...
Start defines the starting point for job execution. Every job must have one (and only one) Start. Unconditional job hops only are available from a Start job entry.
The start job entry settings contain basic scheduling functionality; however, scheduling is not persistent and is only available while the device is running. The basic scheduling functionality and the repeat option are only functional within the main job and not within a sub job.
The Data Integration Server provides a more robust option for scheduling execution of jobs and transformations and is the preferred alternative to scheduling using the Start step. If you want the job to run like a daemon process, however, enable Repeat in the job settings dialog box.
...