A known failure mode of vaccines is if the person who is being vaccinated lacks the right B cells for creating antibodies matching the antigens on the surface of a pathogen. If the B cells a person has lack the right B Cell receptors, they can't generate the correct antibody response to an antigen, because they can't bind with/identify an antigen in the first place. Additionally, they may lack a sufficient number of antigen naive T cells, which haven't yet become differentiated as T memory cells and effector T cells.

But that doesn't mean that it is impossible to create a vaccine: this problem exists already for successful vaccines. For example, the Hepatitis B vaccine has a relatively high failure rate, but is still provides immunity 90+% of the time. It just means that some smallish percentage of the people you vaccinate aren't actually contributing to population immunity.

See below for more info:

https://en.wikipedia.org/wiki/B_cell

https://en.wikipedia.org/wiki/Memory_B_cell

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3760154/ --> this one is particularly good. A lot of good info in here on how vaccination and "memory" interact.

https://www.sciencedirect.com/science/article/pii/S0264410X0...

EDIT: this one covers the major routes of vaccination failure --> https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4962729/#!po=60...

It’s possible and desirable for a vaccine to stimulate antibody production where a “natural” infection would not, which is one of the things that will be tested during their development. An extreme example is infection with the tetanus-causing bacterium clostridium tetani, which doesn’t provide natural immunity at all.