Tips for using component trace external writer

Consider the following when using the component trace external writer:

• Do not use the same writer to trace more than one TCP/IP stack, TELNET, or OMPROUTE application. If you need to trace multiple stacks or applications, use separate writers.
• If your external writer fills up and the wrap option is on, the writer overwrites itself. If the nowrap option is on, the writer stops.
• Use REGION=0K on the trace writer procedure EXEC statement. This helps ensure that there is enough virtual memory for trace buffers.
• Use TIME=1440 on the EXEC statement. This prevents S322 abends.
• Use CONTIG on the disk space allocation of the trace data when using the WRAP option. For example: SPACE=(1024,(4096,100),,CONTIG). This ensures that the space for the trace data set is available.
• Do not specify DCB parameters for trace data sets. The writer optimizes the logical record length and block size for new trace data sets.
• Ensure that the dispatching priority of the writer is equal to or greater than the application that is being traced.
• The CTRACE or packet trace formatter messages ITT10016I, PTRPT14I or PTRPT15I indicate lost buffers. This normally occurs when packet tracing is set up with no capture filters, causing all protocols for all IP addresses to be captured. The result is that some trace buffers are unable to be processed and are flagged as lost. The reason behind these lost records is the amount of data that is being traced over-running the CTRACE writer. This trace data is arriving at the CTRACE writer faster than it can be written out. Setting filters such as IPADDR or PORTNUM will help eliminate lost records.