IDMS: REORG - getting started

Products

• IDMS
• IDMS - Database

Issue/Introduction

There is a great deal of detailed documentation on the REORG utility. To fully understand what the utility does and how to use it, clients should take the time to carefully read sections CREATE DSMODEL and REORG of the IDMS Utilities manual.

This document contains a very brief description of the utility and provides sample JCL that was used to expand a dictionary DDLDML area from 5000 to 7000 pages.

Environment

Release: All supported releases.

Resolution

The REORG utility unloads and reloads an IDMS database using parallel processing to reduce the amount of time it takes to reorganize data. In order to enable parallel operations, the utility divides the source and target databases into slices and groups of system-owned indexes. Each slice or index group can be processed (unloaded and reloaded) in parallel by concurrently executing jobs.

In this example, the utility is told to divide the DDLDML area into four slices.

This example also makes use of the CREATE DSMODEL utility statement that provides characteristics for dynamic allocation of the intermediate work files that are passed from one step to the next during the REORG.

For this example, the DMCL contains all database file DSN definitions so the only database DD/DSN in this example is the override for the new, larger database in the RELOAD step of the JCL. In addition to the JCL that is submitted, the utility will dynamically submit jobs as needed using JCL that it will read from a dataset specified in the submitted JCL.

Please note that in this case, the UNLOAD and RELOAD steps are run separately, and the override for the new database file must be present in the RELOAD phase only. This avoids the problem described at Unexpected database errors in REORG.

The REORG job can run with the CV down or if CV is up, all affected areas must first be varied to Retrieval or Offline. Keep in mind that the procedure described here is one way to do this. It is the method that was used by IDMS Level 1 tech support to expand one of their dictionary areas.

Sample JCL is provided in the .zip file attached below. Download it and consult the readme.txt file in conjunction with the following steps.

Steps taken to expand APPLDICT.DDLDML from 5000 to 7000 pages using REORG:

1. CREATE DMCL RELDDMCL as the Reload DMCL. This DMCL is an exact copy of the the live DMCL used by CV.
   
   
2. ALTER the AREA and FILE definitions to define the new page and block ranges.
   
   
3. GENERATE DMCL RELDDMCL, punch and link to DBA.LOADLIB.
   
   
4. Allocate the new file and FORMAT using RELDDMCL.
   
   
5. Allocate RORGCTL file for REORG (PS,F,8192,8192). See allocctl.txt. The file size will vary depending on what is being unloaded.
   
   * This is the Control file for the entire REORG process. If running the utility in several steps, this file, initially populated by the SETUP phase, must be available to all steps. A report can be run from this file at any point in the process to indicate the status of the REORG.
   
   * IMPORTANT NOTE: If any problems arise, be sure to retain the RORGCTL file as well as all files created by the utility.
   
   
6. Follow the instructions in the readme.txt to prepare the JCL.
   
   
7. Run the REORG utility as described. In this example, the utility is being run in two jobs. The first job stopped after UNLOAD and the second job stopped after CLEANUP. It is critical to be sure that the unload and all of its submitted parallel jobs have completed successfully before commenting in the DD override of the new database file in common.txt and submitting the reload.
   
   
8. Run whatever programs you would use to verify database integrity.
   
   
9. GENERATE DMCL yourDMCL to bring in the changes to the AREA and FILE affected by the REORG. Rename yourDMCL in the DBA.LOADLIB in case there is any reason to restore back to pre-REORG. Punch and Link the updated yourDMCL to DBA.LOADLIB.
   
   
10. Rename or delete the old DDLDML dataset (typically a ".OLD" extension is added to the DSName to retain the old file temporarily). Rename the new dataset name to match what the CV and DMCL expect.
    
    
11. Bring CV up or DCMT V DMCL NC to bring in the new definitions and again verify that the data in the area can be accessed successfully.

NOTE: Determining sizes for intermediate work files is difficult for a number of reasons and will not be discussed in this document. For more information on this subject, go to the REORG documentation page and scroll down to Sizing work files.