A NAS gateway environment has a NAS controller (or NAS head) that is connected to a storage area network (SAN). The SAN is usually a Fibre Channel (FC) infrastructure, and a block storage system will provide the storage resource for the NAS controller, as well as being used for other block storage purposes.
Migrating data from one NAS gateway system to another NAS gateway system is the same as migrating between NAS systems. . .with one exception. Because the gateway will utilize storage systems with replication features providing the storage resources, those storage system features can be used to copy the data from one storage system to another without going through the NAS gateways.
There are four options for migrating data between the two NAS gateway systems:
- A server-based program can copy data from one NAS system to the other. This approach has potential data integrity issues. For example, if accesses are allowed during the copy process, data may not get reflected on the target system when updated in a location after it had been already copied. To assure integrity, the NAS file systems must be taken offline during this copy, which may affect availability requirements.
- Some NAS systems have additional software (usually costing extra) that can migrate data from a donor NAS system while maintaining availability. This is usually accomplished by starting the software on the target NAS system and then directing all accesses (usually through the mount for the file system) to that NAS system. The migration starts, and if an access for data that has not been migrated occurs, the data is pulled from the donor system to honor the request. This method ensures availability but may not provide a fallback condition if the updates that occurred during the migration were not applied to both the target and donor systems. This may be a concern for some if the migration is interrupted for some reason.
- As part of a gateway system, the block storage LUNs used for the file system may be migrated from one storage system to another using the storage network. This option is complex and the controls for access during the migration may require coordinating software between the block storage systems and the NAS gateways. This may be better served as an offline operation rather than dealing with the complexity of coordination. If it is done offline, it will not meet the availability needs.
- An external device may be put in front of the two NAS systems to serve as another level of intelligence that can migrate the data and direct the accesses during the migration. Whether this system (sometimes called a network file virtualizer) can perform simultaneous updates to provide a fallback condition in case the migration is interrupted will depend on the vendor implementation.
About the author: Randy Kerns is an independent storage consultant focusing on storage and storage management software, including SAN and NAS analysis. In the past, he served as vice president of strategy and planning for storage at Sun Microsystems Inc., .