Managing a file system with multiple IBM Storage Protect servers

Overview

A single IBM® Storage Scale system can be migrated to two or more IBM Storage® Protect servers. Two or more IBM Storage Protect servers are required for file systems that contain more objects than can be managed by a single server. As the file system grows beyond the capacity of existing servers, you can add more servers to manage the file system.

A file system that is managed by multiple IBM Storage Protect servers must be GPFS™ version 3.4 or later and must be enabled for Data Management Application Programming Interface (DMAPI).

A file in a multiple-server environment is coupled with the IBM Storage Protect server to which it is initially migrated or backed up. The coupling between the file and the server is persistent over the life of both the file and the IBM Storage Protect server, and ensures data consistency. The file cannot be migrated or backed up to another IBM Storage Protect server until the file is uncoupled. Automatic migration and transparent recall processes contact the server with which a file is coupled.

If a file has not been coupled with a server, you can specify the server when you do a selective migration of the file.

Automatic migration uses the GPFS policy engine to determine which IBM Storage Protect server to use for the initial migration. By default, the space management client distributes files to all servers in a round-robin fashion. You can customize the distribution scheme to suit your environment.

A single space management client can use only HSM daemons (dsmmonitord and dsmscoutd) or IBM Storage Scale policy to manage automatic migration. A space management client cannot use both HSM daemons and IBM Storage Scale policy to manage automatic migration.