5.3.1 Cluster operation with the log file trapping function

In a cluster system, you must start the log file trapping function separately on each physical host. You cannot start the function by specifying a logical host. However, you can specify whether JP1 events will be forwarded to the event service on a logical host where they can be centrally managed. Configure a forwarding destination that works well with how your system is run.

To register them in the event service of a logical host, execute the jevlogstart command, specifying the event server name of the logical host in the -s option. If you execute the jevlogstart command without the -s option, JP1 events will be registered with the event service on the physical host.

The following describes how to monitor log files on shared and local disks.

Organization of this subsection

(1) Monitoring log files on a shared disk
(2) Monitoring log files on local disks

(1) Monitoring log files on a shared disk

To monitor log files on a shared disk, you must coordinate the start and end of the log file trapping function with the startup and shutdown of the logical host. In the event of a failover on the primary node, stop the log file trapping function on the failed server, and then restart the function on the server that has taken over as the active system.

By using a log-file trap startup definition file, you can collectively stop and restart log file traps in the event of a failover. The following figure shows how the system stops and restarts log file traps as a batch:

Figure 5‒2: Collectively stopping and starting log file traps

To stop and restart log file traps as a batch, specify the log file traps that you want to stop and restart in a log-file trap startup definition file. You can then configure the cluster software to execute the jevlogstop and jevlogstart commands with the -cluster option in the event of a failover. When a failover occurs and the jevlogstart -cluster command is executed, the server that has taken over as the primary node reads the log-file trap startup definition file and starts the log file traps specified in the file. When the jevlogstop -cluster command is executed, JP1/Base stops the log file traps that were started by the jevlogstart -cluster command on the failed host.

If you use a log-file trap startup definition file, check whether the required log file traps have started by viewing the startup record (KAVA3661-I) and startup results (KAVA3662-I) output to the log-file trap startup execution results log.

To configure log file traps to start and stop collectively when a failover occurs:

Edit the log-file trap startup definition file.

In the START_OPT_CLS parameter of the log-file trap startup definition file on the primary and secondary nodes, specify the log file traps that you want to stop and restart automatically when a failover occurs. Specify the START_OPT_CLS parameter in the same way on the primary and secondary nodes. In an environment that incorporates multiple cluster systems, specify the log file traps to be started and stopped for each system.

For details on the log-file trap startup definition file, see Log-file trap startup definition file in 16. Definition Files.
Register the jevlogstop and jevlogstart commands (both for cluster environments only) with the cluster software.

For details on these commands, see jevlogstop (cluster environment only) and jevlogstart (cluster environment only) in 15. Commands.

Leave the shared disk allocated so that it can be accessed while log files are being monitored. If you change the shared disk allocation during file monitoring, problems such as errors in the monitoring process and control failure in disk space allocation and deallocation could occur.

An example of a configuration for monitoring log files on a shared disk is shown in the following figure.

Figure 5‒3: Configuration example for monitoring log files on a shared disk

To Page Top

(2) Monitoring log files on local disks

To monitor log files on the local disk of both the primary and secondary nodes, the JP1 events converted from log data must first be registered in the event service of the physical host. Also, use the forwarding settings file (forward) to perform settings to forward to the event service on a separate host as necessary. For details on the forwarding settings file, see Forwarding settings file in 16. Definition Files.

An example of a system configuration for monitoring, on a logical host, log files on local disks is shown in the following figure.

Figure 5‒4: Configuration example for monitoring, on a logical host, log files on local disks

To Page Top