Jump to content

IBM System Management Facilities

From Alfawiki
IBM System Management Facility (SMF)
Company typeSoftware component
IndustryMainframe operating systems
ProductsSystem activity recording and monitoring for IBM z/OS
ParentIBM

IBM System Management Facility (SMF) constitutes an essential element of IBM's z/OS operating system tailored for mainframe computer environments, providing a uniform framework for capturing system activities into data sets. SMF comprehensively monitors core operations within the IBM mainframe OS, recording metrics related to I/O processes, network communications, application usage, error events, and CPU consumption.

A prominent application utilizing SMF is IBM's Resource Measurement Facility (RMF), which oversees system performance and resource allocation, including processors, memory, disk storage, cache, workload distribution, virtual memory, XCF, and the Coupling Facility. RMF is an optional priced feature within z/OS. Competing vendors such as BMC provide alternative solutions like CMF.

SMF serves as the foundation for numerous monitoring and automation systems. Each SMF record is assigned a specific numeric type (for instance, "SMF 120" or "SMF 89"), with customers having control over the volume of data collected. Records generated by third-party software generally use type numbers 128 or above. Certain record types include subcategories; an example is Type 70 Subtype 1 records that capture CPU utilization data via RMF.

SMF record types

Frequently encountered SMF record types comprise:

  • RMF-related records (types 70–79), complemented by Type 30 (subtypes 2 and 3) which provide granular performance metrics.
  • Security event logs such as those from RACF are recorded under type 80; other security products like TopSecret also utilize type 80 records. ACF2 produces comparable security data primarily in type 230 records.
  • Type 89 records document software usage pertinent to sub-capacity licensing compliance.
  • IBM Db2 emits records of types 100, 101, and 102 contingent on subsystem configurations.
  • CICS generates type 110 records in line with its setup.
  • Websphere MQ outputs types 115 and 116.
  • WebSphere Application Server for z/OS uses type 120; notably, Version 7 introduced subtype 9 to support unified request-based monitoring while minimizing overhead.

Evolving records

Core record categories, particularly those associated with RMF, undergo frequent updates with each new z/OS release introducing additional fields. Differences in processor architectures and Coupling Facility versions also influence the structure of these records.

SMF data recording

SMF captures system data through two primary mechanisms:

  • The conventional method employs buffers managed by the SMF address space alongside preallocated VSAM datasets (commonly SYS1.MANx).
  • A more recent technique utilizes System Logger log streams, enhancing write performance, eliminating buffer shortages, supporting multiple concurrent log streams, and enabling efficient data duplication.

Only one recording approach is active at any given time; the alternate method remains available as a fallback option.

Data is periodically offloaded to sequential files (such as tape media) using utilities IFASMFDP or IFASMFDL (the latter for log stream mode). Both utilities generate equivalent output formats, allowing seamless switching between them through JCL modifications without impacting downstream processing workflows.

SMF data collection and analysis

The collection of SMF data is facilitated by IBM Z Operational Log and Data Analytics along with IBM Z Anomaly Analytics with Watson. IBM Z Operational Log and Data Analytics transforms raw SMF records into formats suitable for consumption by external analytics platforms like Elastic Stack and Splunk, or its own proprietary operational analytics environment for comprehensive insight generation. IBM Z Anomaly Analytics with Watson aggregates SMF data from multiple IBM Z systems—including IBM Db2 for z/OS, IBM CICS Transaction Server for z/OS, and IBM MQ for z/OS—leveraging historical logs and metrics to establish baseline behavior models and applying real-time analysis to identify anomalies and proactively notify IT teams.

IBM Z Operational Log and Data Analytics supports three collection modes for SMF data; IBM Z Anomaly Analytics with Watson utilizes the first two:

  • Log stream mode using the SMF in-memory buffer
In this mode, the Common Data Provider extracts SMF data directly from the in-memory buffer via the real-time SMF interface.
  • Data set recording mode
When operating under this mode, the Common Data Provider streams SMF data through user exit interfaces.
  • Batch mode
The System Data Engine component can execute standalone batch jobs to read SMF datasets and either write output files or forward data to the Data Streamer.

Analysis of SMF information is available across multiple platforms:

  • The Z Data Analytics platform within IBM Z Operational Log and Data Analytics delivers search and visualization tools across extensive operational datasets to facilitate early detection and troubleshooting of issues.
  • Enterprise-grade platforms such as Splunk, Elastic Stack, Apache Kafka, or Humio can process operational data but typically require specialized or imported analytics tailored for IBM Z environments.
  • IBM Z Anomaly Analytics with Watson applies machine learning techniques on collected logs and metrics to detect irregular patterns.
  • IBM Db2 Analytics Accelerator for z/OS offers query-driven reporting functionalities.
  • IntelliMagic Vision for z/OS provides expert-level diagnostics along with actionable recommendations based on deep system expertise.

External links

References

  • IBM Redbooks. ABCs of z/OS System Programming, Volume 2, International Technical Support Organization, July 2008.
  • BMC CMF Monitor official product documentation
  • PACSYS examples demonstrating SMF Reports
  • Streaming z/OS IT Operational Data with IBM Z Common Data Provider, Planet Mainframe, August 2020
  • Documentation and product information for IBM Z Operational Log and Data Analytics
  • IBM Z Anomaly Analytics with Watson official documentation and product details
Retrieved from "https://www.alfawiki.org/index.php?title=IBM_System_Management_Facilities&oldid=28137"