Interpreting extended-format CICS system trace entries

CICS system trace entries made to the internal trace table, the auxiliary trace data sets, and the GTF trace data set can all be formatted to give the same type of information.

About this task

There are two slightly different extended trace entry formats. The short style resembles the format used in earlier releases of CICS, and gives FIELD A and FIELD B values. The other long style uses a different format, described below.

Procedure

  1. Look at the trace point ID. This is an identifier that indicates where the trace point is in CICS code. In the case of application (AP) domain, the request type field included in the entry is also needed to uniquely identify the trace point. For all other domains, each trace point has a unique trace point ID.

    Its format is always a two-character domain index, showing which domain the trace point is in, then a space, then a four-digit (two-byte) hexadecimal number identifying the trace point within the domain.

    The following are examples of trace point IDs:
    AP 00E1         trace point X'00EE' in Application Domain
    DS 0005         trace point X'0005' in Dispatcher Domain
    TI 0101         trace point X'0101' in Timer Domain
  2. Look at the interpretation string. It shows:
    • The module where the trace point is located
    • The function being performed
    • Any parameters passed on a call, and any response from a called routine.
  3. Look at the standard information string. It shows:
    • The task number, which is used to identify a task uniquely for as long as it is in the system. It provides a simple way of locating trace entries associated with specific tasks, as follows:
      • A five-digit decimal number shows that this is a trace entry for a task with a TCA, the value being taken from field TCAKCTTA of the TCA.
      • A three-character non-numeric value in this field shows that the trace entry is for a system task. You could, for example, see “III” (initialization), or “TCP” (terminal control).
      • A two-character domain index in this field shows that the trace entry is for a task without a TCA. The index identifies the domain that attached the task.
    • The kernel task number (KE_NUM), which is the number used by the kernel domain to identify the task. The same KE_NUM value for the task is shown in the kernel task summary in the formatted system dump.
    • The time when the trace entry was made. (Note that the GTF trace time is GMT time.)
    • The interval that elapsed between this and the previous trace entry, in seconds.
    The standard information string gives two other pieces of useful information:
    • The CICS TCB ID and the address of the MVS TCB (field TCB ) that is in use for this task. This field can help you in comparing a CICS trace with the corresponding MVS trace. As there can be multiple OTE TCBs, the TCB ID for an OTE TCB is in the format ccnnn where cc identifies the type of OTE TCB (e.g. X9, L8 etc.) and nnn is a sequence number identifying which of the OTE TCBs is in use.
    • The return address (field RET), passed in Register 14 to a called routine. This field helps by showing what invoked the module that is making this trace entry.
  4. Read the data fields, that contain information relevant to the function being performed.

    For short trace points, these are shown as fixed length (4-byte) FIELD A and FIELD B values in the same line as the interpretation string. Both the hexadecimal data values and any printable EBCDIC characters that they represent are shown. Some short trace entries also have a RESOURCE field. When provided, it is usually the name of a resource associated with the request being traced. For example, for program control requests, it is the program name.

    For long trace points, 1–7 variable-length data fields can be given. They are shown immediately below the standard information line. Any printable EBCDIC characters represented by byte values in the data fields are shown on the right of the trace.

    Some of the data fields in the new trace entries contain material intended for use by IBM® support personnel and you cannot interpret them directly. However, there is enough information to resolve user errors and for IBM support personnel to resolve most system errors in the interpretation string for the entry.



dfhs14z.html | Timestamp icon Last updated: Monday, 5 June 2017