Stray nodes

When you implement disaster-recovery strategy, IBM Business Automation Workflow cluster members are distributed across the primary data center and the secondary data center, but only the cluster members in the primary data center are active. The cluster members in the secondary data center are activated when a disaster occurs, and the primary data center becomes invalid. The data in the database is replicated from the primary data center to the secondary data center by using database technology (for example, Db2® HADR) or SAN replication.

IBM Business Automation Workflow integrates case capabilities into products. If an external Content Platform Engine is designed to store document and related objects to use the flexibility of a stray nodes disaster-recovery approach, one sample topology can use multiple Content Platform Engine clusters that are configured with advanced storage area.

In such a set-up (see Figure 2), IBM Business Automation Workflow cluster members are distributed across the primary data center and the secondary data center. The configuration changes in the IBM Business Automation Workflow clusters are replicated to the secondary data center automatically through the node agent. The data changes in the database are replicated to secondary data center by using database replication technology.

Each data center configures a Content Platform Engine cluster. The advanced storage area is configured across the Content Platform Engine clusters in different data centers, automatically assure the content that is replicated between the data centers. While IBM Business Automation Workflow persists documents to an external Content Platform Engine, document objects are persisted into a replica of the advanced storage area in the primary data center. The advanced storage area automatically replicates these documents to peer replicas.

To enable Content Platform Engine clusters to service, data requests in the primary data center, request forwarding is configured. IBM Business Automation Workflow cluster members are located in primary data center to communicate with the Content Platform Engine cluster in same primary data center. IBM Business Automation Workflow cluster members that are located in secondary data center communicate with the Content Platform Engine cluster in same secondary data center.

To access the external Content Platform Engine, map one virtual hostname to a different Content Platform Engine node in different data centers. The virtual hostname is used during configuration with IBM Business Automation Workflow to access the external Content Platform Engine.

During Content Platform Engine configuration, the servers in the Content Platform Engine clusters are organized into different Content Platform Engine sites and the object stores are associated to the site located in primary data center. When activating the secondary data center, the related components must be associated to the site located in the secondary data center.

To move the related components from the Content Platform Engine site located in the primary data center to another Content Platform Engine site located in the secondary data center, use the IBM Administration Console for Content Platform Engine. At the same time, remove the request forwarding configuration from the Content Platform Engine configuration, restart the Content Platform Engine clusters that are located in secondary data center, and then start the IBM Business Automation Workflow cluster members that are located in the secondary data center.

Figure 2. Stray Nodes disaster recovery