Understanding and Designing Serviceguard Disaster Recovery Architectures

Figure 17 SADTA Configuration in Continentalclusters
Continentalclusters
Node 1 Node 2
Disk Array
Active Application
Configuration
Primary Cluster Recovery Cluster
Data Replication
Passive Application
Configuration
Site A
Site A
App. Pkg
Site Safety
Latch
Site Safety
Latch
Site A Mount Point MNP
Site A Disk Group MNP
Site
Controller
Node 1 Node 2
Disk Array
Site B
Site B Mount Point
Site B CFS Sub Cluster
Site A CFS Sub Cluster SG CFS SMNP
Site B
App. Pkg
Application
Data Disk
Application
Data Disk
Site
Controller
Site B Disk Group
Features of SADTA
Simplified Disaster Recovery for complex workloads
SADTA, by means of Site Controller, offers a single interface to move a complex stack of
inter-dependent packages. It ensures the applications are started in the right order after preparing
the replicated storage.
Site Preferred failover policy
In a Metrocluster SADTA configuration, when a node or a critical package fails Site Controller
tries to start the failed packages on any viable node within the same site. Moving complex workloads
running mission-critical applications across sites unnecessarily is disruptive and causes downtime
while the complex workload starts up on the recovery site. The Site Preferred Failover Policy can
help to mitigate unnecessary failover across sites.
Support for Oracle RAC and CFS
SADTA with Site Controller is the only way to configure Oracle RAC and CFS in a Metrocluster.
It is also the recommended way to configure Oracle RAC and CFS in Continentalclusters.
40 Metrocluster and Continentalclusters