7.4 Data collected when a problem occurs
If a problem cannot be solved even after you have taken the actions prescribed in 7.2 Troubleshooting, you need to collect data to investigate the cause of the problem, and then contact the system administrator. This section explains the data that needs to be collected when a problem occurs.
Performance Management provides a command for batch-collecting the necessary data. To collect PFM - RM for Virtual Machine data, use the jpcras command. Data that can be collected using the jpcras command is indicated in the tables below.
- Note:
-
Data that can be collected using the jpcras command differs depending on the option specified for command execution. For details about the options that can be specified for the command and the types of data that can be collected, see the chapter that explains commands in the manual JP1/Performance Management Reference.
- Notes about operations in a logical host environment
-
Note the following points when Performance Management is running on a logical host:
-
When Performance Management is running on a logical host, the Performance Management logs are output to a shared disk. If the shared disk is online, you can use the jpcras command to batch-collect the logs from the shared disk.
-
To investigate a problem that occurred during a failover, you need the data that was present before and after the failover. Therefore, you need data from both the active server and the standby server.
-
To investigate Performance Management when it is running on a logical host, you need data from the cluster software. Because the operations of starting and stopping Performance Management are controlled by the cluster software in this case, you need to compare the behavior of the cluster software with that of Performance Management.
-
- Organization of this section