The new strategy for particle identification samples in Run 2 at LHCb

For Run 2 of LHCb data taking, the selection of PID calibration samples is implemented in the high level trigger. A further processing is needed to provide background-subtracted samples to determine the PID performance, or to develop new algorithms for the evaluation of the detector performance in upgrade scenarios. This is achieved through a centralised production which makes efficient use of LHCb computing resources. This poster presents the major steps of the implementation.