Once BristolTopGroup/NTupleProduction#152 is done we need to re-enabled the PU reweighting.
For now we can use the 8 TeV data 'true' pileup histogram, but we need to produce the true PU profile for the Spring15 MC.
Some documentation:
https://twiki.cern.ch/twiki/bin/view/CMS/PileupJSONFileforData