Allocating database data sets

The data set names are created from the data set name prefix that you specify when you define a partition and the data set name algorithm. You must use these names when you allocate your data sets.

Allocate the database data sets for each partition, including the indirect list data set (ILDS).

The information that is generated by the HALDB Migration Aid utility (DFSMAID0) can assist you in determining how large to make the data sets. The utility reports the number of bytes used by the existing segment data in each partition. You will need to add additional bytes to allow space for segments to be added, for the free space you want to include, and for bitmaps.

When migrating to HALDB, you might want to increase your free space parameters. Non-HALDB databases sometimes have suboptimal free space values due to data set size limitations. Additional free space allows you to reorganize your databases or partitions less frequently. In some cases, additional free space can eliminate the need for reorganizations altogether.

Although databases that do not have logical relationships or secondary indexes do not use the ILDS, in some cases HALDB requires that an ILDS exist in each partition anyway. An online IMS system does not allocate an ILDS if the database does not require one; however, batch jobs do allocate an ILDS, so you must define an ILDS. IMS issues message IEC161I 152-061 when an empty ILDS is opened, but this message does not indicate a problem.

When no secondary indexes or logical relationships are defined, the HD Reorganization Reload utility (DFSURGL0) does not update the ILDS.

If you do not use the ILDS, you can allocate a very small amount of space for the data set, such as one track.

All ILDSs have fixed-length 50-byte records. Keys are 9 bytes at zero offset. The following example shows IDCAMS DEFINE statements to allocate an ILDS.

   DEFINE CLUSTER(                    -     
      NAME(JOUKO3.HALDB.DB.PEOPLE.L00001) - 
      INDEXED                         -     
      CYL(1 1)                        -     
      RECORDSIZE(50 50)               -     
      SHAREOPTIONS(3 3)               -     
      SPEED                           -     
      KEY(9,0)                        -     
      FREESPACE(10,10)                -     
      CONTROLINTERVALSIZE(8192)       -     
      VOLUMES(TOTIMN)                 -     
   )

You must specify REUSE on the DEFINE statement for all HALDB VSAM data sets.

The output of the DBDGEN utility is useful when you allocate PHIDAM primary indexes. The output of the DBDGEN utility for the PHIDAM database lists the required parameters for the IDCAMS definition, as shown following the label RECOMMENDED VSAM DEFINE CLUSTER PARAMETERS. The required parameters are the INDEXED parameter, the values of the RECORDSIZE parameter, the REUSE parameter, and the values of the KEY parameter. The following example shows the IDCAMS DEFINE statement to allocate a PHIDAM primary index data set.

   DEFINE CLUSTER(                    -     
      NAME(JOUKO3.HALDB.DB.RZL.X00001) - 
      INDEXED                         -     
      CYL(10 5)                       -     
      RECORDSIZE(20 20)               -     
      SHAREOPTIONS(3 3)               -     
      REUSE                           -     
      KEY(14,5)                       -     
      FREESPACE(25,10)                -     
      CONTROLINTERVALSIZE(8192)       -     
      VOLUMES(TOTIMN)                 -     
   )