E.1 Example of hardware configuration and overview of HACMP for AIX behavior

When HACMP for AIX detects a failure in a monitored node, it switches processes to the standby node to continue services. This behavior is called a takeover. A takeover has the following switching functionalities:

Node
Application
Network and network adapter
Disk and disk adapter

Examples of a network failure and application failure are as follows.

Organization of this subsection

(1) Example of a network failure
(2) Example of an application failure

(1) Example of a network failure

In a double LAN adapter configuration, define one adapter to run for application services (as the service adapter), and the other as a backup service adapter (as the standby adapter).

In this configuration, when a failure occurs in the service adapter, the connection will be switched to the standby adapter that resides on the same node. The following figure shows an example of double adapter configuration.

Figure E‒1: Double adapter configuration

To Page Top

(2) Example of an application failure

The following example shows a double node configuration.

The example is a double LAN configuration consisting of one active node (service node) and one backup node (standby node), where the two nodes are connected using RS232 signals as a monitoring path. The disk storage device is shared between nodes.

If a failure occurs in the service node and HACMP for AIX determines that a failover to another node is needed, HACMP for AIX removes the service node and continues the services on the standby node. The shared disk can continue to be used on the standby node. The following figure shows an example of multiple node operation.

Figure E‒2: Example of multiple node operation (HACMP for AIX)

To Page Top