Poly-Lab experience; comments/questions

Hello All,

I want to provide some feedback on the Polymorphic Lab Assessment product, as well as offer some suggestions and hopefully get some answers to a few questions I have.  I thought this would be the best spot for this thread.

First I want to commend INE as the product is pretty good.  I think its an excellent resource and compliments the rest of the CCIE 2.0 product suite really well.  I have attempted to complete 6 Poly Labs so far.  Twice I didnt get them graded which was my own fault and once it failed to load the initial configurations.  The Graded Labs and INE support are pretty bloody great so far.  Response times are good and they were quick to refund the tokens if the problem wasnt mine.

I want to address a couple of items below.  Its likely some of the points specific to questions answers being marked wrong are in fact me answering incorrectly, hopefully someone can point them out for me.

(a) Approximately 6 weeks or so ago I remember there being a Session Estimator on the homepage where you generate a lab.  This doesnt seem to be there anymore from what I can tell.  Did it disappear for everyone, just me, or am I blind?

(b) When a lab fails to load or you fail to have it graded it remains in the section 'Labs In Progress'.  Their state is either Setup Not Completed or Expired.  From what I can tell you cant reset them and load/take them again.  Is there a way to do this and if not is there a way to remove them?

(c) The physical topology is static like the workbook labs which is fair enough.  The logical topology is also relatively static; same IP addressing, location of IGP's etc.  This is also fair enough I guess seeing as though the questions must be generated with certain constraints around what the network could look like.  I can see myself after doing another 2 or 3 poly labs getting way too familiar with it.  I am already getting the same tasks and am configuring them blindly based on previous experience.  Dont get me wrong it is still a great learning tool, however, what happens when I am expert in all of the sections?  I notice that the topology diagram has 'Assessor Lab Exam 1' in the bottom left corner.  Is it fair to assume that there will be additional releases?  If so do you have any estimate on this?  Yell out if you want beta testers also ;-)

(d) Would it be worthwhile including a question/task ID for each of the tasks so that they can be referenced in threads?  Are tasks pre defined or can a tasks include various dynamic points?  Just a thought not a big deal..

(e) It doesnt look like the grading scripts check full reachability, do they?  It would be awesome to have some criteria it checks full reachability in IGP and BGP from a couple of devices.

(f) I had a question requiring CBAC.  I successfully completed the tasks however was marked incorrectly as I didnt use an explicit 'deny ip any any log' in the INBOUND ACL.  Is it okay to rely on the implicit deny all or do I need to re-read the documentation.  If not could the grading be updated?

(g) I had the following task:

3.4 BGP Routing Stability   (3 pts)
    * A flapping link inside AS 54 causes frequent updates and removals of BGP prefixes.
    * Configure R6 and SW2 in AS 100 to suppress oscillating routes and set the exponential decay parameter to 5 minutes.

I immediately thought BGP dampening.  I can see the exponential decay parameter in the 'sh ip bgp dampening parameters' output although I hadnt heard of it before.  This was one of the labs I unfortunately didnt get graded.  Can someone link me some good information on this as I dont fully understand how the half-life, re-use timers etc affect it.  I couldnt find adequate information in the documentation.

(h) I had the following task:

1.3 VLAN Creation   (3 pts)
    * Create VLANs in all switches using the diagram provided as your reference. Do not change VLAN names from their default values.
    * SW1 should advertise new VLANs and SW3 should learn them dynamically.
    * Ensure that SW2 and SW3 do not participate in dynamic VLAN distribution.
    * Use a key value of CISCO to secure VLAN information exchange and use the domain name CCIE.

I was slightly confused as to what the task wanted and ended up making SW1/SW2/SW3/SW4 serv/tran/cli/tran.  I got the marks however thought perhaps * 3 above should be SW2 and SW4?  Maybe not..

(i) I had the following task:

2.9 RIP Filtering   (2 pts)
    * Enable RIPv2 to exchange routing updates between SW2 and BB3.
    * SW2 should not accept routes with an odd second octet from BB3.
    * Do not use the offset-list or the distribute-list commands to accomplish this.

I used:

distance 255 204.12.9.254 0.0.0.0 29
access-list 29 permit 0.1.0.0 255.254.255.255

However the grading marked me incorrect.  The assertion it is using is:

show running-config | include distribute-list.*Vlan8

(j) I had the following task:

5.3 Multicast Testing   (2 pts)
    * Ensure that R2 only accepts IGMP joins on its Ethernet interface from groups in the ranges 239.0.0.0/16, 239.2.0.0/16, 239.4.0.0/16, and 239.6.0.0/16.
    * Use access-list number 22 and one access-list entry to accomplish this task.
    * In order to facilitate a multicast test, join the Ethernet interface of R2 to group 239.0.0.2.
    * Ensure you can ping the above group from R5 across your multicast domain.

The grading wanted 239.0.0.0 0.4.0.0, wouldnt it want 239.6.0.0?  Once again I could just be over tired.

(k) I had the following task:

6.1 Congestion Avoidance   (3 pts)
    * Cconfigure R6 to randomly drop packets before congestion occurs on the Serial interface output queue.
    * Ensure that traffic marked with critical precedence will not be dropped unless there are 50 packets in the output queue.
    * If there are 70 critical packets in the output queue, R6 should randomly drop 4 out of every 16 of these packets.
    * In the case that there are more than 70 critical packets in the output queue, they should all be dropped.
    * Do not use any MQC commands to accomplish this task.

Should the mark probability denominator be 8 or 4?

I would of thought:

 random-detect
 random-detect precedence 5 50 70 4

(l) I had the following task:

7.2 Filtering   (3 pts)
    * Recently you have noticed a large number of fragmented packets coming from behind BB3.
    * This type of attack impacts the performance of your servers located on VLAN 17.
    * Configure SW2 to drop all fragmented packets toward VLAN 17 IP addresses as they enter SW2.
    * Use a named access-list NO_FRAGMENTS and the minimum amount of access-list entries to accomplish this task.

Should this be reachable-via rx or any and why?  I would of thought any.

(m) I had a question requiring CBAC.  I believe I only had it graded incorrectly as the parse in the grading script is not allowing for spaces.  See below:

  R3: Check the inspection global configuration settings.
This is a Match All type of assertion with the following parameters;
one-minute (sampling period) thresholds are [60:120]
max-incomplete sessions thresholds are [100:200]
max-incomplete tcp connections per host is 30. Block-time 1 minute.
tcp synwait-time is 10 sec

Rack9R3(tcl)#show ip inspect config
Session audit trail is disabled
Session alert is enabled
one-minute (sampling period) thresholds are [60 : 120]
connections max-incomplete sessions thresholds are [100 : 200]

Is it grading it incorrectly as their are spaces in the [x : y] output?


I hope this is a suitable arena for voicing these comments and questions.  It would be great to get some feedback so I can fix my mistakes and get 100%.  I would definitely recommend the Poly Labs as part of your overall preparation they are a great tool.

Thanks,
Tom.

Comments

  • Tom,

    On your item "G" for BGP dampening, the Decay Time is affected by the half-life and/or max-suppress values...here's an example:

    R3#conf t
    R3(config)#router bgp 3
    R3(config-router)#bgp damp 7 750 2000 15
    R3(config-router)#do sh ip bgp d p
     dampening 7 750 2000 15
      Half-life time      : 7  mins       Decay Time       : 300 secs
      Max suppress penalty:  3310         Max suppress time: 15 mins
      Suppress penalty    :  2000         Reuse penalty    : 750


    For item "J" on multicast, I computed the following one line ACL: access-list 22 permit 239.0.0.0 0.6.255.255


    HTH,
    Jeff

  • Hey Jeff,

     

    Thanks mate.  You are right I should of had 239.0.0.0 0.6.255.255 not 239.0.0.0 0.6.0.0, not sure what I actually had in the config.  The grading script still is incorrect with x.4.x.x though.  I would really like to hear from the developers or anyone at INE about having the grading script changed for items like these which we identify.  Otherwise I am going to constantly get 80% or thereabouts as I always have the possibility of getting marked incorrectly for 3 or 4 questions.

     

    Do you have any other reference material for the BGP dampening feature?  All I could find was:

    http://cisco.com/en/US/docs/ios/iproute/command/reference/irp_bgp1.html#wp1012660

    It doenst reference decay at all so I was just modifying timers and trying to get 300 seconds.  What made you use 15 for the max suppress, is that the default?

     

    Thanks :) !

  • I would encourage anyone who finds errors to email [email protected] to report any issues.  I have found that they are very responsive in reviewing the ticket and making any necessary changes.

    That being said, you probably will find cases where it marks you wrong, simply because you may have used a different method, which doesn't violate the requirement.  That is one of the flaws of an automatically grading system.  However, like you said, it is a good tool, and gives a good bit of flexability.  I wouldn't be too concerned that you are getting the same tasks over and over, unless you are only using the poly labs.  The last few weeks before my lab, I only used the poly labs.  While certain sections were the same, the overall test was different, and for me, was enough to drive it home.  One benefit of similar tests is it helps you with your speed and accuracy.

     

     

Sign In or Register to comment.