Think of the last two weeks before a big football match. You play one practice match against a tougher team to see where you stand, then train your weak spots, play a couple more friendlies, and in the last two days you do nothing new: check your boots, pack your bag, sleep. On match day you arrive early, warm up, and play your plan.
For the CKA: week 29 starts with killer.sh session one (timed, then redone untimed while it's open), then your weakest domain's drills and mocks. Week 30 is the redo queue, mock 3 twice, and the docs pages opened once each. Two days before: killer.sh session two. The day before: nothing new, the PSI system check, sleep. On the day: early check-in, three passes, exit every node session.
Weeks 29-30, and the day itself
The problem. The last two weeks decide whether the drilling turns into a pass: a plan for practice, the technical check of your room and computer, and a calm routine on the day.
What you need to know already: everything in this chapter; the wrong-answer log (19.42).
The two weeks, as the plan lays them out
when
what
Week 29, day 1
killer.sh session one, timed, as the real thing. Do not look anything up that the exam would not allow.
Week 29, days 1-2
Same session, untimed (it stays open 36 hours): redo every question until clean. Read killer.sh's solutions even for the ones you got right - they are often faster. Log everything.
Week 29, rest
The drills of your weakest domain first; the random block every other day; mock 1 and mock 2.
Week 30
The redo queue from the log; mock 3 twice; the docs pages opened once each.
Exam day - 2
killer.sh session two, timed. Log, redo the misses.
Exam day - 1
Nothing new. System check. Sleep.
killer.sh is deliberately harder than the exam. A killer.sh score that would fail the exam is common among people who then pass comfortably. Look at which tasks you missed and why, not the total.
Booking and the system check
Book at a time you are sharp, on a day where a failed platform (it happens) or a retake would not wreck your week.
Run the PSI system check on the same machine and network you will use, a few days before. The PSI Secure Browser needs to be installed; corporate laptops with endpoint security often cannot run it - use a personal machine.
The room: clear desk, no second screen connected (or turned away and unplugged), no notes, closed door. A proctor will ask you to show the room with the webcam.
A government photo ID matching the name on your booking.
Exam day
Log in early; the check-in takes time.
First minute: open the terminal and the docs, check copy/paste works (Ctrl+Shift+V in the terminal).
Read the first task, run its ssh/context line - and start. Leave the speed kit for the task hosts (export do=... on each).
Three passes: quick ones, flagged ones, verification.
Exit every node session and root shell before moving to the next task.
With ten minutes left: stop starting new work; verify what you have.
Afterwards
Results arrive by email, usually within 24 hours, with the score. If you pass: the certificate is valid for two years. If you do not: the retake is included (within 12 months of purchase); the score report shows domain percentages - put the weakest one at the top of the log and book the retake one or two weeks out, not the next day.
Then
The CKA proves you can operate a cluster quickly. The interview questions in the debrief are what proves you understand it - practise saying them out loud, they are the other half of the job hunt.
Why it helps
A good plan for the last two weeks turns preparation into a pass. Most failures in the final stretch are avoidable: cramming new topics instead of drilling misses, reading a hard killer.sh score as a verdict instead of calibration, or discovering on the day that the corporate laptop can't run the PSI Secure Browser.
The logistics matter as much as the knowledge: a personal machine that passed the system check, a clear room, an ID matching the booking, a time slot where a platform failure wouldn't ruin your week. And the plan for "afterwards" (the retake is included, use the domain percentages) takes the pressure off. It's the same thinking as preparing for a production change window: rehearse, check the environment in advance, and know the rollback.
FAQ
My killer.sh score would fail the exam. Should I postpone?
Not necessarily. killer.sh is deliberately harder than the real exam, and many people who score below the pass mark there then pass comfortably. Look at which tasks you missed and why: knowledge gaps in a heavy domain are a reason to wait; reading mistakes, slow but correct tasks, or killer.sh's extra-hard questions usually aren't.
Can I take the exam on my work laptop?
Often not. The PSI Secure Browser needs to take over the machine, and corporate laptops with endpoint security software or restricted admin rights frequently can't run it. Use a personal machine, run the system check on it and on the same network a few days before, and fix any problem then rather than at check-in.
What should I do the day before?
Nothing new. Run the system check, prepare the room (clear desk, no second screen connected, no notes, closed door), have your government ID ready with the same name as the booking, and sleep. Last-minute cramming adds anxiety, not points; the drills have already done their work.
What's the first thing to do when the exam starts?
Open the terminal and the docs browser, and check copy and paste works in the terminal (Ctrl+Shift+V). Then read the first task, run its ssh or context line, and start. Set up export do=... on each task host as you go, since variables don't follow you through ssh.
What happens if I fail?
The exam includes one free retake within 12 months of purchase. The score report shows your percentage per domain: put the weakest at the top of your log, drill it, and book the retake one or two weeks out, not the next day. If you pass, the certification is valid for two years.
In an interview Junior
How do you prepare for a high-stakes change, like a production upgrade?
The same way as for the exam - rehearse, check the environment, then follow a calm routine on the day:
Rehearse under real conditions: a dry run on a copy (a staging cluster, a mock exam), timed; redo every step that went wrong until it is clean, and write down why it went wrong.
Check the environment days before: access, tooling, versions, backups taken and a restore tested - the equivalent of the exam's system check.
Pick the moment: a time you are sharp, where a failure or a retry does not wreck the week; nothing new the day before.
On the day: a written order of steps, verify after each one, never skip the last steps (uncordon, restart, persist) - on exam day that includes the setup on each host (export do=...) and exiting every node session, and stop starting new work near the end of the window - verify what is done.
Afterwards: record what happened, and the weakest part goes to the top of the list for next time.
Also asked: How do you stay calm and effective under time pressure? · Why did you decide to get the CKA, and was it worth it? · What would you do if the platform failed in the middle of a timed task?