Long PCE dispatches can be an indication of a tuning problem, a complex process that takes a long time to complete or an error situation. If the JES2 main task is not getting enough CPU resource, it could appear to the monitor as a loop. Use the D A,JES2 command or a performance monitor to ensure that JES2 is indeed consuming CPU. When you have eliminated CPU performance problems, check if the PCE displayed is the command processor and examine the command that was entered. Some commands can take a long time to complete, such as a command to alter all 200,000 jobs or 500,000 JOEs in the system. If progress is being made (job number displayed is being updated) for a complex command, then this might be normal processing.

First, deal with the current situations by either waiting for the command to complete or ABEND JES2 and then restarting it. When this situation has been addressed, consider entering multiple commands with job ranges, for example $PJQ(1–20000), $PJQ(20001–40000), and so forth). As each command completes, JES2 can perform other processes preventing JES2 from being able to process other requests.

If the problem is not a complex command, determine if progress is being made by examining the job ID in the message. If no progress is being made, there might be a problem. First, issuing the z/OS® MVS™ DUMP command to capture a dump of the JES2 address space. Then issue a $PJES2,ABEND command followed by a $PJES2,ABEND,FORCE command to terminate JES2. When JES2 terminates, you can then restart it. Analyze the dump to determine what code is running and the cause of the long dispatch. If the problem is in exit code, or exit–invoked code, correct the exit logic. Some logic, for example, that performs adequately when there were only 20,000 jobs in the job queue might have problems with 200,000 jobs. If the problem is caused by IBM®–distributed code, contact the IBM support center.