Replication policies

A replication policy defines how replication is configured between systems.

A replication policy is defined between two systems in a partnership that supports policy-based replication. The partnership must be fully-configured and each system must be connected. Each replication policy identifies a single I/O group on each system. To replicate volume groups from multiple I/O groups, you need to create another replication policy. You can use the management GUI interface and command-line interface (CLI) for creating replication policies.

Policies that use asynchronous replication must also specify an RPO (recovery point objective) alert. The system always tries to achieve the lowest possible recovery point. However, if the recovery point increases beyond this value an alert is raised. If the RPO exceeds that value, an alert is sent.

A replication policy defines three key attributes:

A set of locations
Defines the I/O groups on the partnered systems that contain a replicated copy of the volume group or storage partition. The location defines where data is replicated.
A topology
Represents organization of the systems and the type of replication that is completed between each location. The topology defines how data is replicated between the locations.
A recovery point objective (RPO)(for asynchronous replication topologies)
Defines a maximum acceptable RPO for the asynchronous replication between locations.
The following rules apply to replication policies:
  • An asynchronous replication policy can be assigned to one or more volume groups.
  • A high availability replication policy can be assigned to one or more storage partitions.
  • A replication policy cannot be assigned to a volume group that is in a storage partition.
  • Replication policies cannot be changed after they are created. If changes are required, a new policy can be created and assigned to a volume group.
  • Each system supports up to a maximum of 32 replication policies.
Two uses cases for replication policy:
Policy-based replication: Asynchronous

For more information, refer to the Replication section.

Policy-based High Availability

For more information, refer to the High availability section.

Creating a replication policy

To create a replication policy, use the Policies > Replication Policies panel in the management GUI. You can also create replication policy by following the Copy services > Partnerships and remote copy panel in the management GUI and click Setup policy-based replication.

On the left under Location 1, select the first system to use in the replication policy. This system always appears on the left when managing and monitoring replication. On the right under Location 2, select the second system to use in the replication policy. This system always appears on the right when managing and monitoring replication. If either system contains multiple I/O groups, select the I/O group that is the caching I/O group for all volumes that are replicated with this policy.

Assigning an asynchronous replication policy to a volume group

To assign an asynchronous replication policy to a volume group, use the Volumes > Volume Groups panel in the management GUI. Select the volume group that needs to be associated with the replication policy, and under the Policies tab, click Assign Replication Policy.

Assigning a high availability replication policy to an existing storage partition

To configure a high availability replication policy on a new storage partition, use the Copy services > Partnerships and remote copy panel in the management GUI and click Setup policy-based replication.

To assign a high availability replication policy to an existing storage partition, use the Storage partitions panel in the management GUI. Choose a partition that does not already have a replication policy assigned and click Add high availability.

For more information on the command-line interface, see mkreplicationpolicy command.

Managing policy-based replication

Replication policies are a key concept for policy-based replication. A replication policy defines how replication is configured between I/O groups in partnered systems, and how replication is applied to a volume group and therefore to all volumes within that group. The replication policy is defined between two policy-based replication capable systems that are already in fully configured partnership. A replication policy can be associated with any number of volume groups. However, a volume group can have at most one replication policy associated.

You can manage replication policies with either the management GUI and the command-line interface. To manage replication policies in the management GUI that are assigned to existing volume groups, select Volumes>Volume Groups. Select the volume group and select Policies.

Replication policies are immutable, and you cannot change a replication policy after it is created. The noneditable property of policy name parameter guarantees that both systems always have the same definition of how replication is configured. For any change, you can create a new policy by using mkreplicationpolicy command with the required changes, and reassigned policy to the volume groups where the updated replication policy is required.