Hybrid beamforming that combines analog phase shift and digital processing is a promising solution to alleviate the implementation cost and power consumption in millimeter wave (mmWave) massive MIMO systems. To achieve sum rate optimization, an iterative scheme that considers both the analog and digital precoders in the optimization metric is proposed. The digital precoder is mainly used to remove multiple access interference for enhancing the multiplexing gain. Then, the coordinated beamforming procedure between the base station and mobile stations is performed to exploit the array gain. Although the iterative scheme is used, only few iterations are required. This scheme can also be extended to the systems using precoding and combining vectors from quantized codebooks. Furthermore, with the channel reciprocity property or the two-phase training strategy, the iterative scheme can obtain the required channel information. From the simulation results, the sum rate is improved, especially when quantized codebooks are used.