Each CRG defines its own recovery domain. The recovery domain identifies the current role and preferred role of each node within the CRG. While the current role of a node may change, the preferred role is identified when the CRG is created.
The recovery domain also determines the order in which nodes can become the primary access point in the event of an outage.
Nodes can have the following roles:
Primary - The node is the primary access point for the resources associated with the CRG. Only one primary node is allowed. For an application CRG, this is the node where the application is currently running. For a data CRG, this node contains the principle copy of the resources. For a device CRG, this node is the current owner of the devices in the CRG. The primary role is not supported for a peer CRG.
Backup - A backup node will take over the role of the primary access point for resources associated with the CRG in the event of an outage on the primary node. A backup node either contains a copy of the resources that is kept current by replication software, or contains the IP takeover address and activation instructions necessary to access the resources. The recovery domain determines the sequence in which the exit program will attempt to activate backup nodes during a switchover or failover. This role is not supported for peer CRGs.
Replicate - A replicate node contains a copy of the resources associated with the CRG, but cannot participate in a switchover or failover. Replicate nodes are optional. They are reserved for those systems which are either not powerful enough to host applications, are used for queries and reports only, or are perhaps used as a data warehouse server. For peer CRGs, nodes defined as replicate represent an inactive access point.
While any CRG type can be defined and managed through Assure MIMIX for PowerHA, data CRGs are associated with replication activities.