CodeRunner doesn't support CBM - it always uses its own custom behaviour, an extension of Adaptive. If all the questions in the quiz are CodeRunner, then when the quiz review attempts to summarise the results by computing the average accuracy and average cbm, it comes up with 0 / 0 for both.
Similar problems apparently occur with a quiz containing only essay questions. See the bug report here: DivisionByZero while reviewing attempt in quiz using CBM behaviour.
I think (but please check) that the problem will go away if you add just one non-CodeRunner non-essay question. Given that CodeRunner questions don't present the user with an option to specify their confidence in their result, is there a reason to use CBM behaviour for the quiz that contain only CodeRunner questions?
You may want to add a comment to the bug report linked above, pointing out that the issue occurs with CodeRunner questions as well.
As to why you're getting the problem on one of your systems and not the other: my guess is that one of the servers is set to display warnings, the other not.