This information describes:
- Creating component trace entries (CTEs) and placing them in trace
buffers
- Understanding the fields in a CTE
- Organizing
the variable data in a CTE.
Your trace entries can be in any format. However, to be able to
use IPCS to display them, use the ITTCTE mapping macro to create CTEs.
You can trace a maximum of 60KB of application-unique data in each
CTE.
Figure 1 is an example of a CTE
with a trailing length,
which is a length field at the end of the entry.
Figure 1. Component Trace Entry in
a Trace Buffer
Follow this procedure when creating your CTEs
and placing them
in buffers:
- Start the trace buffer with a complete CTE at
offset 0.
- Continue adding CTEs without any gaps between the
trailing length
field (CTELENE) of the previous CTE and the starting length field
of the next CTE (CTELENP).
- When the buffer reaches a stage
where the entire next CTE cannot
fit, IBM® recommends that you
clear the remaining portion of the trace buffer to zeros. At this
point, you have the choice of either using a new buffer or wrapping
and reusing the same buffer. If you are using the external writer,
use a new buffer so that component trace can capture your full trace
buffer asynchronously while you are writing trace entries to a new
buffer. If you do not require trace data to be captured asynchronously,
or if you can afford to lose old trace data, you might reuse the same
buffer.
Once the application creates all its trace
entries, places them
in trace buffers in storage, and externalizes them either in a dump
or by using the external writer, IPCS can format each entry according
to instructions you specify in a CTRACE format table. Use the FMTTAB
parameter on CTRACE DEFINE to specify the name of the load module
that contains the CTRACE format table. See z/OS MVS IPCS Customization for
information about creating
a CTRACE format table.