Back when I worked on IBM's J9, the nightmare scenario was either a dropped memory fence, or the map keeping track of links between register / stack references

Back when I worked on IBM's J9, the nightmare scenario was either a dropped memory fence, or the map keeping track of links between register / stack references and GC heap objects getting screwed up. Things could go on a very long time before those errors manifest visibly.