Hitachi

JP1 Version 12 JP1/Performance Management - Remote Monitor for Oracle Description, User's Guide and Reference


7.5.2 In UNIX

Organization of this subsection

(1) Execute the data collection command

Use the jpcras command to collect the information needed to investigate an error. The following procedure shows how to execute the data collection command. Note that you must perform these operations as an OS user who has root privileges.

To execute the data collection command:

  1. Log in to the host on which the service whose data you are collecting is installed.

  2. Execute the jpcras command, specifying the data to collect and the directory in which to store it.

    For example, to have the jpcras command store all collectible data in the directory /tmp/jpc/agt specify the command as follows:

    /opt/jp1pc/tools/jpcras /tmp/jpc/agt all all

    The data collected by the data collection command is compressed with the tar and compress commands and stored in the specified directory. The file is given the following name:

    jpcrasYYMMDD.tar.Z

    YYMMDD indicates the date.

When you execute jpcras command, jpctool service list -id * -host * command is executed to acquire the PFM service list and check the startup status of these services. If the execution-host-name and other Performance Management System host communicate across a firewall or if the scale of the system is large, it may take time to execute jpctool service list -id * -host * command. In that case, by setting JPC_COLCTRLNOHOST environment variable to 1 jpctool service list -id * -host * command is not executed so that jpcras command can take shorter time.

For details about the jpcras command, see the chapter on commands in the manual JP1/Performance Management Reference.

(2) Execute the data collection command (for logical host use)

When you run Performance Management in a logical host environment, the data is stored on a shared disk. In this case, collect data from both the executing node and the standby node.

Use the jpcras command to collect the information needed to investigate an error. The following procedure shows how to execute the data collection command. Note that you must perform these operations as an OS user who has root privileges.

To execute the data collection command in a logical host environment:

  1. Mount the shared disk.

    For logical hosts, data is stored on a shared disk. On the executing node, make sure that the shared disk is mounted before collecting data.

  2. Execute the jpcras command on both the executing node and the standby node, specifying the data to collect and the directory in which to store it.

    For example, to have the jpcras command store all collectible data in the directory /tmp/jpc/agt specify the command as follows:

    /opt/jp1pc/tools/jpcras /tmp/jpc/agt all all

    The data collected by the data collection command is compressed with the tar and compress commands and stored in the specified directory. The file is given the following name:

    jpcrasYYMMDD.tar.Z

    YYMMDD indicates the date.

    If you execute the jpcras command without specifying the lhost argument, data relating to Performance Management is collected from all physical and logical hosts on that node. If any Performance Management programs are running in a logical host environment, their log files will be collected from the shared disk.

    If the shared disk is not mounted on the node on which you execute the jpcras command, you will be unable to acquire the files on the shared disk. However, the command will end normally without generating an error.

    Note

    Collect data from both the executing node and standby node by executing the data collection command on each node. Data from both nodes is required to investigate problems leading up to and following failover.

    When you execute jpcras command, jpctool service list -id * -host * command is executed to acquire the PFM service list and check the startup status of these services. If the execution-host-name and other Performance Management System host communicate across a firewall or if the scale of the system is large, it may take time to execute jpctool service list -id * -host * command. In that case, by setting JPC_COLCTRLNOHOST environment variable to 1 jpctool service list -id * -host * command is not executed so that jpcras command can take shorter time.

    For details about the jpcras command, see the chapter on commands in the manual JP1/Performance Management Reference.

  3. Collect data about the cluster software.

    This data is required to determine whether a problem is caused by Performance Management or the cluster software. Collect data from which you can determine the results of any instructions, such as start and stop requests, that the cluster software issued to Performance Management.

(3) Check the operation information

If an error occurs while an operation is being performed, check and record information about the operation. Check the following information:

(4) Collect error information

Collect the following error information:
  • Messages output to the console when an error occurs during command execution

(5) Collect other information

Also collect the following information:
  • Arguments specified in the command when an error occurs during command execution