MIGRATIONSUBTASKS: Enabling multiple migration subtasks

Explanation: MIGRATIONSUBTASKS(NO | YES ADDITIONALSUBTASKS(nn)) is an optional parameter that allows DFSMShsm to run multiple subtasks concurrently. Each subtask manages the migration for a single data set, and these run under migration tasks for primary space management, on-demand migration, and interval migration on level 0 volumes that migrate data sets to ML1 or ML2 volumes. Note that, when MIGRATIONSUBTASKS(YES) is specified, this will apply to all migration tasks running in DFSMShsm.

The ADDITIONALSUBTASKS(nn) subparameter allows you to dynamically change the number of additional subtasks that the system can use, running under each migration task. These additional subtasks add to the number of subtasks that the system already uses when the MIGRATIONSUBTASKS parameter is specified. Note that the actual number of total subtasks used can vary. In general, the total migration subtasks used will be lower if a large value is specified with the maximum migration tasks (MAXMIGRATIONTASKS) parameter. Conversely, the total migration subtasks used will be higher, up to 15, if a smaller value is specified with the maximum migration tasks (MAXMIGRATIONTASKS) parameter.

Enablement occurs at DFSMShsm startup, with SETSYS MIGRATIONSUBTASKS(YES) specified (or defaulted to YES if SETSYS MIGRATIONSUBTASKS is specified without the YES/NO keyword). Disablement occurs at DFSMShsm startup with SETSYS MIGRATIONSUBTASKS(NO) specified, or if the command is not specified at all. Additional subtasks can be added or removed dynamically with the ADDITIONALSUBTASKS(nn) subparameter, with no DFSMShsm startup necessary.

You may find that adding additional subtasks is advantageous, based on different possible factors in your system environment. If tape units limit the maximum number of migration tasks, for instance, then you may find that specifying additional migration subtasks increases the amount of concurrency. As another example, you may find that if many data sets being migrated are small, then additional migration subtasks will provide a improvement in overall throughput.

You must weigh such gains, however, against the possibility that specifying additional subtasks may increase the amount of below the line storage used by DFSMShsm.

Note that when specifying MIGRATIONSUBTASKS(YES), USERDATASETSERIALIZATION must also be specified. Specifically, SETSYS USERDATASETSERIALIZATION must be specified in the ARCCMDxx parmlib member, prior to the SETSYS MIGRATIONSUBTASKS(YES) command.

Note, also, that when MIGRATIONSUBTASKS(YES) is specified, concurrent cell pool activity may increase in cell pool 5. It is therefore recommended to review the value specified by the CELLS parameter in the DFHSM startup procedure for CELL POOL 5. See "Specifying the size of cell pools" and "Adjusting the size of cell pools" in the z/OS DFSMShsm Implementation and Customization Guide for more information.

SMS relationship: Parameter has the same meaning when applied to SMS-managed or non-SMS-managed DASD volumes or data sets.

SETSYS default: None.

DFSMShsm default: If you do not specify this parameter on any SETSYS command, the DFSMShsm default is to not use migration subtasks. When MIGRATIONSUBTASKS is specified without the YES or NO keyword, the default is YES. When the ADDITIONALSUBTASKS option is not specified, DFSMShsm will not use any additional subtasks.