6.3 Changing the behavior of JP1/AJS3 if a JP1/AJS3 process terminates abnormally
If JP1/AJS3 stops for any reason, it is no longer able to execute jobs, adversely affecting the jobs in a distributed system.
If a process stops because of an error, JP1/AJS3 attempts recovery by automatically restarting the process. If the recovery fails, JP1/AJS3 operates at a reduced level without stopping any processes other than the process that caused the error. For details about automatically restarting a process that has terminated abnormally, see 6.3.1 Restarting an abnormally terminated JP1/AJS3 process.
In the reduced-operation mode, only some functions are available, depending on the running processes.
To display the processes in the JP1/AJS3 service, you can use the jajs_spmd_status command.
For details about the functions of processes, see B. List of Processes in the manual JP1/Automatic Job Management System 3 Troubleshooting.
Because processes of the agent service function (job execution and event jobs) do not stop when a process of the scheduler service management function stops, the execution of jobs and events jobs requested by other hosts is still able to continue.
You can also use the jajs_spmd_status command to check whether the JP1/AJS3 service is currently in a reduced-operation mode.
If you specify the -HA option in the jajs_spmd command or if an error occurs in the JP1/AJS3 service on a logical host in a cluster configuration, there is no reduced-operation mode. If any one of the processes for the JP1/AJS3 service displayed by executing the jajs_spmd_status command stops, the entire JP1/AJS3 service will stop.
To detect abnormal condition without causing JP1/AJS3 on a physical host to run in the reduced-operation mode, specify the jajs_spmd command with the -HA option specified and start JP1/AJS3. You can also stop the entire JP1/AJS3 service.
In addition to the above operations, as a method for detecting failures and taking appropriate action, JP1/AJS3 provides a function that issues a JP1 event if JP1/AJS3 detects an abnormal process. You can specify this function beforehand to issue a JP1 event if a process terminates abnormally.
When issuance of JP1 events is enabled, you can use JP1/AJS3 with other programs, such as JP1/IM, to automatically notify users when a JP1 event is issued.
For details about the function that issues a JP1 event when an abnormal process is detected, see the subsequent subsections.