HP-UX SNAplus2 Diagnostics Guide

Using Logging and Tracing
Tracing
Chapter 386
application first writes to it. Normally, the file is reset
and its contents discarded when an application writes
its first trace message to the file. If you are tracing two
or more applications to the same file, or if you want to
trace two or more runs of the same application to the
same file, you can prevent the file from being reset by
setting the SNATRCRESET environment variable to NO.
If you are tracing to two files, the files continue to be
reset as normal when themaximum file size is reached,
but they will not be reset when an application starts
tracing for the first time. If you are tracing to one file,
setting SNATRCRESET to NO means that the file will
never be reset automatically. To avoid taking up too
much disk space, delete it manually from time to time.
To cancel the setting of SNATRCRESET and return to the
default setting so that the file is reset when an
application first traces to it, set SNATRCRESET to a null
string.
Trace File Format for API Tracing
The trace data for a single message can occupy more than one line in the
trace file. Each individual message is preceded by a horizontal line
indicating the time the trace entry was made. The following describes a
SNAplus2 API trace file:
• The process ID of the component being traced appears at the start of
each line. After the process id is the thread id (separated by a period).
For a single-threaded application, this will always be 00. The process
ID is followed by an indicator of the component type being traced (for
example, APPC, for an APPC TP).
• Message data is shown in the following formats in separate columns
to ensure that a character string in the message data appears as
readable text in either the EBCDIC or the ASCII column, according
to its character set:
— Hexadecimal
— Interpreted as EBCDIC
— Interpreted as ASCII
The format of trace data varies slightly among the APIs: