The brightest and longest-lasting supernovae radiate about about a hundred times more energy than typical events. These “superluminous” supernovae were discovered only about 15 years ago, and their stellar progenitors and the mechanisms by which they radiate so much energy are under much debate. SN 2006gy was one of the first observed such supernovae, and still one of the brightest ever seen. It showed signs of circumstellar interaction with a massive hydrogen-rich envelope that must have been ejected shortly before the supernova. In 2008 spectroscopy of the ejecta revealed emission lines that had not been seen before in any supernova and could not be identified. I present a resolution to this long-standing mystery, showing that the lines can be convincingly identified with neutral iron. Further, the properties of the lines point towards a large mass of iron in the ejecta, at least a third of a solar mass.  I discuss how this new discovery means problems for most models proposed so far for SN 2006gy – all involving the explosion of a very massive star – and outline a path toward a new scenario instead involving a white dwarf spiralling into the envelope of a red giant or supergiant companion.