7.6.9 Monitoring events or messages issued by JP1/AJS3
When monitoring JP1/AJS3 messages, monitor the message ID and the first part of the message, rather than the entire message text. If you monitor the entire message text, JP1/AJS3 could fail to detect the target message if it has extra information added at the end.
You might want to set a JP1/AJS3 event job to monitor JP1 events or Windows events issued by JP1/AJS3, or to monitor the messages output to log files. However, such an event job might be unable to detect its monitoring targets as events depending on the status of JP1/AJS3, or might be executed repeatedly depending on the monitoring condition settings. For example, the following can occur:
-
Like a normal job, an event job is executed as a job in a jobnet. Monitoring of events does not commence until both JP1/AJS3 and the jobnet have started. Because the jobnet-initiating event is issued after the first job starts, an event job cannot monitor a jobnet-initiating event of JP1/AJS3 which is executing the event job.
-
If a problem occurs in JP1/AJS3 and the jobnet in which that event job is defined ends abnormally, the event job itself also ends abnormally. If this situation occurs, events cannot be monitored.
-
An event job also issues JP1 events or Windows events, or outputs messages to log files, as part of its job processing. Because the event job monitors the JP1 events, Windows events, and messages that the event job itself outputs to the log files, monitoring continues in an infinite loop.
When you want to execute actions by using events from JP1/AJS3 or messages in log files as triggers, you can avoid these problems by using use the automatic action functionality of JP1/IM. Alternatively, you could use a JP1/AJS3 manager that is different from the one executing the event job. The preventive measures when you want to execute commands by using events from JP1/AJS3 or messages in log files as triggers are described below for two cases: when there are multiple JP1/AJS3 managers (including a configuration where multiple instances of JP1/AJS3 are concurrently running on multiple logical hosts) and when there is only one JP1/AJS3 manager (only one instance of JP1/AJS3). The following also gives notes on monitoring events and messages.
Note that the description in this section applies to monitoring targets such as JP1 events issued by JP1/AJS3 itself, Windows events, and messages output to log files or the syslog. The description does not apply when the Send JP1 Event job is used or when JP1 events are sent as a user job.
- Organization of this subsection
(1) When there are multiple JP1/AJS3 managers (including a configuration where multiple instances of JP1/AJS3 are concurrently running on multiple logical hosts)
The following describes the measures to take to monitor events issued by JP1/AJS3 on another logical host or messages in log files when there are multiple JP1/AJS3 managers (or when multiple instances of JP1/AJS3 are concurrently running on multiple logical hosts). In this case, JP1/IM is not needed.
(a) Monitoring JP1 events issued by JP1/AJS3 on another logical host, or when there are multiple JP1/AJS3 managers
Under normal circumstances, you cannot monitor JP1 events issued by an instance of JP1/AJS3 running on another host (logical host). However, you can monitor such events if you transfer them using the event service functionality of JP1/Base. To do this, transfer JP1 events to an event service on another host (logical host) that is not managed by the JP1/AJS3 that issued the JP1 events. You cannot monitor the events if you transfer them to the event service running on the host managed by the JP1/AJS3 that registered the JP1 events.
The following figure shows an example flow of JP1 events.
|
In the above example, if you want to monitor the JP1 events issued by JP1/AJS3 on HOSTA, transfer them to the event service on HOSTC or HOSTD. You cannot monitor the events if you transfer them to the event service on HOSTB.
(b) Monitoring Windows events issued by JP1/AJS3 on another logical host
Windows events can only be monitored on logical hosts that use the JP1/Base event log trapping function.
When you want to monitor Windows events issued by JP1/AJS3, set up a separate logical host and monitor the events from a different manager. In this case, you cannot use the Monitoring Event Log job from the monitored JP1/AJS3.
(c) Monitoring log file messages output by JP1/AJS3 on another logical host
Reference the monitored log files from JP1/AJS3 on another logical host.
(2) When there is only one JP1/AJS3 manager (only one instance of JP1/AJS3)
The following describes the measures to take to monitor events issued by JP1/AJS3 or messages in log files when there is only one JP1/AJS3 manager. In this case, JP1/IM is needed.
(a) Monitoring JP1 events issued by JP1/AJS3
Use the automatic action functionality of JP1/IM.
(b) Monitoring Windows events issued by JP1/AJS3
Use the Windows event trapping functionality of JP1/Base and the automatic action functionality of JP1/IM.
(c) Monitoring log file messages output by JP1/AJS3
Use the log file trapping functionality of JP1/Base and the automatic action functionality of JP1/IM.
(3) Notes on monitoring events and messages
The following provides precautions for monitoring events or messages.
(a) Restrictions when using an event job to monitor a JP1/AJS3 - Manager host from another JP1/AJS3 - Manager host
The following describes the restrictions that apply when HOSTB is monitored from HOSTA by using an event job as shown in the figure below. In this case, JP1/AJS - Manager host (HOSTA) is set as the manager of HOSTB (monitoring host), and another JP1/AJS3 -Manager host (HOSTB) is set as the agent of HOSTA (monitored host).
|
- Restrictions
-
-
When JP1/AJS3 is not running on HOSTB, you cannot use an event job to monitor HOSTB.
The manager provides the agent functionality. Therefore, the agent of HOSTB monitored by HOSTA is started or stopped in tandem with the manager of HOSTB starting or stopping.
-
If JP1/AJS3 on HOSTB terminates abnormally, it cannot provide information to HOSTA (you cannot monitor HOSTB from HOSTA).
The manager provides the agent functionality. Therefore, the agent of HOSTB monitored by HOSTA stops in tandem with the manager of HOSTB ending abnormally.
-
To monitor a host in this configuration, transfer events to a JP1/AJS3 - Manager host that is neither the monitoring JP1/AJS3 - Manager host nor the monitored host.
|
(b) Monitoring events and messages when multiple JP1/AJS3 - Manager hosts use the same JP1/AJS3 - Agent host as an agent
When the same JP1/AJS3 - Agent host is configured as an agent for multiple JP1/AJS3 - Manager hosts, JP1/AJS3 - Agent monitors the same log files (JP1/AJS3 log files to which the instance of JP1/AJS3 on the Agent host outputs information) for events and messages. Therefore, you cannot monitor the log files correctly even from a Manager on a different host.
If you want to monitor events and messages in this configuration, transfer the events and messages to a different JP1/AJS3 - Manager host from the one managing the JP1/AJS3 - Agent.
|