Configuring the space management client

Before you configure space management for a file system, you must configure the space management client itself. You also must configure some options that apply to all space-managed file systems.

The space management client is configured by setting options in the dsm.sys and dsm.opt files. The options generally affect all space-managed file systems, and determine the following things:

• The IBM Storage® Protect servers to which your files migrate or premigrate.
• The include-exclude file that applies for each IBM Storage Protect server.
• How often space usage is checked on your file systems.
• How often your file systems are automatically reconciled.
• How often candidates are searched for automatic migrations.
• How many automatic migration processes for each file system can migrate files in parallel.
• How many days to keep an obsolete file copy in IBM Storage Protect storage before the copy expires.
• Whether to use two-way orphan check reconciliation (for IBM® Storage Scale only).

During installation of the IBM Storage Protect for AIX® and Linux™ backup-archive clients, sample options files (dsm.sys.smp and dsm.opt.smp) are placed in the following directories.

• On AIX: /usr/tivoli/tsm/client/ba/bin64
• On Linux: /opt/tivoli/tsm/client/ba/bin

If you install both the backup-archive client and the space management client at the same time, copy and rename the sample options files. Modify them for both clients. If you previously installed the backup-archive client and you set up your options files, modify them for the HSM client.

After you select options for the space management client, you must restart all space management daemons to activate the changes. All space-managed file systems must be mounted either automatically or manually each time you restart your system to start space management.

The space management client shares the following common files and code with the AIX and Linux backup-archive clients:

• Communication protocols
• The dsm.opt options file
• The dsm.sys options file
• The include-exclude file
• Node registration
• Server file spaces

Data deduplication of HSM data

If client- or server-side data deduplication is used with HSM data, ensure that there is more than a single copy of the HSM-migrated file on the IBM Storage Protect server.

When a deduplicated data chunk is damaged on the IBM Storage Protect server, and the deduplicated chunk is part of a migrated file, you cannot recall the file because the migrated file is the only copy on the IBM Storage Protect server. On the other hand, if you previously backed up the migrated file, a backup copy of the file also exists on the IBM Storage Protect server. In this situation, two identical versions of the original file exist on the server: a backup copy and an HSM-migrated copy. When data deduplication is enabled, the IBM Storage Protect server removes all duplicate content, and only one copy of the file remains on the server. If that copy is ever lost or damaged, the original file cannot be recovered.

To prevent the potential loss of data, use one of the following methods:

• Place either the HSM data or backup data for a file space in a nondeduplication pool.
• Store the HSM and backup data on different deduplication pools.