Replication and disaster recovery

Replication and Disaster Recovery Administration Guide for Hitachi NAS Platform

Version
15.1.x
Audience
anonymous
Part Number
MK-92HNAS009-29

Object replication provides replication of file systems, including the replication of related access points, such as SMB shares and NFS exports, as well as tools to automate disaster recovery.

Object replication has several important concepts:

  • Primary file system: The "primary file system" is the file system that network clients access. The primary file system is the "live" file system and the source of the replication.
  • Direction of replication: When the primary file system from site "A" is replicated to another server at site "B," the direction of the replication determines which file system is designated the "source" and the "target." The primary file system from site "A" is always the replication source. The target file system is the replica on the server at site "B" (which may located at the same physical site as "A" or at a remote location). Note that the target file system at site "B" may also be used as a replication source to a third site (site "C").
  • Swapping roles: Moving the network client access from the file system at site "A" to the replicated file system at site "B." Roles are swapped when the replication either stops (for example, if site "A" goes offline for some reason) or the direction of the replication is intentionally reversed (a planned role swap).
  • During the role swap, the file system at site "B" is promoted to primary, and the file system at site "A" is demoted. (If site "A" is accessible, the file system at site "A" typically becomes the target file system.)
  • As a part of the role swap, access point (SMB share or NFS export) settings are deleted from the server at site "A" and, along with other configuration settings, are applied to the server at site "B" so that network clients now access the primary file system, which is now physically located at site "B." Clients accessing the file system now communicate with the server at site "B" and read from and write to the file system at that site, which has become the primary file system.
  • Note that a single server can host many file systems, and could be providing "primary" access to several file systems while other file systems hosted by the same server could be target file systems. Primary access for any file system can be moved independently of any other file system on the same server.

Object replication is most often used in the following situations:

  • A planned promotion of the file system at site "B" to primary. In this case, it is possible to ensure that the file systems at sites "A" and "B" are exact replicas (though this would require a period of read-only access at site "A"). If both sites are functional, it may be possible for the server at site "B" to access the server at site "A" to retrieve information such as configuration settings.

    In the case of a planned promotion, the administrator puts the primary file system into syslocked-mode, then schedules an incremental replication to the target to ensure both file systems are synchronized. Once the replication is complete, the promotion can proceed, and after the transfer of primary access, the clients access the newly promoted primary file system.

  • An unplanned promotion of the file system at site "B" to primary (also known as "disaster recovery". If, for any reason, the primary file system at site "A" becomes inaccessible, the file system at site "B" is promoted (becomes primary). In this case it is unlikely that the file system at site "B" will be an exact copy of the file system at A at the time of the outage (because the replication is asynchronous). Server B must already have access to all the information necessary to function as the primary.