If more than one node application has been started for a UTM cluster application, each node application is monitored by a different node application.
The following are dynamically defined when a node application is started:
what other node application is to be monitored by this node application,
and what other node application is to monitor this node application.
These monitoring relationships are entered in the cluster configuration file. When the node application is terminated, the relationships are canceled.
Monitoring process
The availability of a node application is monitored. Heartbeat monitoring is performed using messages which are exchanged over a special connection. If errors occur during communication, the system checks whether the KDCFILE of the monitored node is still open.
Only when the result of all these checks indicates failure is it assumed that the monitored node has failed.
You can specify the following individual aspects of monitoring (in the UTM generation):
the interval between the monitoring messages,
the time that the application waits for a response to the message,
the retry factor, the number of retries before level 2 of monitoring takes effect if no response is received to a message.
openUTM manual “Generating Applications”, CLUSTER statement
You configure mutual monitoring between the node applications using the following operands:
CHECK-ALIVE-TIMER-SEC=
COMMUNICATION-REPLY-TIMER-SEC=
COMMUNICATION-RETRY-NUMBER=