5.1.1 Considerations related to the monitoring methods used in JP1/AJS3 - View
This subsection describes the considerations necessary for monitoring in JP1/AJS3 - View.
- Organization of this subsection
(1) Planning which work tasks to monitor in the Summary Monitor window
By designating selected jobnets as monitoring targets in the JP1/AJS3 - View window (Summary Monitor window), you can monitor the execution status of the jobnets, including the execution progress and the statuses of subordinate jobs. You can also view a simulation of the jobnet's end time, calculated from the jobnet's execution times in the past. This can be useful when any delay in starting or ending the work task would have a major effect on the execution schedule for subsequent work tasks.
JP1/AJS3 - View also allows you to centrally monitor jobnets that are distributed over different jobnet hierarchies or are governed by different scheduler services.
(2) Jobnet delay monitoring considerations
JP1/AJS3 provides the following methods for detecting delays in jobnet execution.
(a) Monitoring by date and time specification
With this method, you specify a date and time, and JP1/AJS3 monitors whether a jobnet starts and ends at the scheduled time. JP1/AJS3 provides two methods of monitoring for delays by date and time specification: Monitor start delay and Monitor end delay.
Use these methods when a delay in starting or ending the jobnet will have a major effect on the execution schedule for subsequent jobnets.
- Monitor start delay
-
Set a date and time for judging when a start is delayed. If the jobnet starts within a certain time after the scheduled start time, it is deemed normal. If the jobnet has still not executed when the grace period elapses, some abnormality is assumed to have occurred.
- Monitor end delay
-
Set a time for judging when completion is delayed. If the jobnet ends within a certain time after the scheduled start time, it is deemed normal. If jobnet execution has still not ended when the grace period elapses, some abnormality is assumed to have occurred.
You can designate both a start delay and an end delay, or just one or the other.
(b) Monitoring by time-required-for-execution
You can also monitor for delays based on the time required for execution of the jobnet.
If the time set is exceeded, the color of the work task icon changes, and an event is issued to report the delay.
(3) End delay monitoring based on time-required-for-execution
You can also monitor for end delays based on the time-required-for-execution of a job. In this case, JP1/AJS3 monitors how long the job takes to execute on the manager host, which may differ from how long it takes on the agent host. This discrepancy can affect whether an end delay is detected.
Timeout monitoring, on the other hand, is based on how long the job is active on the agent host. For this reason, there is a time difference between timeout detection and the detection of an end delay.
The following are examples of situations where a job has a different time-required-for-execution on the manager host and agent host:
-
The job is started on the agent host while the scheduler service is stopped.
If a job begins executing on the agent host while the scheduler service is stopped (that is, the scheduler service stops before the manager host receives notification from the agent host that the job has started), the job does not enter Now running status until the scheduler service restarts. Here, the job will have a longer execution time on the agent host than on the manager host. For example, suppose the job's time-required-for-execution is 10 minutes, and it has been executing for 20 minutes on the agent host. In this case, if the job finishes five minutes after the scheduler service restarts, an end delay will not be detected.
-
The job ends on the agent host while the scheduler service is stopped.
If a job that was already running on the agent host ends while the scheduler service is stopped (that is, the scheduler service stops before the manager host receives notification from the agent host that the job has ended), the job does not acquire an end status until the scheduler service restarts. Here, the job will have a shorter execution time on the agent host than on the manager host. For example, suppose the job's time-required-for-execution is 10 minutes, and it ends on the agent host after five minutes. Because the job is considered to be running the entire time the scheduler service was stopped, an end delay will be detected if the scheduler service is stopped for longer than 10 minutes.
-
The job starts and ends on the agent host while the scheduler service is stopped.
If a job starts and finishes on the agent host while the scheduler service is stopped (that is, the scheduler service stops before the manager host receives notification from the agent host that the job has started and ended), the job enters Now running status then ends as soon as the scheduler service restarts. Here, the job execution time on the manager host will be zero minutes. For example, suppose the job's time-required-for-execution is 10 minutes, and it took 20 minutes to execute on the agent host. In this case, an end delay will not be detected because the job's execution time on the manager host is zero minutes.
- Supplementary notes
-
-
When an end delay is detected for a job, you can check the job's execution time on the agent host by using the ajsshow command. Execute the command with the format specifiers %V and %Q to acquire the execution start and end times of the job. For the command syntax, see ajsshow in 2. Commands in the manual Job Management Partner 1/Automatic Job Management System 3 Command Reference 1.
-
If you specify yes in the JOBDELAYWARNMSG environment setting parameter of the scheduler service, the message KAVS0249-W The scheduler services stopped before execution of the job began. is output to the integrated trace log when a job being monitored for an end delay starts executing while the scheduler service is stopped. You can identify the job name and execution ID from the content of the message.
To verify an end delay based on the time-required-for-execution of the job on the agent host, use the ajsshow command to acquire the execution start and end times of the job identified in the message KAVS0249-W.
For information about the environment setting parameter JOBDELAYWARNMSG, see 2.2 Setting up the scheduler service environment in the Job Management Partner 1/Automatic Job Management System 3 Configuration Guide 2.
-