Planning an application trace

This information describes the design considerations needed to implement component trace. Decisions are needed for the following tasks to provide application tracing:

You Decide:	Where to Look:
When to trace and whether to provide minimum options (MINOPS) that cannot be turned off.	See When to trace.
Where the trace points will be and what the trace entries will contain.	See Where and what to trace.
How to create and control trace buffers that contain trace entries.	See Creating trace buffers.
Whether to provide multiple traces, which are called *sublevel traces*, and how many traces your installation will need.	See Using multiple traces.
Whether to set up user-defined trace options. Component trace allows you to categorize your trace entries and create only those you select at a given time.	See Setting up user-defined options.
Whether a display exit routine is required.	See Coding a display trace exit routine.
How to provide default tracing options	See Starting, stopping and changing the trace.
How you will externalize your data. You can: Request a dump Combine the trace with other trace output Use the component trace external writer (called the external writer in the remainder of this text) to write data directly to a data set.	See Externalizing trace data in a dump, or Externalizing trace data through the external writer.
How you will format your data once it is externalized. Component trace supports formatting through IPCS. If you do not use IPCS, you will have to format the data yourself.	See z/OS MVS IPCS Customization and z/OS MVS IPCS Commands for information on using IPCS.

Understanding the Sequence of Trace Activities: The following is a simplified description of tracing using component trace.

The programmer plans and sets up the user-defined control structures.
The application defines itself to component trace.
The operator turns the trace on, either because of a problem or because of a message from the application.
The system passes control to the start/stop exit routine to start tracing.
The application starts tracing by creating trace entries and placing them in trace buffers.
The application or operator (or both) dumps the trace buffers.
The operator turns off the trace.
The system passes control to the start/stop exit routine to stop tracing, and optionally dump the trace buffers.
The application deletes itself from component trace.
The system programmer formats the trace buffers through IPCS.

Using component trace involves a number of trace-related activities. The following tables summarize the trace-related activities and how to accomplish each activity.