Setting Up API Trace Data Sets

Before you start using the API trace facility, you must have at least one pre-allocated, sequential data set to contain the trace data. APPC/MVS limits the number of API trace data sets per z/OS system to fifty. This limit is imposed to control the possible impact on performance; collecting and writing trace data are resource-intensive activities that might affect not only APPC/MVS work, but also other system work. On a test system, performance might not be a major concern; on a production system, you may use API tracing options that further limit tracing activity to an acceptable level, with little or no impact on performance.

In any case, because of the limited number of trace data sets, your installation might have to consider various approaches to setting up the API trace data sets, to ensure greatest efficiency and ease of use. Installation set-up affects what you need to know before using the API trace facility:
  • Security requirements for programmers who use the data sets
  • Characteristics to select when you allocate the data sets
  • Techniques for avoiding loss of data through wrapping
  • Techniques for avoiding loss of data when APPC/MVS suspends tracing activity
  • Suggestions that might increase efficient use of the trace data sets.