Understanding and Designing Serviceguard Disaster Recovery Architectures

Figure 14 Cascading Failover Data Center Distribution Using Metrocluster
Continentalclusters
Primary Cluster Recovery Cluster
Quorum Server
IP Router
IP Router
Host A Host B Host C
Primary Secondary Recovery DA
Recovery
Site 3
Site 1 Site 2 Site 4
WAN
SRC (A)
Local Mirror SRC (B’) Local Mirror SRC (C’)
DEST (C)
App. X
DEST (B)
Features of Continentalclusters
Cluster Event Notifications
The Cluster Event Notification feature allows you to configure notifications to indicate changes in
cluster status. Administrators can customize these notifications according to their solution needs.
Push-button Recovery
The recovery process is manually initiated but the recovery is automated. The automation of the
recovery process reduces the chances of human errors. The recovery packages can be customized
by using user defined modules or by external scripts integrated with the recovery package to
automate all aspects of the recovery process. For example, the DNS values can be updated to
reflect the change of servers where the production application is running.
Highly available monitoring
The primary cluster is monitored by the monitoring software. This software runs as a Serviceguard
package in the recovery cluster. This ensures that the monitoring software is highly available even
in case of node failures in the recovery cluster.
Understanding Continentalclusters 33