EPV Technologies: May 2026 Newsletter

In the May 2026 edition, EPV Technologies delivers critical insights into mainframe performance evaluation following an IBM upgrade, analyzes key IBM MQ log read performance metrics, and addresses a frequent customer technical question regarding SMF 30 records.

Key Technical Insights

1. Is My z17 Performing as Expected? (Part 1)

When upgrading to a new IBM machine, organizations heavily rely on LSPR benchmarks and tools like IBM zPCR to plan capacity. However, post-upgrade reality doesn't always clearly match post-upgrade expectations.

EPV Technologies has introduced a comprehensive methodology designed to help performance analysts:

• Set and accurately check post-migration performance expectations.
• Select the correct baseline data and relevant metrics to compare old vs. new hardware.
• Apply these steps to real-world migration case studies, providing a reliable guiding framework for your team.

2. IBM MQ Performance Metrics: Log Reads

During normal mainframe activity, MQ logs are only read when applications roll back uncommitted, persistent messages. To optimize these requests, EPV highlights three distinct locations where unit-of-recovery records are retrieved, along with their associated metrics:

• QJSTRBUF (In-Memory Buffers): The optimal scenario. Most read requests should be satisfied here to maximize efficiency.
• QJSTRACT (Active Log Data Sets): Indicates data is no longer in memory. This often points to a long-running task failing to commit frequently enough, or an OUTBUFF parameter set too low.
• QJSTRARH (Archive Log Data Sets): The least efficient location. This indicates a very long-running task that has delayed commits for an extended period, or that the size/number of active logs needs adjustment.

Technical Tip: If you notice high active or archive log reads, activate the Class 3 accounting trace and evaluate the QWACBACK (rollbacks) and QWACCOMM (commits) fields to pinpoint the culprit application.

3. Customer Q&A: Measuring Job Initiator Delays

• The Question: Is there a metric in SMF 30 records that reports the delay in assigning an initiator to a submitted job when experiencing class saturation?
• The Answer: Yes. The SMF30SQT field measures "initiator wait time". It tracks the exact elapsed time a batch job was eligible to run but sat waiting for an available JES initiator in its assigned job class. High values indicate it is time to review your JES initiator configurations.

Request the Full Technical Papers

You can download the newsletter here for more info. To read the complete white papers, access full technical documentation, or review past editions, visit the EPV Technologies website.