The methyl and methylene protons are magnetically nonequivalent by the chemical nonequivalence criterion. In other words, both sets of protons are not identical and thus they produce two different signals.
If they were identical, there would still be two signals, only they would overlap one another and the integration would reflect this feature.
As Borek was asking, there's not a broad singlet somewhere? Usually heteroatoms like O or N produce broad singlets when bonded to H.