7.10 Notes on using custom event jobs
This section gives notes on using custom event jobs.
-
Custom event jobs do not support operations that use execution agent groups. If a custom event job without a specified execution agent is included in a root jobnet or nested jobnet for which an execution agent group is specified, JP1/AJS3 will attempt to use the agent group specified for the jobnet as the execution agent for the custom event job. If an execution agent with the same name as the agent group exists, the custom event job will be executed by that execution agent. If there is no execution agent with the same name as the agent group, an error occurs and the following message is output to the integrated trace log: KAVT0403-E The specified agent is not defined in the job execution environment. (host=exec-agent, maintenance-information). Therefore, if you want to specify an execution agent group for a root jobnet or nested jobnet, make sure that an execution agent is explicitly specified for the custom event job in the jobnet.
-
You cannot specify the name of an execution agent group in the Exec-agent field of the detailed definition of a custom event job.
-
For a custom event job waiting for a start condition to be satisfied, if JP1/AJS3 - Manager stops during start condition monitoring, you can resume event monitoring after JP1/AJS3 - Manager restarts. If multiple events were being monitored for a start condition, you can hold the reception information of the events that satisfy the condition even after JP1/AJS3 - Manager restarts.
-
When a custom event job is defined at the beginning of a jobnet, the jobnet is executed after the condition is satisfied, in the same way as a jobnet with a start condition. A jobnet with a start condition stays in Wait for start cond. status while event reception is being monitored. A jobnet with a custom event job defined at the beginning is kept in Now running status while event reception is being monitored. When you define a custom event job at the beginning of a jobnet, we recommend that you use the custom event job to wait for events you expect to occur.
-
If a timeout period is specified for a custom event job, the counting is done by the execution host. Therefore, when event monitoring resumes after the execution host is restarted due to a power failure or similar condition, counting restarts upon the restarting of the execution host.
Note that you can verify the restart of counting and the restart time in the message KAVT0603-W Elapsed time since restart-time is used for time-out due to temporary interruption of monitoring. that is output in the custom event job execution result details.
If you want to stop monitoring at an absolute time regardless of the status of the agent host, define a custom event job as a jobnet start condition, and then specify the effective range of the start condition by using an absolute time. For details about a start condition, see 3.4 Defining a start condition in the manual JP1/Automatic Job Management System 3 Overview.
-
When using custom event jobs, system load or temporary network failure might cause a time lag between the job execution time and the time that the execution agent actually starts monitoring the events. The system detects only events that occur after event monitoring begins. Therefore, you must provide some leeway when executing a custom event job, to allow the system to be ready and monitoring when a target event occurs.
-
When you execute custom event jobs (including those with a start condition), definition data for the executed custom event jobs and information about events that satisfy monitoring conditions is transmitted between processes such as the event/action control manager and the event/action control agent. If transmission fails due to some problem such as a temporary network error or because the remote process is busy, the information is saved to a file and transmission is attempted again after a predetermined interval. In JP1/AJS3, this is referred to as an unreported information file. At successful re-transmission, the information is deleted.
-
In a manager/agent configuration, if the event/action control manager and the event/action control agent are unable to communicate due to a network error or other problem, inconsistencies in custom event job monitoring (including custom event jobs defined as start conditions) can occur if you perform any of the operations listed below. In this case, the custom event job will continue monitoring on the agent despite having ended on the manager.
-
Killing a custom event job
-
Killing a jobnet that has a start condition
-
Changing the status of a custom event job to Ended
These operations can result in problems such as failure to restart the custom event job where the inconsistency occurred, and delays to processing of other normal custom event jobs.
For this reason, if you perform one of the above operations in a system affected by a network error or similar problem, execute the jpomanjobshow command on the manager host, and the jpoagtjobshow command on the agent host. Compare the results of the commands to see whether any custom event jobs that have ended on the manager are still monitoring on the agent.
If you discover a custom event job that is in Now monitoring status only on the agent host, restart the JP1/AJS3 service on the agent host, and then terminate the custom event job that is still monitoring on the agent.
-