Subscribe
THE DAILY DIGEST

The daily edition.

Essential reading from Rocket Enterprise. Straight to the source.

Archived edition
DAILY EDITION1 story· 4 min read

Mainframe Subsystem Support (MSS) & JES Batch Emulation

In traditional IBM z/OS mainframes, batch processing is orchestrated by the Job Entry Subsystem (JES2/JES3), which handles job submission, scheduling, initiator allocation, and spool output.

IN FOCUS

Mainframe Subsystem Support (MSS) & JES Batch Emulation

In traditional IBM z/OS mainframes, batch processing is orchestrated by the Job Entry Subsystem (JES2/JES3), which handles job submission, scheduling, initiator allocation, and spool output. [1]

Read the full briefing

1. Capability Overview

In traditional IBM z/OS mainframes, batch processing is orchestrated by the Job Entry Subsystem (JES2/JES3), which handles job submission, scheduling, initiator allocation, and spool output.

Rocket Enterprise Server provides native JES emulation within its Mainframe Subsystem Support (MSS) architecture. This enables standard mainframe JCL jobstreams—including JOB, EXEC, and DD statements, conditional execution logic (COND, IF/THEN/ELSE), cataloged procedures (PROCs), generation data groups (GDGs), and standard utilities (e.g., IDCAMS, SORT, IEFBR14)—to execute on distributed Linux/UNIX and Windows environments without refactoring them into shell scripts (SEPs for JCL).

2. Architecture & How It Works

  • Service Execution Processes (SEPs) & Class Matching: Rather than executing as unmanaged background operating system processes, batch workloads run inside specialized Batch Initiator SEPs (SEPs for JCL).
    • Each batch initiator is configured with one or more assigned Job Classes (e.g., Class: ABC) (JES Configuration).
    • When a JCL job is submitted to the input queue with a corresponding //... JOB ...,CLASS=A, an idle initiator assigned to Class A dequeues the job, resolves dataset allocations, and executes each job step sequentially.
    • Companion Batch Printer SEPs monitor the output queue to dispatch jobs based on MSGCLASS or SYSOUT routing definitions.
  • Mainframe Dataset Catalog Emulation (catalog.dat & GDGs): Enterprise Server maintains a system catalog that emulates z/OS dataset management. When a JCL step executes with DISP=(NEW,CATLG,DELETE), the dataset is cataloged with its physical storage path, organization (PS, PO, VSAM, DA), and DCB attributes (record format and logical record length) (Catalog List - Enterprise Server).
    • Generation Data Groups (GDGs): Fully supported relative generation references (e.g., (+1), (0), (-1)) automatically roll in new dataset generations at job step or end-of-job boundaries according to standard mainframe rules.
  • Spool Architecture & Automated Housekeeping: All job execution logs, JCL listings, and SYSOUT datasets are stored in the Enterprise Server Spool subsystem (Managing and Viewing Spool Queues). To prevent spool volume growth from consuming host storage, Rocket provides standard housekeeping utilities—such as MVSSPLHK and MVSP0027—which can be scheduled as routine batch jobs to archive, compress, or purge historical spool outputs based on age or retention limits (Spool Housekeeping, MVSSPLHK Spool Housekeeping Process).

3. Practical Configuration & Operational Tidbits

  1. Enabling JES Subsystem in ESCWA: To configure a region for batch processing, open ESCWA (Enterprise Server Common Web Administration), go to Operation > Directory Server > [Region] > General > Properties, and ensure both MSS Enabled and JES Enabled are checked under the Region Features panel (Configuring the JCL Environment for Database File Handling).
  2. Dynamic Initiator Control via Command Line: During peak processing periods (e.g., overnight batch windows), operators can dynamically spawn additional batch initiators without stopping the region using the casout utility: This command starts an initiator handling classes A, B, and C against the ESDEMO region (To start a batch initiator SEP from the command line).
casout -iABC -rESDEMO
  1. Database-Backed Catalogs for Multi-Region Concurrency: While smaller installations use a local disk-based catalog.dat, enterprise and clustered (PAC) environments can host the catalog inside an RDBMS (such as Db2, Oracle, or PostgreSQL) via the Database File Handler (DBFH). Running dbfhdeploy migrates catalog metadata into database tables, eliminating file lock contention when hundreds of batch initiators access datasets concurrently (Configuring the JCL Environment for Database File Handling).
  2. Mainframe-Consistent Filtering: In the ESCWA Catalog interface, checking the option to disable additional qualifiers replicates z/OS catalog search behavior (e.g., searching for A.B returns only dataset A.B, rather than implicitly matching A.B.**) (Catalog List - Enterprise Server).

4. Reference Citations (Rocket Enterprise Server 12.0 Manuals)

Enterprise Server · 12.0 · Linux/UNIX
Documentation references
That’s the briefing.You’ve reached the end of this edition.
More editions

From the archive

View all
Sep 21

XA Distributed Transaction Coordination & Resource Managers (XAR)

Sep 20

External Security Facility (ESF) & Security Manager Stacks in Rocket Enterprise Server

Sep 19

Scale-Out Performance and Availability Clusters (PAC) in Rocket Enterprise Server