Interpreting the formatted global trace table

The formatted Db2 trace tables are printed in time sequence from the oldest to the most current. To begin an analysis, review the most current entries (these appear at the end). Continue analyzing the information by scanning backward.

The following fields are the most significant fields in each trace entry:

RMID
The resource manager ID that made the trace entry.
Event number
The event number unique to the resource manager.

Refer to Figure 1 and follow these guidelines for interpreting that trace table:

  1. Determine the address of the failing EB (execution block). This helps pinpoint some of the significant trace entries to examine.

    The address of the failing EB is listed under Key 14 of the VRA Diagnostic Information Report (at the beginning of the formatted dump) and at offset X'30' in the SDWA (system diagnostic work area). Printing and analyzing dumps provides more information.

    For the trace table shown in the example, the failing EB is at X'01E30480'. (The SDWA and VRA Diagnostic Information Report are not shown.)

  2. Turn to the end of the formatted trace table. The last entry is most current, but might not always correspond to the failing EB.
  3. Scanning backwards through the table, locate the most recent entry for the failing EB.

    In the example, the last entry is the most recent entry corresponding to the failing EB. The EID (event identifier) indicates this entry is for the agent services manager (RMID = X'02'), event number X'0001'. The EID trace code tables show this involves an entry in the SUSPEND execution unit, and that the address of the resource options block (ROB) is recorded (to the right of the DATA column). There is no corresponding exit from SUSPEND (because this is the most recent entry and also the end of the table).

  4. You can continue to scan backwards through the trace table, interpreting the entries for the failing EB as needed.