gtpc3m19 | Concepts and Structures |
Errors can occur at any point during TPF operations. There can be
programming errors, such as incorrect macro parameters, or hardware
malfunctions, or a variety of unusual conditions. The levels of error
are:
- Hardware malfunctions that are overcome by retrying the I/O
operation. In this case, error statistics are recorded but the Entry is
protected from the problem (for example, unit checks).
- An error detected by the system from which the programs related to an
Entry may be able to recover. In this case, the ECB-controlled program
regains control (for example, a record identification check).
- An error detected by the system from which the Entry cannot
recover. In this case, the Entry is forced to exit (for example, an
addressing exception generated by an ECB-controlled program).
- An error detected by the system that makes continued operation of the
system inadvisable. This is called a catastrophic error; for
example, an operation exception (programming error) within the control
program. Such a failure is detected and handled by various components
of the control program and may require a system restart.
The basic objectives for TPF system error processing are: