I know this reply is quite late but I saw the problem and wanted to answer in case anyone else was confused about this.
Integrating along the x-axis instead of the rod is, of course, absolutely fine, however in this case your infinitesimal mass element dm should be dm = D*(dx/cosB), as each...