Data Migration Methodologies and Approach: Firstly, it is important to differentiate between data migration and other data management strategies such as backup, snapshotting, cloning, archiving, or Hierarchical Storage Management (HSM). Data migration is usually defined as the one-time movement of data between source systems and target systems. Once migrated, source data is not normally used, and is often deleted or archived. Data migration can be accomplished at several different levels. Levels of Data Migration can include:
- Disk or volume level
- LUN level
- Filesystem Level
- File Level
- Disk Array controller or Appliance level
- SAN or Network level
- Application level
Migration between filesystems, ZFS to/from VxFS, for example, can be complex, and requires a good understanding of the following:
- Available migration tools.
- Source and Target Filesystem type and version.
- Source and Target data consumer Operating System (OS) type, version, and patch levels.
- Application level data management impact, for example Oracle database considerations.
- Performance, Availability, Security, and Reliability Impacts on Applications using filesystems during, and post migration.
- Risk mitigation and problem recovery options.
Commonly used migration tools range from the simple to quite complex, and may include:
- Commercial Data Migration Software, including HSM or Backup based software, such as Veritas NetBackup Storage Migrator, CA Brightstor, EMC DiskExender. Array or Appliance Based Migration Tools. These are vendor specific, and are typically used for LUN level migration.
- Application based utilities, for example Oracle's RMAN, ASM, DataPump export/import.
- Filesystem based commands and utilities, such as dump and restore variants, for example vxdump, fscdsconv, zfs snapshot send/receive.
- Storage Pool migration commands and utilities, such as zfs zpool export / import. Volume Manager commands and utilities, such as VxVM deport / import. OS based commands/utilities, for example XXCOPY, sync, cp/dd/cpio, export/import.
Most of these tools are vendor, OS, or filesystem specific, and each will have specific hardware and software dependencies, as well as individual performance and availability implications. Deciding which data migration tool is most appropriate and formulation of a migration strategy and project plan is not a simple task. It requires foremost an understanding of the business and application requirements related to the end user data. In all but the simplest of environments, expertise and experience in data migration and knowing the pros and cons of the selected methodologies are vital for a successful data migration.