Understanding and Designing Serviceguard Disaster Recovery Architectures

The following are the main components of Site Aware Disaster Tolerant Architecture:
• Sites
• Complex Workload
• Site Controller Package
• Site Safety Latch
Terms and Concepts
Sites
Site, in SADTA, is a collection of Metrocluster nodes in the same location that are connected to
the same disk array. The site information must be provided in the cluster configuration file. The
nodes within a site form a sub-cluster and multi-node packages can be restricted to run within this
sub-cluster.
Complex Workload
Complex workload is a set of multi-node and failover packages with dependencies among them
and running within a single site. In any SADTA Metrocluster, there are two complex workload per
Site Controller. These complex workload sets are identical to each other, one production and the
other recovery. The production complex workload is called the active complex workload set and
the recovery one is called the redundant complex workload set.
Critical and managed packages
The packages in a complex workload can be marked either as critical packages or managed
packages. If a package marked as critical, fails, the Site Controller halts all packages including
other critical packages and starts up the redundant complex workload packages. If no packages
in a complex workload are marked as critical, the Site Controller starts up the redundant packages
only when all the packages of the currently running complex workload fails.
Root package in a complex workload
The packages in a complex workload that do not have a dependency on any other package is
known as the root package in a complex workload.
Understanding Site Aware Disaster Recovery Architecture Concepts 37