Before you can use your existing data in SAN File System, you must first migrate it to the SAN File System global namespace.
You must perform data migration from the client machine. For example, data from a Windows-based application must be migrated from a Windows-based client. This also applies to UNIX®, AIX®, Solaris, and Linux™.
You can use standard copy commands or utilities that are provided by the client operating system to migrate your data (for example, cp, cpio, or tar commands on UNIX; and the xcopy command or Explorer on Windows®). You can also use backup applications to restore data from the latest backup into the SAN File System as the destination. These methods work best when migrating large numbers of small files.
The migratedata utility attempts to preserve all hard links to regular files. For UNIX-based clients, if SAN File System is set up in such a way that hard links traverse across a fileset, the hard links are replaced with a symbolic link.
You can use the migratedata utility to migrate files that use file-system-based compression (as with NTFS) or file-based compression (for example, files produced by utilities such as compress on AIX systems and ZIP utilities on systems running the Microsoft® Windows operating system). Note that during the migration process, all compressed files are expanded.
Migrating data using the data migration utility is a disruptive process. This means that, to guarantee data integrity, you must stop all applications and users from modifying the data being migrated (including database and application servers) until the migration is complete. Only the data being migrated must remain unchanged. To minimize the impact of a migration, you can migrate your data in chunks rather than all at one time.
If you are migrating a database environment, the procedures vary depending on whether your environment is file-system based or contains raw configuration devices. For a file-system-based environment, use the migratedata utility to migrate your files, and then reconfigure the database application to reflect the data movement. For raw configuration devices, perform an unload command (such as the dd command) to move data out of the raw devices to a temporary holding location, and then perform a load operation to place the data in its location in the global namespace. After the data is loaded into the global namespace, it is file based.
Data-migration considerations
This topic describes considerations for data migration.
Parent topic: Concepts
Related tasks
Migrating data
Related reference
Data-migration considerations