Datacom Log Area (LXX) Spills Automated Operations Routine Management

Products

• Datacom
• Datacom/AD
• Datacom/DB

Issue/Introduction

Automated operations monitoring and management tools play a critical role in 'lights-out' data centers. These tools respond to console messages from application jobs and system tasks to execute commands traditionally performed by human operators. This article provides guidance on using these routines to manage Datacom Log Area (LXX) spills on z/OS systems to prevent system halts caused by log capacity issues.

Environment

• Product: Datacom/DB, Datacom/AD
• Release: All Supported Releases
• Platform: z/OS

Resolution

Datacom supports automated management of Log Area (LXX) files. Automated routines capture specific console messages to determine if a Log Area requires a spill to a Recovery File (RXX). These routines then build and submit the necessary spill job directly. 

Job Submission Note: Do not submit the spill jobs through scheduling applications that interface with Datacom databases. If the MUF LXX reaches 100% full, all database activity ceases. In this state, a database-dependent scheduling application cannot initiate new tasks, preventing the spill job from being submitted.

Configuration Considerations

MUF Message Formatting

By default, Datacom displays messages in the following format:

[JOBNAME]:DB0xxxxI [MESSAGE_TEXT]

To simplify automation scripts across multiple MUFs, the jobname prefix can be disabled. This allows a single routine to manage all MUFs, making logic distinctions based on the job issuing the message. To disable this prefix, use the MUF Startup Option below and restart the MUF:

MUFMSG NO,NO,NO

Threshold Tuning

Unless modified by the LOGSPILL startup option, messages appear on the console based on the following default thresholds:

• 75% Full: Initial message
• 80% – 90% Full: Messages every 5%
• 90%+ Full: Messages every 1% until 100%

Automated Processing Messages

Configure automation software to respond to these primary Datacom messages:

• DB00308I - LOG AREA IS nnn% FULL, nnnnnn BLOCKS SPILLABLE
• DB00309I - LOG AREA IS nn% FULL, JOB [JOBNAME] [ID=monitor] PREVENTING SPILL
• DB00310I - LOG SPILL rrrrr-ccccc, BLOCKS xxxxx-xxxxx, TIMES ccyy/mm/dd hh.mm.ss-ccyy/mm/dd hh.mm.ss
• DB00311I - LOG AREA IS nn% FULL, SPILLING BLOCK nnnnn, OUT OF nnnnn

Steps to configure the system

1. Set the automation routine to trigger when message DB00308I is displayed.
2. Program the routine to submit a predefined JCL member for the MUF issuing the message. Alternatively, use a 'skeleton' JCL member customized by the automation routine prior to submission.
3. Monitor DB00309I to identify specific jobs preventing a spill. Configure alerts for this message to allow for manual intervention if the log continues to fill.
4. Refer to the following technical documentation for additional logging and recovery details:
• Using Logging
• Operating with Logging Active (paragraph Processing With an Inactive Recovery File)
• Datacom Messages Documentation

 

Additional Information