Introduction › How to Convert Data Sets from Multiple to Single Volume › Multiple Volume Analysis Best Practices

Multiple Volume Analysis Best Practices

Multiple volume detail segments are defined as data sets with the same volume table of contents (VTOC) name which are found on different volumes and are diagnosed as parts of the same data set.

For Virtual Storage Access Method (VSAM), this diagnosis means that the data component itself or an index component itself is multiple volume. In the case where a VSAM key-sequenced data set (KSDS) has its data component on one single volume and its index on another volume, it is not considered multiple volume by the storage engine. Each component is single volume.

Observe the following points when analyzing multiple volume data sets:

• In the Multivol Data Sets For System object, the Volume sequence numbers greater than 1 indicate it is part of a multiple volume data set. For example, a value of 3 means it is part 3 of a multiple volume data set.

  The Multivol Data Sets For System object is located in the Diagnostics & Investigations, Data Set Management, Multi Volume Data Sets directory in the Storage object tree in the CA Chorus Investigator. The Seq Nbr field displays the Volume sequence number.

• The Summary of Multvol Data Sets object displays the total number of parts spread across all Volumes in the Vols field. For example, if a data set has parts that are spread across eight volumes, the Vols field shows a value of 8.

  The Summary of Multvol Data Sets object is located in the Diagnostics & Investigations, Data Set Management, Multi Volume Data Sets, Summaries directory in the Storage object tree in the CA Chorus Investigator.

• Multiple volume data set segments form the detail input to the Summary of Multvol Data Sets object.

  This object combines the relevant fields data, such as space allocated, from all the multiple volume segments (parts) that are found on different volumes. For example, if a data set has parts that are spread across eight volumes, this object shows a line which is the total for all eight parts of the data set.

• All parts of a multiple volume data set must exist.

  This point means that all sequence numbers should be found. If a sequence number is missing, something is probably wrong. However, if volumes have been excluded from processing, the excluded volumes can contain the missing portion of a multiple volume data set.

  You can list all the parts, (to see that all the sequence numbers exist), from the Summary of Multvol Data Sets object by selecting the item in the table, then selecting the Navigation action Show Multivol Data Sets for System.

• The VTOC entry for parts 2, 3, and so on, should all identify volume 1.

  When a data set extent is going to another volume (making it a multiple volume data set), DADSM functions record the VOLSER of the first volume within the VTOC entry for all the multiple volume parts. That is, the VTOC field DS1DSSN in parts 2, 3, ... N of the data set all identify the volume where part 1 should be found. By zooming to a dump of the VTOC entry, you can see the DS1DSSN field identifying volume 1 where part 1 should be found.

• In most environments, the catalog should reflect the correct volumes.

  The volumes in an IDCAMS listcat should match the volumes in the sequence number order indicated in the object display. If they do not, something is probably wrong. However, non-SMS data sets do not have to be cataloged, and the non-SMS environment allows VOLSERs to reference an uncataloged data set in the JCL.

• Clean bad multiple volume data sets, but do so cautiously.

  As evident from the preceding points, identifying good from bad multiple volume data sets is not always easy. However, bad entries can easily waste disk space.

• Parts of multiple volume data sets can disappear.

  The parts can disappear for many reasons. The following are some of the common reasons:

  • A full-volume restore can back level a volume to a state before part N of the multiple volume data set being created on it. All parts on other volumes are still found, but the part on the restored volume is now missing.
  • Reinitializing a volume can cause a part to disappear. Historically, multiple volume data sets were not often used or understood. As a result, some vendor and homegrown utilities do not handle them correctly, which can lead to VTOC and catalog errors for these data sets.