Sizing the ACQDWK03 file - syncsort_capacity_management - 12 - 12.40

Syncsort Capacity Management Installation Guide

Product type
Software
Portfolio
Integrate
Product family
Syncsort
Product
Syncsort Capacity Management
Version
12.40
Language
English
Product name
Syncsort Capacity Management
Title
Syncsort Capacity Management Installation Guide
Topic type
How Do I
Copyright
2023
First publish date
1985

If you are only collecting volume or storage group data, the size of the ACQDWK03 file will be fairly small. A useful estimate is 1 cylinder of space for each 3,000 disks online to your system.

If you choose to generate dataset groups the size of the ACQDWK03 file will need to be considerably larger than if you are just collecting volume or storage group information.

There are two kinds of dataset records generated by DCOLLECT.

  • Type “D” records (for all datasets) are each 350 bytes long

  • Type “A” records (for active VSAM dataset components) are each 208 characters long

 

Note that a type “A” record is generated for all components of a VSAM dataset - the data portion, the index portion, the alternate index portion etc. - so you may have many records for a single VSAM entity.

For a system with 100,000 datasets, 20,000 of which are VSAM components, you will need an allocation of about:

(100,000 * 350) + (20,000 * 208) = 39 MB or about 50 cylinders

These values are for a fairly modest size of system - most will have much larger numbers of datasets, and this is why experimenting with Acquire/DASD in single shot mode is recommended before running in cyclic mode.

In single shot mode, you run DCOLLECT yourself and so you can see how big an input file is generated for Acquire/DASD, and also make sure your groupings work as you expect in the resulting output data.

When you run Acquire/DASD in cyclic mode, and are collecting dataset groups, you may want to set MSGLEVEL=(2,0) to suppress system messages for the job or started task.

Without it, the system generates many allocation/deallocation messages which can lead to a large JES2 spool occupancy if Acquire/DASD is run over a long period of time.