Bukola:
@Craig Shea So it is marked for me as correct now after I added the scheduler name to the command
Craig Shea:
@Bukola Interesting, because I did add that parameter, otherwise, other “bad” things happen 
I’ll have to go back and try. 
Glad you got it to work!!
Craig Shea:
Hmm, I think question 3 is not being marked as correct for me is because I left --leader-elect=true and added --lock-object-namespace=kube-system --lock-object-name=my-scheduler, because in a highly available cluster scenario (e.g. more than one master node), you’d want your custom scheduler to be highly available, too, and distributed on other master nodes–but only one should be “active” at a time.
Craig Shea:
Going to try removing those and see if it works…
Craig Shea:
Nope–that wasn’t it. No matter what I do, my scheduler is running and can schedule pods, but question 3 will not be marked as correct.
Craig Shea:
Even switching to use only HTTP (vs. HTTPS) did not change the result. Question 3 does not ever pass for me.
Craig Shea:
I wonder–@Bukola, what version of kube-scheduler are you using? The answer file uses v1.16. But the version deployed on my cluster is v1.19. I wonder if the test is looking for a specific version of the kube-scheduler…guess I’ll try changing the image version…
Craig Shea:
Nope…that wasn’t it either. Oh well. For me, question 3 never passes…but the scheduler is working, as you and I have both observed.
Bukola:
okay let me share my manifest here and you can compare
Craig Shea:
DUH…I know what the problem is…the name of the pod comes up as my-scheduler-controlplane…
Craig Shea:
Definitely…you can see it in my original YAML.