Abnormal termination (ABEND or ABENDU) procedure

A malfunction in the system can cause an abnormal termination (abend). Abend categories are:
  • User abend
  • System abend
  • Program check abend.

User abends originate in the application program. HCL Workload Automation for Z abend codes are documented in Abend codes and Messages and Codes

System abends can occur, for example, when a system supervisor-call instruction (SVC) is issued in a program with an incorrect event-control-block (ECB) address. Refer to z/OS Codes and z/OS Messages

Program check abends are hardware-detected error conditions, such as a store to an invalid address or an attempt to execute an invalid instruction.

To correct the problem, you need to know which part of the code the problem is in. For any abend, you need to determine which module failed. For example, was the failure in:
  • An HCL Workload Automation for Z module?
  • User-written code, an exit, or an HCL Workload Automation for Z program interface (PIF) application?
  • Another IBM® licensed program running in the HCL Workload Automation for Z address space?
  • A PIF application running in another address space?
  • The Dynamic Workload Console?
If your HCL Workload Automation for Z problem is an abnormal termination, you should:
  1. Determine the name of the active HCL Workload Automation for Z task. The active task can be determined only if you have a dump data set. You can use the SYSMDUMP dump data set and format the output using interactive problem control system (IPCS).
  2. Check the name of the failing load module. If the first 3 characters are EQQ, an HCL Workload Automation for Z module has abended.
  3. Locate and record the abend code.
  4. Locate and record the name of the failing module and the offset of the failing instruction in the current CSECT of that module. The address of the failing instruction is given by the program status word (PSW) at entry to the abend information.

    Locate the failing instruction in the dump. The base register will normally be register 11 or 12. Determine the values of these registers by locating the Registers at entry to ABEND area in the dump. Either register 11 or register 12 will have a value that is less than the address of the failing instruction but within X'1000' of this address. This register is the base register, which you use to find the beginning of the failing CSECT in the dump. At offset 5 in the failing CSECT, you will find a module identification that contains the name of the failing CSECT and the compilation date of this CSECT.

  5. Record the name and compilation date of the failing CSECT. Calculate and record the offset of the failing instruction in the failing CSECT.
  6. Locate the save-area trace information in the dump. You should be able to find a line containing AT EP entry-point-name in the save-area trace. The entry-point-name is the module identification of the failing CSECT. The first word on the same line is the name of the failing load module.
  7. Record the name of the failing load module.
  8. Compile the required information described in Information needed for all problems.

HCL Workload Automation for Z abend codes are listed in Abend codes.

Note: This procedure does not apply to the end-to-end server address space when the failing module is a USS process.