Recall volume selection

When DFSMShsm receives a non-directed recall request (either by command or by allocation with no volume specified), it builds a list of up to five primary volumes with the most available space. The space on the volumes is based on the last space check performed on the primary volumes. DFSMShsm selects the primary volumes as follows:

  1. If the data set being recalled has a high-level qualifier that matches that of a data set pool, the five volumes come from the data set pool.
  2. If the volume from which the data set migrated is part of a volume pool, the candidate volumes come from the volume pool.
  3. If the data set is not restricted to be recalled to a user-defined pool, DFSMShsm selects volumes that were assigned to the default pool. Whether a volume is selected is based on:
    • The subparameters selected for the RECALL parameter of the SETSYS command
    • The volumes having the AUTORECALL subparameter specified in the ADDVOL command and having MIGRATE as the space management technique
    • The volumes being online
    • The use attribute of the volume
      • Attribute of storage for the RECALL(ANYSTORAGEVOLUME) parameter
      • Attribute of storage, public, or private for the RECALL(PRIVATEVOLUME) parameter

In building the list of candidate volumes from the default pool, DFSMShsm considers the recall attributes described in Recall attributes and the LIKE|UNLIKE subparameters specified with the RECALL parameter of the SETSYS command. If LIKE is specified, DFSMShsm builds the list only from volumes with recall attributes that match the recall attributes of the volume from which the data set migrated. If fewer than five volumes have like attributes, the list of candidate volumes contains fewer than five volumes.

If UNLIKE is specified, DFSMShsm builds the list from all the volumes in the default pool in the following order:

  1. If five or more volumes exist with like recall attributes, all the volumes in the candidate list have recall attributes that match those of the volume from which the data set migrated.
  2. If not enough like volumes exist, DFSMShsm selects as many volumes with recall attributes that match those of the volume from which the data set migrated as possible, then selects unlike volumes in the following order:
    1. Volumes with matching backup attributes (AUTOBACKUP|NOAUTOBACKUP and BACKUPDEVICECATEGORY)
    2. Volumes with matching migration attributes (AUTOMIGRATION|NOAUTOMIGRATION)
    3. Volumes with no matching attributes

    Within this group, each subgroup is placed in order by the amount of available space on the volume. Thus, the volumes with like attributes are placed (in order by free space) ahead of the volumes with unlike attributes.

DFSMShsm never requires matching recall volume attributes for data sets under the following conditions:

A primary volume managed by DELETEIFBACKEDUP or DELETEBYAGE space management techniques will never be selected by DFSMShsm as the target of a nondirected recall.