In this paper, we propose a dynamic and reliability- driven real-time fault-tolerant scheduling algorithm on heterogeneous distributed systems (DYFARS). Primary-backup copy scheme is leveraged by DYFARS to tolerate both hardware and software failures. Most importantly, DYFARS employs reliability costs as its main objective to dynamically schedule independent, non-preemptive aperiodic tasks, therefore system reliability is enhanced without additional hardware costs. A salient difference between our DYFARS and existing scheduling approaches is that DYFARS considers backup copies in both active and passive forms; therefore, DYFARS is more flexible than the existing scheduling schemes in the literature. Finally, simulation experiments are carried out to compare DYFARS with existing similar algorithm, experiment results show that DYFARS is superior to existing algorithm regarding both schedulability and reliability.