I failed question 2 in the 3rd mock exam in the CKA course due to a trailing space. I used the the following command:
k get nodes -o jsonpath='{range .items[*].status.addresses[?(@.type=="InternalIP")]}{.address}{" "}{end}'
While in the solution, they had:
k get nodes -o jsonpath='{.items[*].status.addresses[?(@.type=="InternalIP")].address}'
Can we expect the main exam to also be this pedantic?
If you think the grading software in our Mock Exams are a black box, well, wait until you take the real exam 
In truth, I honestly don’t know, and I’ve taken a mess of them. There is a procedure to appeal, but I believe it costs money, and I don’t know anybody who’s (1) tried and (2) succeeded. Just do your best to follow a question’s instructions, and hope 
I hear you. I guess the 66% to pass gives leeway for such.
I’d also imagine that in the main exam, they also look at the command history to see if someone simply did:
echo 'ip1 ip2' > file
If I pass with a 67%, I’ll be the first one to post here (anonymously of course). Can’t have a potential employer seeing my low passing score and take that to mean I’m borderline incompetent - tsk, tsk…
PERISH THE THOUGHT!!! We maintain our collective competence around here.
Here’s another one for you, @rob_kodekloud
I failed this one because I was completely unaware that I could claim a volume directly using volumeName - 
.
Anyway, after the exam was graded, I went looking to understand why my VOLUME field wasn’t being populated, found the answer, deleted and recreated the PVC but it still failed to give me a pass - I’m obsessed with seeing 100% after testing again. However, not in this this case. It utterly failed to humor me.
Anyway, can you spot the issue - if any? I’ve been looking at it for a while and can’t see it!
TBH, I don’t see a problem with your solution either: the node affinity on the PV is the same, the parameters of the PV and the PVC look both the same and compatible with binding. And of course, the PV and the PVC bound, which you show. I’ll try the problem and see if I do any better
I guess this is an Ultimate Mock Exam, which is hard for me to examine (due to the randomization). The only thing I can think of is that given the WaitForConsumer property, that perhaps binding would be delayed; except IIRC it does not if you specify volumeName. If you’re really curious, we have an engineer that’s been tracking issues like this in the Ultimate Mocks. But I think you understand the concepts here, and don’t really need to worry about this as an exam topic as a result.
Yep - it’s in the Ultimate Mock Exam. Glad you confirmed it wasn’t it wasn’t a case of dyslexia on my part. And yes, no need to get an engineer involved for this. But maybe a heads up for them to verify the grading software isn’t in a drunken stupor - 
.
I’m thinking there might be something odd going on in the clusters. I just started another one and it took close to 1 minute to get things ready for this question and then I got this:
Might be time to get an engineer involved to take a look.
It might be the universe telling me it’s time to take a breather…
I’m wondering, is it okay to post screenshots with questions from the mock series in the community forum?
I had issue with sysdig utility during CKS exam, it just didn’t work. I had 64% from 67 required. When I complain about it they said - it’s your problem, our environment is ok. If you want to get deep analysis of your env - it will cost 150$
If you 100% sure - you can try. If they found that issue on their side - you will get 150$ back
@ak.cka on that etcd restore question, looks like you’re using the right context and have ssh’ed into the correct node. that etcd db file is just not there when it should be
Indeed @reteer
In fact, there’s another post reiterating the same so it appears to be a bug somewhere. Especially so as it takes quite a long time to set things up for this question. My money is on SSH and changing IP’s.
