gear without a plan.  Also, this seems to generally be frowned upon universally by anyone who knows about the CCIE lab.

 

In the past, I've been drawing tracking tables (tasks and points) and making my own L2 diagram and L3 sketch diagram.  Then I've gone through the lab and kind of "skim" it.  After all this analysis and prep, which takes about 20-30 minutes, I start configuring and track my points and minutes per point.

 

I've been trying something a little different lately and I'm curious if anyone has any experience with my old way (above) or the new way (below).

 

The new way of analysis and prep I've been doing is the same except for the "skim."  In the new way, I have notepad side by side the lab scenarios and I type a really rough translation from the scenario into notepad.  On the good side, I have to look at the scenario less and can just focus on converting my sketch notes into actual commands.  On the bad side, the prep and analysis is taking about 45 to 50 minutes instead of 20-30.  With the new way, I also feel like I have a better grasp of the overall lab as a whole before I start configuring.

 

For example, I asked myself "why not do 1.3 of the Vol2 Lab2 right at the begining so the switches can reboot?  I can work on 1.1 while they are rebooting."

 

Does anyone have an opinion about the two approaches?  50 minutes seems like a lot of config time to give up vs 20-30 minutes if people have found it doesn't work well.

 

Here is an example of a few notes I made from one of my Vol2 Lab2 run throughs:

 

1.1 --------------------------

19 - 21 on sw1
13 - 15 on sw4
one l2 link
sw4 respond to sw1
sw1 is decision maker 802.3ad
isl for logical l2 link

2.1 --------------------------

ospf 1
r1 r2 r3 r4 no DR election
no neighbor command
r1 and sw1 clear text pw CISCO

Another way to ask my question is, would it be more efficient just to expand those notes into commands right as I'm reading them first pass.  I guess this would be a 3rd option, vs the old way and the new way.

 

Any thoughts or input would be helpful and appreciated.

 

Comments

  • Hi,

     

    I usually do your first way, draw my own l2,l3 diagram with redistribution points and bgp multicast diagram. I also do a tab with task, points and checkpoint to see what task I've done what task I should come back and take a look again.

    Your second approach looks good the only problem I'd say , for me, you are just rewriting the task on the easy way to understand and do, its good but you just have to be carefull with details.

    The best way for me is doing the L2,L3 Diagram according to my evolution on the tasks.....for example:

    I start my lab, recheck all configuration to see if there is any ip address mismatch, pvc wrong, etc, and then I do a L2 diagram.....

    Start doing the tasks , when I'm doing the L3 tasks I do the L3 Diagram according to the task, it helps me to identify the loops prevention points, where could be a RPF Failure in the future............

     

    I'm not telling you that your second approach is not good, it is very good, but everyone have your own way and the best way to understand and do the things.....Your approach is very good as well, but for me it seems like rewriting the question.....

     

     

     

  • JoeMJoeM ✭✭✭

    I have not taken the lab yet, but here is my thinking, after having a mentor (multiple-CCIE) look over my shoulder during labbing.

    • For the real-lab, this new approach may be too time-consuming....and that is if it is written out correctly the first time.
    • For practice labbing, it is good method that can be optimized with practice.  This is just a variation of writing everything in notepad.

     

    For most of my workbook lab practice, I did something similar as the 2nd approach. Everything was in notepad, with an outline of the task requirements and  the final config. But this was only for practice.   In reality, it was a way for me to intrepret/translate what the task is asking us to do.  At first, everything looked to me like it was written in Greek. I did not understand what was being asked.

    But as I have come further along and closer to the lab date, I am going back to a more optimal approach, as what was recommended to me. 

          I was told, "Don't do everything in notepad. There is not enough time."

     

    There are many tasks that become super clear.  1,2,3 DONE    When this is the case, why not just put it directly into the console and save the config?  TASK checked off with a smiley.  ;-)

     

    MAPS

    My drawings are close to being optimal. Solid L3/L2 maps (with colored IGP) is less than 15 minutes (17min on a foggy brain day).

    This gives me about 3-6 minutes to do the redistribution and BGP maps later. I also have these drawings optimized for time/readibility, but I do these later, when the tasks actually come up. 

    My thinking here, is that when the config clock starts, I fully expect to be in a small panic/race mode to get the L2/L3 maps done and clean/usable.  This allows me to start building my confidence (relax a little) by beginning to complete tasks and marking them off on my check-list.   Settling into the actual lab configs -- and remembering to relax and breathe, and not be panic'd about the clock.

    TASK REVIEW and POINTS CHECKLISTS

    This part I still need to refine for myself.  But everything I have read, is that it should be simple.  Skim/review and draw the columns.

     

    I am at a point, where these lab topics on strategy are important.  Any corrections or suggestions are appreciated.

    Thanks SPNetwork for your opening post.

     

     

     

  • Thanks for the thoughts Renato and JoeM.

     

    You guys are 100% correct that re-writing the task has a major flaws:

     

    You either have to blindly trust you have rewritten it correctly or you have to double-check the actual task.  If you didn't re-write it correctly, you could lose a lot of time, or worse - points.  If you have to glance back at the task, what was the point of re-writing it.  As I've been doing it more, I've found it can actually waste a lot of time.

     

    However, what I'm trying to find is a way of doing the SKIM of the "old" method described while getting a good feeling for the overall lab.  That's the big thing I've noticed from the doing the short version rewrite is the big picture of the lab in question is very clear.

    Regarding notepad, I have to disagree somewhat.  I think if you have  a task which requires one line added on one router, then just type it.  But if you have the same or similar config to be added to 5 routers, you would be putting yourself at risk during the exam and wasting time, IMO.  An example would be enabling PIM NBMA on the FR Hub router (just type it in) vs converting all interfaces on the frame-relay network to point-to-point and multpoint (better get notepad out).

     

    Maybe you meant the same thing.  Also, I did not pass my attempt #1, so I'm only qualified to tell you for sure how to not pass. :)  But I didn't feel I would ever have been able to get all the work done that day without notepad and beloved copy paste.

     

    The other benefit to me of practicing notepad is I have gotten a LOT better with command accuracy.  Before I started working almost 100% in notepad, I was abusing the "?" key. :)  And I found it did not help me get the commands burned in my brain.  Notepad has.  I've been doing my best guess and passing it into the routers.  When I get the "^" I try again until I get it right.  It's really helped me a lot.

     

    Again, thanks for the replies and thoughts.  I agree with you I'm at a point where strategy is very imporant too.  It was clear to me it's a HUGE part of passing the exam.

     

     

  • JoeMJoeM ✭✭✭

    spnetwork,

    Well, you have the advantage of already experiencing the lab. That should help you a lot on the next attempt, don't you think?

    I agree completely with you about the efficient use of notepad.

    In practice, I think it is one of the best tools for memory. Everything into notepad first.  But, I have been using notepad++ (tabs).  Now that I am at my 60 day mark (only one attempt at v4), I am going to windows notepad (arrrgh).

    Fortunately, my use of notepad is narrowing down to just a few things:

    • identical configs/commands that are needed in many/all devices.
    • tricky/lengthy configs that I may need to review/adjust
    • math/binary
    • (practice) memory/doubts on configs that I need review later

     

    For simple configs in multiple devices, I have started using a different method:

    1. type the config into a device
    2.   show history
    3. copy and paste (using the ALT key for text selection)

     

    SKIMMING/REVIEWING THE COMPLETE LAB?

    I am definitely with you, in regards to a method to SKIM/REVIEW everything.  I feel overwhelmed by the information, and I am not sure I am going to be able to optimize this.  I think my points chart will be okay, but I am not fast enough to review the details in each topic.

    I remember a video with Brian Dennis, where he was joking about pointing our finger at each word on the screeen -- being sure to understand EXACTLY what the task is looking for.  I laughed, because I feel like that some times (with the mouse cursor).  I feel completely dyslexic when I have to review so much -- along with the time crunch.  I keep thinking, "The clock is ticking...get going  Time to get config'ing"

    Any advice or suggestion on this part of the lab is much appreciated.  This may be my strongest weakness in the lab, aside from DocCD as a backup strategy for all of the Services. 

    What do we preview in such a race for time?  Can we narrow it down to topics that deserve caution (for dependencies)?

  • Everyone has different methods but here is what worked for me.

     

    No extra diagrams. The ones provided were good enough. If there was some advanced L2 task or similar it might be worth drawing. If it really helps you and you are fast at it then go ahead.

    Skim through the entire lab. Look for things like if you need to bump MTU on switches or enable IPv6 on switches. Those are tasks that require a reload. Make those changes directly while you go on with other stuff. Look for any tasks where you think you might have to use the DOCCD. Maybe a BGP task looks tricky. Open up BGP in the DOCCD and let it load, that could take a while. Go on with other things.

    Make a scoring tracker. I just made an easy one with task number, points and a check mark if I had done the task. An extra check mark if I had verified it twice.

    I wrote all the configuration in Notepad. Do whatever works best for you. If config was very similar I would copy it and just change IP addresses etc.

    I copied text from the workbook such as the VLAN IDs and names and created those. Since they are case sensitive this makes sure you don't get anything wrong. Verify connectivity straight away, it's better to have this done than get nervous when OSPF is not working but you got the VLAN wrong or something.

    Do all the tasks, hopefully by lunch you have at least completed L2 and some of the L3. On my passing attempt I was completely done with L3 before lunch.

    During lunch people won't be too cheery. Don't let anyone elses mood influence yours. The lunch break gives you some well needed rest and it's also a good time to think about a task you might be stuck on. Being out the lab room could give you a new perspective so use this time. You are not paying 1500$ to get a free lunch anyway.

    Keep doing tasks and write them down on the scoring tracker. Always verify your tasks as you go on, don't leave this to the end.

    Hopefully you have done all the tasks and then you may start another round of verification. Depending on how much time you have left you have to decide if you want to verify tasks or try some task you get stuck on. This decision will be based on the number of points that you think that you have and how much those tasks are worth. In some situations verification could give you more points than solving tasks. You will probably have done some easy mistake like getting a name wrong or missing an entry in an ACL or something stupid like that. Stupid or not that's still points out the window.

     

    Submit the exam, give the proctor a smile and wait for the e-mail with your number... :)

    If not, come back, optimize strategy and get it next time.

  • JoeMJoeM ✭✭✭

    Thanks Daniel. Really appreciate your input.


    By the way, congratulations on your VIP status at Cisco.   I read everything you post.   Your blog is great.

  • JoeM and Daniel,

     

    Thanks for the replies and great notes.  It's definitely given me a lot to think about.  I like the idea of focusing on dependencies and major trouble areas during the skim.  Maybe even make a little note somewhere if you run across a topic you really are strong in so you know you can definitely get points there.

     

    JoeM,

     

    As for having the experience of a failed exam, you are certainly right that it helps to have been there.  I don't think anything can prepare you for that experience.  The main thing I got from it, and the reason I am so curious about this post subject, is you cannot waste any time.  Perhaps if you are so strong with everything you can finish Config in four hours and have time to spare and triple check, then strategy isn't that important.  I feel if I'm not strong enough to bang out an INE Vol2 Lab in four hours, then strategy and points will be as important as my expertise.

     

    Thanks for the input and thoughts!

     

Sign In or Register to comment.