Job trace

Use the job trace tool to trace data in any job to help identify your problem.

Job trace is a problem analysis tool that allows you to look at what any application is doing. Use job trace as a first step in locating a problem with an application. You can turn on job trace in any job and see the call and return flows of that application. Job trace records the raw data and then stores it in a set of database files.

The job trace is performed using a series of CL commands, such as STRTRC (Start Trace), ENDTRC (End Trace), and PRTTRC (Print Trace). Starting a job trace consumes relatively few system resources. However, ending the job trace and printing the job trace require more time and processing resources. If you have a limited amount of interactive capability available on your system, you might want to submit the ENDTRC and PRTTRC to batch.

Note that if the application code is created with OPTIMIZE(40), the optimization disables call and instruction tracing. Although you can specify LICOPT (CallTracingAtHighOpt) to enable job call tracing, the optimization might still disable some calls. Therefore, job trace might not be effective when using OPTIMIZE(40).

Use job trace in these situations:

• You want to debug any job on your system. See the server table to understand the correlation between the servers and the applications and jobs they represent.

• You want to troubleshoot your sockets application.

• You are developing an application for i5/OS^® and encounter a problem. By tracing the application, you can identify the problem.

To use the CL commands to perform a job trace, have *SERVICE special authority, or be authorized to the service trace function of i5/OS through iSeries™ Navigator. See the chapter on user profiles in iSeries Security Reference for more information about this type of authority.

Use the following job trace instructions as a guide. The example demonstrates how to use job trace to troubleshoot a sockets application. Sockets adds information to the job trace output when errors are returned on the sockets APIs. You might need to specify different parameters, depending on the type of application you are troubleshooting. Note that communications trace is also useful when troubleshooting sockets applications.

1. Starting a job trace
   This action starts a job trace for one or more jobs. You can start any number of trace sessions, but active trace session identifiers must be unique across the system.
2. Re-creating the problem
   Re-create the problem by repeating the series of actions you previously took.
3. Ending a job trace
   This action ends the trace and stores the collected trace records in a set of database files. The stored trace records stay in the database files until you run the Delete Trace Data (DLTTRC) command.
4. Printing a job trace
   This action formats and writes the stored trace records to a spooled output file or to a database output file.
5. Deleting a job trace
   This action deletes the trace records that were stored in the database files as a result of the End Trace (ENDTRC) command.
6. Advanced job trace functions
   Job trace offers some advanced functions that enhance the results of the job trace.

Parent topic:

Related concepts
Trace connection Trace TCP/IP application