Configuring Fault Tolerance Parameters

For the Core Engines, you should run the multiple instances across the servers in a load balanced setup. The configuration for load balanced setup is defined in the asg_core.cdd file. See Configuring Core Engines.

For the Cache Cleanup Agent, Global Throttle Manager and Central Logger components, the instances must be deployed as one active engine and one or more stand by agents that run on a separate server. See Configuring Cache Cleanup Agent, Configuring Global Throttle Manager and Configuring Central Logger.

The maximum number of active instances are configured as an agent class configuration parameter. An agent class is an agent type, defined in the CDD file that deploys as an agent instance.

The following parameters define the fault tolerant configuration of runtime components:

Fault Tolerant Configuration Parameters
Parameter Description
Max Active
  • Specifies the maximum number of active agents. This value is used for fault tolerance. Deployed agents that are acting as standbys can take over from active instances that fail. In many cases, there is no need to keep standby instances.
Priority
  • Specifies the priority of the agent for fault tolerant setup of agents. The priority is set at the processing unit level.
  • The priority indicates the order in which standby agents become active, and conversely, the order in which active agents become standbys, when new agents join the cluster. The lower the number, the higher the agent is in the activation priority list. For example, an agent with priority 2 has a higher priority than an agent with a priority of 6. This value determines the order of each instance of an agent class for startup, as well as fail over and fail back in fault tolerance situations.
  • If the priority values are same for agents in fault tolerant setup (active passive), the agent which is started first gets the higher priority to become active agent instance.

This section explains the steps for setting the maximum number of active instances and priority required for the high availability of runtime components.