Hard - Permanent or physical
In mainframe computing, the term "hard" often describes attributes that are physical, permanent, fixed, or difficult to change without significant system intervention. It contrasts with "soft" attributes, which are typically logical, temporary, or easily modifiable. This distinction is crucial for understanding system architecture, data persistence, and error handling.
Key Characteristics
-
- Physicality: Refers to tangible hardware components like disk drives (DASD), tape units, or printers, which are physical resources.
- Permanence/Persistence: Implies data or configurations that endure across system restarts or job executions, such as data stored on DASD or system parameters defined in
PARMLIB. - Immutability/Fixed Nature: Describes values or configurations that are hard-coded into programs or system definitions, requiring recompilation, reassembly, or system re-IPL to alter.
- Severity in Error Handling: A "hard error" indicates a persistent, often unrecoverable hardware or software fault, requiring manual intervention, component replacement, or a system restart.
- Direct Output: "Hardcopy" refers to physical printouts of reports, logs, or system console messages, providing a tangible, non-volatile record.
Use Cases
-
- Hardcopy Output: Generating physical reports, invoices, or system logs on line printers (e.g., using
SYSOUTto a printer class) for archival, regulatory compliance, or distribution. - Hard-coded Values: Embedding fixed constants (e.g., specific file names, system IDs, fixed thresholds) directly into COBOL or PL/I application programs, making them difficult to change without code modification.
- Hard Disk Storage (DASD): Referring to the underlying physical direct access storage devices where permanent datasets (e.g., VSAM, sequential, PDS, PDSE) reside, ensuring data persistence.
- Hard IPL (Initial Program Load): Performing a complete system shutdown and restart, often necessary after significant system software changes, hardware maintenance, or to clear severe, unrecoverable system states.
- Hard Errors: Diagnosing and resolving persistent hardware failures (e.g., a failing disk drive, a memory error) or unrecoverable software abends that prevent normal system operation.
- Hardcopy Output: Generating physical reports, invoices, or system logs on line printers (e.g., using
Related Concepts
The concept of "hard" is fundamental to distinguishing between physical infrastructure and logical constructs in z/OS. It relates directly to DASD (Direct Access Storage Devices) as the physical medium for permanent data storage, and JCL (Job Control Language) for defining physical resource allocation (e.g., DD statements for datasets on specific volumes). It also contrasts with virtual storage and logical units, which abstract the underlying physical components, and with soft errors that are transient or recoverable.
- Minimize Hard-coding: Avoid embedding literal values directly into application code; instead, use external configuration files, parameter libraries (e.g.,
PARMLIB), or database entries for flexibility and easier maintenance. - Secure Hardcopy Output: Treat physical printouts containing sensitive data with the same security protocols as electronic data, ensuring proper disposal or secure storage in accordance with data privacy regulations.
- Monitor Hard Errors Proactively: Implement robust system monitoring (e.g., using RMF, SMF, or third-party tools) to detect and alert on hardware and software hard errors to enable timely intervention and prevent outages.
- Plan Hard IPLs Carefully: Schedule and plan system IPLs meticulously, especially those involving significant configuration changes, ensuring all necessary system programmers and operators are available and recovery procedures are documented.
- Understand Physical Resource Allocation: For optimal performance and reliability, understand how JCL
DDstatements map to physical