Evaluate your stand-alone dump data set allocations and your IPCS processing of them
Description
As your applications grow in size and use greater amounts of storage, you should evaluate whether the DASD allocated for your stand-alone dump data continues to be adequate.
In z/OS V1R6, support was introduced for extended-format sequential data sets, a form of data set that is SMS-managed and can occupy more than 64 K tracks per volume. In z/OS V1R7, this support was supplemented with support for large format sequential data sets (DSNTYPE=LARGE), a form of data set that is essentially the same as conventional sequential data sets except that more than 64 K tracks may be spanned per volume. If your stand-alone dump data sets are spread over more volumes than you want, both types of support can help you gain better control over the number of volumes used for each stand-alone dump data set.
Table 1 provides more details about the migration action. Use this information to plan your changes to the system.
| Element or feature: | BCP. |
|---|---|
| When change was introduced: | General migration action not tied to a specific release. |
| Applies to migration from: | z/OS V2R1 and z/OS V1R13. |
| Timing: | Before installing z/OS V2R2. |
| Is the migration action required? | No, but recommended because of changes that have been made to stand-alone dump processing (that reorder dump records with the intent of recording more important data early), and especially recommended if you deploy any LPARs with significantly more main storage than previously used. |
| Target system hardware requirements: | None. |
| Target system software requirements: | None. |
| Other system (coexistence or fallback) requirements: | None. |
| Restrictions: | None. |
| System impacts: | None. |
| Related IBM® Health Checker for z/OS® check: | None. |
Steps to take
- Use multivolume stand-alone dump data sets. Adjust the number of volumes and their separation to achieve tolerable stand-alone dump capture times.
- Use extended-format sequential data sets or large format sequential data sets. Copy their contents to an extended-format, compressed, striped data set using the IPCS COPYDUMP subcommand before analysis. Use the same or a larger striping factor than you used for your stand-alone dump data sets. Dump data sets to which stand-alone dump can write may be neither compressed nor striped, but both attributes are advantageous for the target of the copy operation. Starting with z/OS V1R12, stand-alone dump data sets can be placed in track-managed space as well as cylinder-managed space on Extended Address Volumes (EAV).
- Use a large CISIZE and striping for IPCS dump directories, and
use blocking, striping, and compression for the stand-alone dump
data set. Very large stand-alone dumps might require that you
define your directory with the extended addressing attribute, allowing
it to hold more than 4 GB. Tip: Control interval sizes less than 24K have been shown to be more vulnerable to fragmentation when used as IPCS dump directories, and IPCS performance can be degraded when such fragmentation occurs. In this background, warning message BLS21110I will be issued and you might recreate the DDIR by using the CLIST BLSCDDIR.
BLS21110I CISIZE(cisize) is less than 24K. It may degrade IPCS performance
Reference information
- For information about dump data set allocation, extended format sequential data sets, large format sequential data sets, and multivolume dump data sets, see z/OS MVS Diagnosis: Tools and Service Aids.
- For stand-alone dump procedures, see z/OS V2R2 Problem Management.