JCC
When the JCC function is used, the success or failure of an operation cannot be determined by the controller until the JCC task has processed the output on the JES spool. The JCC checks every line of SYSOUT against tables defining conditions to be detected.
Measuring JCC performance
You can measure JCC performance by:
- Using the EQQUX005 exit. Specific call indicators tell you when the exit has been called to start process a new job and when there is no more output to check for a job. Good performance in your exit is vital because it is called for every SYSOUT line unless you tell IBM Tivoli Workload Scheduler for z/OS to stop calling it.
- Observe the output queue using SDSF. The job is highlighted on the queue while it is being checked.
Performance in the JCC is influenced primarily by JES performance and also by the size of the job logs.
Recommendations
Consider these recommendations:
- If you currently route all output for JCC processing to one system in your configuration, consider checking the output on the system where the job ran to balance the JCC workload among the available processors. The JCC task cannot process multiple jobs in parallel.
- Remove unnecessary JCC processing:
- Detection of nonzero return codes, use NOERROR instead.
- Trapping NOT CATLG x, use z/OS to fail the job on step end.
- SKIP sections of the output where you cannot possibly match a condition defined in the messages tables.
- Avoid scanning user SYSOUT data sets if possible.