CKA Mock exams validation are bugged or unclear

Hi,

I’ve been actively studying with kodekloud for CKA certification.
I went for my daily mock exam and started by resetting TWICE because the first question was the infamous “restore a backup that is nowhere to be found”.

Third time was the charm and I was really exited when I clicked the “End exam” button half an hour early feeling really confident since it never happened before.

What was my disappointment when I saw my 64% success rate…
Then I proceed to check where I’m wrong just to find the following (cf attached screen capture at the end) :

  • The first pod for color just went MIA
  • The deployment with incorrect database info as been fixed(the hard way by modifying all the secret without touching the existing deployment because question says to not touch existing environment var, yet solutions does it) and then it’s wrong ? Also why does it suddenly restart the moment i complete the exam ?
  • I’m still trying to find where I missed a requirement for my sidecar (A.K.A the question is not specifying all the requirement ?). Also, the main container is not created correctly ? Really ?
  • Sleep 3600 missing …
  • And here I actually not so confident cause Network is my nightmare, but i’m pretty sure it is accessible.

So please correct me if the mistake is on my side.

Also, I’d like to ask if the real exam question are formulated the same way because I often wonder about the meaning details ( for instance, I got the wrong answer when asking for image name, because I wrote “busybox” without the tag which may or may not appear depending of the way you display the container yaml/json.

Finally, is the exam validating question in the same way(automated regexp, etc, without human factor) ? I’m thinking of the sidecar mainly where I feel the error comes from an unclear requirement.

Best regards,
Hopefully I’m not confidently wrong, in such case, i’m really sorry.

There are various internal tickets raised against the labs of this course. Those of us answering support questions have no idea on the lead time for lab fixes.

As for the real exam, yes it marked by scripts. Questions are usually clear about what is required. Be sure to read exam questions very carefully, as points scored are binary. Something is right or it is not.

You need to also do killer.sh before you seat for exam. This is the closest experience you will get to the real exam as it is the official exam simulator.

Hi,

Thanks for the answer.
I learned to deal with most of these issues with enough practice even though sometimes I still face some strange issues where I’m really not sure I’m the problem.
for instance:

  • The sleep 3600 is most often asked as “command” hence it break from my “command: - sleep” + "args: - “3600"”
  • a few days ago, I realized I couldn’t use “kubectl top node --cluster=a_diff_cluster_from_current” because it suddenly gave Authentication errors.

About killer.sh, I see that question are in fact split into task and each task is scored as completed or not.
I guess this mean for the real exam, you should not straight up skip question you can’t finish because you can still get partial points ?

Best regards,

Yes you can get partial scores for the parts of a question that can be checked separately, e.g. “Do X, Y and Z”, so you could score for X but if Z depends on Y and you get Y wrong, then Z would also be wrong.