Data striping considerations when preventing abends with Allocation

Sequential data striping was introduced with DFSMS version 1.1 and was originally limited to QSAM and BSAM files.

Beginning with DFSMS version 1.2, VSAM KSDS files can also benefit from the data compression that data striping provides.

Sequential data striping extends sequential data sets over multiple volumes, preferably on different channel paths and control units to improve performance. Beginning with DFSMS 1.2, a change was introduced that allowed single stripe, permanent data sets to extend to new volumes after initial allocation. The number of stripes can be controlled by the DFSMS Storage Group constructs that you specify for the data set. Refer to the IBM Storage Management Library documentation for more details. Another benefit of extended sequential data sets is that you can have 123 extents per volume.

Temporary data sets can be allocated as extended-format data sets, but they cannot be extended to multiple volumes; otherwise, they are not allocated as striped. Refer to APAR OW03815 for further details.

When extended-format data sets were first introduced, these data sets had to be allocated on volumes that were attached to the 3990-3 Storage Control Unit with the extended platform and had to be connected to the host by way of ESCON channels. For DFSMS 1.2, DFSMS/MVS SAM/VSAM hardware compression requires extended-format data sets, but IBM has made modifications to the Licensed Internal Code (LIC) of several direct-access storage subsystems so that ESCON attachment is no longer a requirement for DFSMS/MVS SAM/VSAM hardware compression.

Note

DFSMS version 1.1 still has the ESCON requirement. Refer to IBMLINK item RTA000031564 for details.

Several variations of the extended-format (striped) data sets now exist as follows:

  • SAM Single Stripe
  • SAM Multistripe
  • VSAM Single Stripe
  • VSAM Multistripe

The following table summarizes the functions and the processing that is performed on each type of extended-format data set.

Functions processing methods

Function

Processing method

SAM

Primary space reduction (SPACPRIM)

  • Single Stripe—The primary allocation can be reduced to allow the allocation to remain on the current volume. This process can be done even if it drives utilization above the volume’s high-water mark.
  • Multistripe—This process works the same as for single stripe data sets. The allocation is divided among the number of stripes. When you use ISPF option 3.4, the catalog entry shows the total allocations for all stripes. You can specify the data set name and volume for the specific allocations to a stripe.

Note: Allocation reduces the primary allocation before the stripe count is changed.

VSAM

Primary space reduction (SPACPRIM)

  • Single Stripe— reduces the primary allocation of the data set even if it forces the volume utilization above the high water mark.
  • Multistripe—Multistripe VSAM data sets are not currently allowed. This situation is an IBM limitation.

VSAM files are limited to 123 extents per volume.

SAM

Secondary space reduction (SPACSECR)

  • Single Stripe—With SPACSECR active for this recovery attempt, reduces the secondary allocation as it did for nonextended files. If SPACSECR is not active for this recovery attempt, forces SPACVOLA processing.
  • Multistripe—With SPACSECR active for this recovery attempt, reduces the secondary allocation as it did for nonextended files. If SPACSECR is not active for this recovery attempt, issues a message, and the job fails with an X37 abend.

VSAM

Secondary space reduction (SPACSECR)

  • Single Stripe—With SPACSECR active for this recovery attempt, reduces the secondary allocation and continues processing. If SPACSECR is not active for this recovery attempt, attempts SPACVOLA processing.
  • Multistripe—With SPACSECR active for this recovery attempt, reduces the secondary allocation and continues processing. If SPACSECR is not active for this recovery attempt, attempts SPACVOLA processing.

SAM

Volume add (SPACVOLA)

  • Single Stripe—No change in functionality has occurred.
  • Multistripe—This type of data set cannot include volumes that were added with SPACVOLA processing.

VSAM

Volume add (SPACVOLA)

  • Single Stripe—No change in functionality has occurred.
  • Multistripe—With SPACVOLA active for this recovery attempt, can add one (when global MULTISTRIPE=NO) or more (when global MULTISTRIPE=YES) volumes to the data set to prevent the extend from failing. With SPACVOLA not active, no further recovery is attempted.

SAM

Add secondary (SPACSECA)

  • Single Stripe—No change in functionality has occurred.
  • Multistripe—No change in functionality has occurred.

VSAM

Add secondary (SPACSECA)

  • Single Stripe—No change in functionality has occurred.
  • Multistripe—Adding a secondary allocation amount does not apply to Multistripe VSAM data sets because the primary allocation amount is used by IBM during extents.

 

Tip: For faster searching, add an asterisk to the end of your partial query. Example: cert*