N. What to Do When the Message KAVL20516-W Is Output Repeatedly

When you restart the system or when a system in a cluster configuration is switched, the message KAVL20516-W can be output repeatedly depending on the time when the PFM - RM for Virtual Machine service starts. This can occur even though a problem has not occurred (for example, the system is temporarily under high load or takes a long time to communicate with the monitoring target as shown in the Figure 1-10 in 1.3.1 (2) General procedure for collecting performance data).

If you restart the PFM - RM for Virtual Machine service, the message KAVL20516-W is no longer output.

In addition, you can configure PFM - RM for Virtual Machine so that the KAVL20516-W message is not output when the system is restarted or switched in a cluster configuration. To do this, use either of the following settings:

(a) Changing the collection interval settings (Collection Interval and Interval) with which performance data is collected (if UseOffset is set to disable (N))

When you set up the instance environment, set the Interval value to half of the Collection Interval value, which enables the collection process of the Remote Monitor Collector service to read a new temporary performance information file.

The following is an example when you set the Collection Interval value of the record to 300 seconds, and the Interval value in the instance environment to 150 seconds.

Figure N‒1: Example procedure for when changing the collection interval value of performance data

(b) Using a zero reference point based on which the collection process starts (if UseOffset is set to enable (Y))

If you set a short time (such as the minimum value of 60 seconds) to the Collection Interval value for each record and the Interval value of the instance environment, you cannot set half of the Collection Interval value to the Interval value. In that case, set UseOffset to enable (Y) in the PFM - RM for Virtual Machine setup file so that the collection process starts based on a zero reference point. For details about the settings, see 2.1.4(2) Setting PFM - RM for Virtual Machine.

The following is an example when you set the UseOffset value to enable (Y), the Collection Interval value of the record to 60 seconds, and the Offset value of the instance environment to 10 seconds.

Figure N‒2: Example procedure for when using a zero reference point

Tip

If you set the UseOffset value to enable (Y), the first collection immediately after the service starts is performed one minute after the collection process starts. Subsequent collections start in the period (in seconds, specified in the Offset value during startup of the instance environment) from the time the next collection process is to be started (which is the zero reference point).

If multiple instances are monitored and you choose to use a zero reference point for the collection process, you might need the settings described below.

Load balancing for the collection process

If you set up more than one instance while using the default Offset value (10 seconds) in the instance environment, collection processes start at the same time. In an environment where a large number of instances exist, the load on the PFM - RM for Virtual Machine host increases significantly, which might cause an operational problem.

To prevent this, set a different Offset value for each instance.

The Offset value can be 1 through 3600. However, taking into account the time it takes for the Remote Monitor Collector service to read a temporary performance information file, specify a time about 10 seconds longer than the Collection Offset value of each record.

Figure N‒3: Example for when the same Offset value is used for all instances

Figure N‒4: Example for when a different Offset value is used for each instance

Changing the Collection Offset value

Setting a larger Offset value close the Collection Interval value (which determines when the Remote Monitor Collector service reads a temporary performance information file) can lead a repeated KAVL20516-W message output. To prevent this, use the Collection Offset setting for each record to change the time for which the service reads the temporary file.

If you do so, use the same Collection Offset value for all records in the instance.

Figure N‒5: Example for when a different Collection Offset value is used for each instance

To Page Top