Setting the size of active log data sets

You can modify DSNTIJIN installation job to change the size of your active log data set.

To choose the most effective size for your active log data set:

• When you calculate the size of the active log data set, identify the longest unit of work in your application programs.
  For example, if a batch application program commits only once every 20 minutes, the active log data set should be twice as large as the update information that is produced during this period by all of the application programs that are running.
• Allow time for possible operator interventions, I/O errors, and tape drive shortages if offloading to tape.
  Db2 supports up to 20 tape volumes for a single archive log data set. If your archive log data sets are under the control of DFSMShsm, also consider the Hierarchical Storage Manager recall time, if the data set has been migrated by Hierarchical Storage Manager.
• When archiving to disk, set the primary space quantity and block size for the archive log data set so that you can offload the active log data set without forcing the use of secondary extents in the archive log data set.
  This action avoids space abends when writing the archive log data set.
• Make the number of records for the active log be divisible by the blocking factor of the archive log (disk or tape).
  Db2 always writes complete blocks when it creates the archive log copy of the active log data set. If you make the archive log blocking factor evenly divisible into the number of active log records, Db2 does not have to pad the archive log data set with nulls to fill the block. This action can prevent REPRO errors if you should ever have to REPRO the archive log back into the active log data set, such as during disaster recovery.

  To determine the blocking factor of the archive log, divide the value specified on the BLOCK SIZE field of installation panel DSNTIPA by 4096 (that is, BLOCK SIZE / 4096). Then modify the DSNTIJIN installation job so that the number of records in the DEFINE CLUSTER field for the active log data set is a multiple of the blocking factor.

• If you offload to tape, consider adjusting the size of each of your active log data sets to contain the same amount of space as can be stored on a nearly full tape volume.
  Doing so minimizes tape handling and volume mounts and maximizes the use of the tape resource.

  If you change the size of your active log data set to fit on one tape volume, remember that the bootstrap data set (BSDS) is copied to the tape volume along with the copy of the active log data set. Therefore, decrease the size of your active log data set to offset the space that is required on the archive tape for the BSDS.