DAE - Dump Analysis and Elimination

Enhanced Definition

DAE (Dump Analysis and Elimination) is a z/OS component designed to automatically analyze and suppress duplicate or non-critical system dumps (SVC dumps) to conserve system resources, reduce DASD consumption, and minimize operational overhead. It prevents the creation of redundant dumps for recurring problems by comparing new dump requests against a history of previously taken dumps.

Key Characteristics

- Automated Dump Suppression: DAE intercepts SVC dump requests and, based on predefined criteria, decides whether to take the dump, suppress it, or take it selectively.
- Dump Directory: It maintains an in-memory dump directory that records key characteristics of recent dumps, such as abend code, program name, PSW, and ASID, for comparison with new requests.
- Configuration via PARMLIB: DAE's behavior is controlled by parameters defined in the ADEDAE00 member of SYS1.PARMLIB, allowing administrators to specify which types of dumps to suppress, for how long, and for which address spaces.
- Message Generation: When a dump is suppressed, DAE issues console messages (ADED0001I, ADED0002I) indicating the suppression and the reason, ensuring transparency and auditability.
- Resource Conservation: By preventing unnecessary dump creation, DAE reduces the demand for DASD space, CPU cycles for dump processing, and I/O operations associated with writing large dump files.
- Selective Dump Taking: DAE can be configured to take a dump only after a certain number of suppressions for a specific problem, or to always take a dump for critical abend codes.

Use Cases

- Reducing Redundant SVC Dumps: Preventing the system from taking multiple identical SVC dumps when a recurring application or system error occurs, which is common in busy production environments.
- Managing Test/Development Environments: Suppressing dumps for known or expected abends in test regions, where frequent program failures are part of the development and testing cycle, without completely disabling dump capture.
- Optimizing DASD Usage: Minimizing the consumption of valuable DASD space that would otherwise be filled by numerous large dump datasets, which can quickly accumulate and lead to storage management issues.
- Improving System Performance: Reducing the overhead associated with dump processing, including CPU time for dump formatting and I/O for writing to dump datasets, thereby contributing to overall system responsiveness.
- Focusing Problem Determination Efforts: Allowing system programmers and support staff to concentrate on unique and critical system issues by filtering out repetitive, non-actionable dumps.

Related Concepts

DAE works closely with the z/OS SVC dump mechanism, acting as a gatekeeper to control which dump requests are honored. Its configuration is part of the overall SYS1.PARMLIB setup, which governs many aspects of z/OS system behavior. While DAE reduces the number of dumps, the remaining dumps are still processed by IPCS (Interactive Problem Control System) for analysis. DAE's actions are recorded in system logs, often integrated with System Logger facilities, providing an audit trail for system administrators. It complements other problem determination tools by ensuring that only relevant diagnostic data is collected.

Best Practices:

Careful Configuration: Configure ADEDAE00 judiciously, ensuring that critical abend codes or system components are never suppressed, or are suppressed only after a specific threshold is met.
Regular Review of DAE Activity: Monitor ADEDAE00 messages in the system log (SYSLOG) to understand which dumps are being suppressed and why, ensuring that no critical issues are being overlooked.
Coordinate with Application Teams: Work with application developers to understand common abends in their programs and configure DAE to suppress non-critical ones, while ensuring that new or unknown issues still trigger a dump.
Periodically Clean Up Dump Datasets: Even with DAE, dump datasets can accumulate. Implement a regular process for offloading, archiving, or deleting old dump datasets to manage DASD space effectively.
Test DAE Changes: Thoroughly test any changes to the ADEDAE00 member in a non-production environment before implementing them in production to prevent unintended dump suppressions.