By Karla Arndt, IBM
Runtime Diagnostics is a “point-in-time” diagnostics tool that detects problems that could be contributing to system issues RIGHT NOW with the goal of finding the problems in sixty seconds or less. It has been available since V1.12 and as of V1.13, diagnoses seven types of problems which are documented in the redbook http://www.redbooks.ibm.com/abstracts/sg248070.html?Open and in z/OS Problem Management.
There are now two new diagnostic events: JES2 Health Exception and Server Health events. The JES2 Health Exception event is available now – starting in V2.1 with APAR OA46531. The Server Health event is available starting in V2.2.
When Runtime Diagnostics is invoked with this APAR via the existing F HZR,ANALYZE command, it gathers information about the JES2 subsystem from the JES2 subsystem interface (SSI). Runtime Diagnostics analyzes the information received, determines a possible corrective action, and presents it to the caller on the system console, the hardcopy log, and optionally, to a sequential dataset. An example of a possible JES2 event appears below:
HZR0200I RUNTIME DIAGNOSTICS RESULT
REQ: 004 TARGET SYSTEM: SY1 HOME: SY1 2015/01/12
INTERVAL: 60 MINUTES
FOUND: 01 - PRIORITIES: HIGH:01 MED:00 LOW:00
EVENT 01: HIGH - JES2 - SYSTEM: SY1 2015/01/12
$HASP9158 JES2 PROCESSING STOPPED, $S NEEDED
ERROR: JES2 CANNOT PROCESS NEW WORK.
ACTION: $S TO ENABLE JES2 TO START PROCESSING NEW WORK.
Also, z/OS V2.2 provides enhanced health-based workload routing in a Parallel Sysplex with new z/OS Workload Manager (WLM) and XCF functions to improve availability. As part of that z/OS V2.2 feature, Runtime Diagnostics invokes a new WLM server health query service whenever you request it to do analysis. If any servers have a current health value less than 100, it displays a SERVERHEALTH event in its output along with all the other events it finds. More information on this new event as well as example event output will be documented in z/OS V2R2 Problem Management when it is available.
As an additional feature, some of the Predictive Failure Analysis (PFA) checks invoke Runtime Diagnostics when PFA determines that the metric value is too low. The new JES2 and SERVERHEALTH analysis types are automatically integrated into that feature and will be returned to PFA by Runtime Diagnostics if they exist. PFA will then include those events in the PFA check's exception report.
The next time your system seems to be having issues, give Runtime Diagnostics a try!