When a user data set meets the criteria to be migrated to an SDSP
data set, DFSMShsm selects the SDSP data set with the most free space.
When an SDSP is allocated and opened, it remains open until the DFSMShsm-managed
volume has been completely processed, the SDSP goes unused for 30
seconds, or if the SDSP is needed for recall or aggregate backup.
When a data set fails to migrate because it will not fit into the
selected SDSP, DFSMShsm selects a different SDSP when it migrates
the next data set to an SDSP. You should consider two error conditions:
- If no SDSP is available to a DFSMShsm host (they have all become
full), the data set migrates, but not into an SDSP data set. To avoid
this type of migration, define more SDSP data sets than the maximum
number of concurrent migration tasks. Because each ML1 volume has
no more than one SDSP data set, you must have more ML1 volumes available
than the maximum number of migration tasks.
- Do not run automatic secondary space management while you are
running automatic primary space management.
- If all SDSP data sets are busy, the migration of the data set
fails.