High availability

Policy-based high availability provides an active-active high availability (HA) solution using two independent systems in different locations. The storage will automatically remain accessible to hosts if there is an event that makes one of the systems unavailable.

Storage partitions are used to implement policy-based high availability and disaster recovery solutions. Partitions contain volumes, volume groups, hosts, and host-to-volume mappings. For more information, see Storage partitions.

When an HA replication policy is associated with a partition, all the objects contained in the partition are automatically configured across both systems that are defined in the replication policy. Hosts that are created in a partition will be able to discover paths to the mapped volumes through both systems. Hosts that are configured with a location prioritize the use of a co-located system for I/O.

Using policy-based high availability, everything in a storage partition is highly available. When required, both management and I/O failover between systems to maintain access.

Each partition that is associated with an HA replication policy has two properties:
  • preferred management system
  • active management system

The active management system is the system from which all configuration tasks must be performed. In the event of a failure that affects the current active management system, the active management system role will automatically failover to the other system in the partnership.

The preferred management system is the system that you would like to be the active management system under ideal conditions and is the system that is preferred to keep the partition accessible if the two HA systems are disconnected. The preferred management system can be changed from the current active management system. Different partitions can be configured with different preferred management systems.

If the active management system and the preferred management system are not the same system, the system will automatically failover the active management system role to the preferred management system when it is able to.

As part of the high availability solution, an IP quorum application must be deployed on a server that is separate from the storage systems. The IP quorum application is used to arbitrate between systems in a partnership, to determine which system will operate as the active management system in the event of a failure.