We report on a detailed analysis of a SO(10) SUSY GUT model of Dermisek and Raby (DR) with a D3 family symmetry. The model is completely specified in terms of only 24 parameters and is able to successfully describe both quark and lepton masses and mixings, except for |Vub| that turns out to be too low. However, a global fit shows that flavor changing (FC) processes like Bs --> mu+ mu-, Bs-mixing, B+ --> tau+ nu, B --> Xs gamma and B --> Xs l+ l- pose a serious problem to the DR model. The simultaneous description of these FC processes forces squarks to have masses well above 1 TeV, not appealing on grounds of naturalness and probably beyond the reach of the LHC.