The job journal is a temporary sequential data set that resides on a JES2 spool volume.It preserves a set of selected job-related control blocks that are necessary for restart processing for the following situations: automatic step (using the RD= parameter on the JCL JOB statement), automatic checkpoint, continue, and JES2 member warm starts.

The job journal is necessary because the scheduler control blocks, maintained in the scheduler work area (SWA), are otherwise lost when a job abnormally ends. Reconstruction of the SWA is possible because the job journal preserves up-to-date copies of essential control blocks.

Unless you specify no job journal by defining JOURNAL=NO on the JOBCLASS initialization statement, each job is provided with a job journal. If you are using checkpoint restart or need to restart a job step, you need a job journal or automatic restarting will not be possible (and jobs that are executing during member failure need to be resubmitted).

However, if you are using automatic restart management to restart a job, you should use non-journaled classes.

Note that job journaling can cause some performance degradation. For further information regarding the JOURNAL parameter, see the JOBCLASS initialization statements in z/OS JES2 Initialization and Tuning Reference.