This chapter describes the automatic migration function and cloud usage.

The automatic migration is performed as part of primary space management, interval, or on-demand migration tasks. The ability to automatically select data that is eligible for cloud migration is vital to maximize central processing unit (CPU) savings that are related to data migration in DFSMShsm.

This chapter describes the following topics:

To enable the DFSMShsm automatic migration to cloud, it is necessary to define the policies that define one of the following targets for migration of a data set:

The definitions for data set migration are included in the SMS Management Class construct. Therefore, new fields are included to allow storage administrators to define conditions that are required for a data set to be eligible for cloud migration.

The data sets cannot be automatically migrated from ML2 to cloud the same way that Migration Level 1 (ML1) data sets can be migrated to ML2. So, the management class was updated to create the rules for deciding the migration level tier.

The following fields are now available on Management Class under the Migration Attributes panels:

Specifies a direct migration to cloud storage. Current possible values are 0 and NOLIMIT.

A blank value also indicates NOLIMIT. A value of NOLIMIT indicates that a data set migrates to Level 2 based on the value of Level 1 Days Non-usage.

A value of 0 for this attribute indicates that the data set migrates to the cloud storage specified by the Cloud Name field if the Primary Days Non-usage value is met and the data set still is on Level 0.

The Level 2 Days Non-usage processing takes precedence over the Level 1 Days Non-usage value. The default value is NOLIMIT.

The “Size Less than or equal” field shows the low data set size threshold in tracks. It is used to take the action described in the Action LTE field.

The value in the Action LTE column shows which action to perform if the data set size is less than or equal to Size LTE. The following values are possible:

If no value is specified, DFSMShsm performs the migration action the same way it is performed today.

The “Size Greater than” field shows the high data set size threshold in tracks. It is used to take the action described on “Action GT” field.

The value in the Action GT column shows which action to perform if the data set size is greater than Size GT. The following values are possible:

If no value is specified, DFSMShsm performs the migration action the same way it is performed today.

The value in the CLOUD NAME shows the name of a previously defined cloud construct for the data set migration to the cloud during automatic migration (Primary Space Management, Interval migration, and On-Demand migration).

The new fields work in a way that is similar to the existing DFSMShsm data set migration exit (MD), which controls the migration level for data sets selected by automatic migration processing. You can configure these fields by using these values in the exit, and you can disable the exit.

EXITON is an optional SETSYS command parameter specifying active installation exits in the DFSMShsm primary address space. For modname, substitute the module name of the installation exit that you want to be active.

For more information about exiting, see the following website:

The use of cloud storage by DFSMShsm can affect the length of your space management windows, as your migration to the cloud can take longer than ML1/ML2 migrations depending on your network usage, latency, and cloud performance.

For some large systems, with several storage groups and DFSMShsm images running, storage administrators might decide to spread the workload between these Hierarchical Storage Manager (HSM) images. There are enhancements to the storage group affinity to allow the selection of storage groups management for specific HSM hosts.

To enable the storage group affinity, you can use the new SETSYS command to identify the storage groups that are managed by an HSM host. Example 10-1 shows a sample SETSYS command to create the affinity between storage group BATCH1 and the HSMIMG2 DFSMShsm image.

By setting up the storage group affinity, the specific HSM usage manages the defined storage groups. Define this command in your ARCCMD parmlib member to save your settings across initial program loads (IPLs).

The SETSYS MIGRATIONAUTOCLOUD parameter enables the host to run cloud migration, and is an optional parameter that defines the possible target tier during automatic migration. The default value of this parameter is NOCLOUD:

This section describes the space management functions for managing SMS-managed storage, which is updated with cloud information:

During automatic primary space management, the following events occur:

Primary space management continues until the SMS-managed volumes have the specified amount of available space. However, if deletion of expired data sets, fast subsequent migration, and space reduction of the remaining data sets achieves the specified volume available space low threshold, no actual data sets are moved.

If data sets are migrated during primary space management, they are migrated in compacted form if possible. Data sets that are expected to be smaller than 110 KB (where 1 KB is 1024 bytes) after they are compacted are candidates for small data set packing (SDSP) if enabled.

Automatic secondary space management schedules TAPECOPY commands for migration tape copy needed (TCN) records, deletes expired data sets from the migration volumes, deletes obsolete MCDs, VSRs, and DSRs during migration cleanup, and moves data sets (under control of the management class) from ML1 to ML2 volumes.

On the first day of the secondary space management cycle, DFSMShsm examines the containers in the clouds that are defined to DFSMShsm, looking for containers that do not have objects. Any empty containers are deleted. If the end of the Secondary Space Management window is reached and not all containers are examined, the processing continues in the next Secondary Space Management window.

When DFSMShsm processes an SMS-managed volume for space management, it records the time at which the volume was processed for cloud or non-cloud migrations. If another DFSMShsm host attempts to process the same volume for space management to cloud or non-cloud migrations, it checks the time at which the volume was last processed. If the volume was processed within the last 14 hours, DFSMShsm does not process the volume again for that type of target. A volume can be processed by one host for non-cloud migrations, and later processed by another host for cloud migrations.

Under Tasks for automatic space management, Specifying the DFSMShsm hosts that process each storage group, see 10.2, “Storage Group affinity enhancements” on page 106.