13.5.2 How to collect data during troubleshooting in a Linux environment
- Organization of this subsection
(1) Executing the data collection command
Use the jpcras command to collect the data needed to determine the cause of an error (note that the procedure described below must be executed by an OS user with the root permission).
To execute the data collection command:
-
Log on to the host where the service subject to this data collection is installed.
-
Specify in the jpcras command the data to be collected and the storage directory for the data, and then execute the command.
The following jpcras command stores all the available information in the /tmp/jpc/agt directory:
jpcras /tmp/jpc/agt all all
The data collected by the data collection command can be stored in the specified directory in a compressed format by using either the tar or the compress command. Example file name:
jpcrasYYMMDD.tar.Z
The date is added at the location indicated by YYMMDD.
When you execute the jpcras command, the jpctool service list -id * -host * command is executed internally in order to obtain a list of PFM services and check their activity status. If there is a firewall between the host where the command is executed and the host for the Performance Management system, or if the system configuration is large, it might take time to execute the jpctool service list -id * -host * command. If desirable, you can reduce the command execution time by suppressing execution of the jpctool service list -id * -host * command. To do this, specify 1 in the JPC_COLCTRLNOHOST environment variable.
For details about the jpcras command, see the chapter that describes commands in the manual JP1/Performance Management Reference.
(2) Executing the data collection command (for logical host operation)
Performance Management information for logical host information is on the shared disk and must be collected in both executing and standby systems.
You use the jpcras command to collect information needed to check the cause of a problem. This subsection describes how to execute the data collection command. This procedure must be executed by an OS user with the root user permission.
To execute the data collection command for logical host operation:
-
Mount the shared disk.
Information about the logical host is stored on the shared disk. At the executing node, make sure that the shared disk is mounted and then collect the information.
-
In both executing and standby systems, execute the jpcras command specifying the information to be collected and its storage directory.
The following example stores all the information collectible by the jpcras command in the /tmp/jpc/agt directory:
jpcras /tmp/jpc/agt all all
The information collected by the data collection command is stored in the specified directory in the compressed format specified by the tar and compress commands. The file name is as follows:
jpcrasYYMMDD.tar.Z
YYMMDD is replaced with the date (year, month, date).
Executing the jpcras command without the lhost argument specified collects all the Performance Management information on the physical and logical hosts at the node. If there is a Performance Management in the logical host environment, the log files on the shared disk are acquired.
If the jpcras command is executed at a node where the shared disk is not mounted, files cannot be acquired from the shared disk, but the command terminates normally without resulting in an error.
- Note
-
Execute the data collection command at both executing and standby nodes to collect information. To evaluate an event before and after failover, you need the information for both executing and standby systems.
For details about the jpcras command, see the chapter that describes commands in the manual JP1/Performance Management Reference.
-
Collect information about the cluster software.
This information is needed to determine whether the problem occurred in the cluster software or Performance Management. Collect the information that provides details about control requests, such as startup and termination of Performance Management from the cluster software, and their results.
(3) If the system log information and CCMS Alert Information are not stored in the default locations
Manually collect the storage and management files from the new storage location.
(4) Collecting information about the SAP system
Collect information about the SAP system which is being monitored.
For details about the data to be collected, see 13.4.2 (3) Information about the SAP system.
(5) Checking information about the operation
You should check and save information about the operation being performed when the error occurred; the following is the information that you should check and save:
-
Details of the operation
-
Time the error occurred
-
Machine configuration (the OS version, host name, and PFM - Manager and PFM - Agent configuration, and so on.)
-
Whether the problem is replicable
-
Performance Management user name used during logon, if the user has logged on from PFM - View
(6) Collecting error information
You should collect the following error information:
-
Messages output to the console, if the error occurred during command execution
(7) Other information
You should also collect the following information:
-
Command arguments that were specified, if the error occurred during command execution