This study is aimed to smooth the job flow between consecutive production shops along a production line. We propose solutions to a factory that wants an upstream shop to complete every job at a timing that the downstream shop can process the job. We model the upstream shop as a reentrant flexible job shop and propose three online executable methods forming a control system. One is a due date setting method that estimates the flow time of a job by making a quick schedule of jobs in bottleneck workstations. Another is a dispatching rule that sequences jobs according to urgency degree. The other is a heuristic that controls the start time of processing the final operation in a job. Simulations by using data from the factory show that each method is superior to its rivals in the literature, and when working together, the methods can not only complete jobs at the desired timings but also reach instantly far better solutions than metaheuristics customized for solving similar problems.