Optimize Distributed Workload - Expert Oracle RAC Performance Diagnostics and Tuning - page 492

ONS

ONS ONS

Public Interface

Cluster Interconnect

ORADB1 ORADB2 ORADB3 ORADB4

VIP VIP VIP VIP

SSKY1 SSKY2 SSKY3 SSKY4

ONS ONS ONS ONS

+ASM1 +ASM2 +ASM3 +ASM4

AV1 AVL9 AV10 AV11 AV12 AV13 AV14 AV15 AV16 AV17 AV18

GRID_DATA PRD_DATA PRD_FRA

Figure 15-3. FAN—ONS communication

As illustrated in Figure 15-3 , FAN uses ONS for server to server and server to client notification of state changes,
which includes up, down, and restart events for all components of the RAC cluster. For example, in Figure 15-3 ,
ONS daemon on node oradb2 notifies of any state changes with respect to components on that node to all other
nodes in the cluster and to all client machines running ONS. The node sends all events, with the exception of a node
failure event, where the event is generated; in the case of a node failure, then one of the surviving nodes will send the
notification.
Based on the notification received, the FAN calls inside the application will proactively react to the situation,
which includes failover of connections to another instance where the service is supported.
Oracle uses the advanced queuing (AQ; illustrated previously in Figure 15-2 ) technology for event notifications
between various servers and clients.