This section explains some points to consider in cluster service operation.
Starting and stopping
When the active node is started, the CEP service and CEP engine will start.
PRIMECLUSTER process monitoring will be performed for the CEP service. Therefore, do not execute cepstopserv unless it is for a cluster switch.
Sending events
Send events to the takeover IP address.
Changing CEP engines and development assets
Set up the CEP engine configurations and deploy the development assets (such as definition information and master data) so that the active node and the standby node have the same resource configuration.
Cluster operation and response at an active node fault
When a hardware fault occurs at the active node and the cluster service detects the abnormality, the active node will be stopped and the CEP engine and CEP service will stop.
After a switch to the standby node, event sending to the takeover IP address can restart.
When an abnormality occurs at the active node, take action to resolve the hardware fault that caused the cluster switch, and then take action to allow a switch back in case an abnormality occurs at the standby node.
