This topic describes multi-cluster federated publishing for disaster recovery in LHC. It covers basic concepts and disaster awareness.
Background information
LHC provides disaster recovery protection using multi-cluster federated publishing. This feature is useful when a site cannot quickly recover an application due to events such as force majeure or device failure. If a site fails, simple configurations enable quick service recovery at a disaster recovery site.
Disaster recovery is a broad concept. Broadly, it is a systems engineering practice that covers all aspects of business continuity. More narrowly, disaster recovery involves setting up two or more identical IT systems that monitor each other's status and can switch over functions. If the primary site stops working unexpectedly, the entire application system can use a secondary site to quickly recover and continue its operations.
The main purpose of disaster recovery is to ensure business continuity when the production system is affected by a natural or human-made disaster.
Data center disaster awareness
To ensure successful publishing for LHC multi-cluster deployments in a disaster recovery scenario, you can determine whether a data center disaster has occurred in the following ways:
Cluster status: From an Operations and Maintenance (O&M) perspective, an unavailable cluster directly indicates a data center disaster.
Deployment unit status: From an application perspective, a data center disaster ultimately results in an unavailable deployment unit (Cell). This can lead to unexpected results when you publish an application service using a release order.
Multi-cluster federated publishing during a data center disaster
Multiple options are available for application publishing in a disaster recovery scenario. Before you execute a release order, you can publish the application by specifying only the active Cells.
For more information, see Create a release order.