jcogencore
- Organization of this page
Function
With the exception of the Central Scope Service (jcsmain), this command outputs dumps in the event of a JP1/IM - Manager process failure. After executing the jcogencore command, you must restart JP1/IM - Manager.
-
In Windows:
Execute this command if the evflow, evtcon, evgen, and jcfmain processes of JP1/IM - Manager have failed.
-
In UNIX:
Execute this command if the evflow, jcamain, evtcon, evgen, and jcfmain processes of JP1/IM - Manager have failed.
To detect failures, use the health check function of JP1/IM - Manager (for details about the health check function, see 7.2 JP1/IM - Manager health check function in the JP1/Integrated Management - Manager Overview and System Design Guide).
When you execute the jcogencore command, it displays a message asking you to choose the processes for which failure data is to be output. Select the processes that have failed. The following table shows the files that are output when the jcogencore command is executed.
OS |
Process name |
Name of output file |
|
---|---|---|---|
Java thread dump |
Core dump |
||
Windows |
evflow |
javacore-process-ID.XXXXXXXXXX.txt |
-- |
jcamain |
-- |
-- |
|
evtcon#1 |
javacore-process-IDID.XXXXXXXXXX.txt |
-- |
|
evgen#1, #2 |
javacore-process-ID.XXXXXXXXXX.txt |
-- |
|
jcfmain |
javacore-process-ID.XXXXXXXXXX.txt |
-- |
|
UNIX |
evflow |
javacore-process-ID.XXXXXXXXXX.txt |
core.evflow |
jcamain |
-- |
core.jcamain |
|
evtcon#1 |
javacore-process-ID.XXXXXXXXXX.txt |
core.java |
|
evgen#1, #2 |
javacore-process-ID.XXXXXXXXXX.txt |
core.evgen |
|
jcfmain |
javacore-process-ID.XXXXXXXXXX.txt |
core.jcfmain core.process-ID.jcfallogtrap |
The output files are stored in the following folders:
- When the process name is not jcfmain
-
- In Windows:
-
Physical host: Console-path\log\
Logical host: shared-disk\jp1cons\log\
- In UNIX:
-
Physical host: /var/opt/jp1cons/log/
Logical host: shared-disk/jp1cons/log/
- When the process name is jcfmain
-
- In Windows:
-
Physical host: Manager-path\log\imcf\
Logical host: shared-disk\jp1imm\log\imcf\
- In UNIX:
-
Physical host: /var/opt/jp1imm/log/imcf
Logical host: shared-disk/jp1imm/log/imcf
In addition to the thread and core dumps that are output, other failure data can be obtained by using the data collection tool.
Once you have executed this command, you must restart JP1/IM - Manager.
-
In Windows:
Physical host: After the command has executed, stop JP1/IM - Manager Service by choosing Control Panel, Administrative Tools, Services, and JP1/IM-Manager Service, and then restart JP1/IM - Manager. After JP1/IM - Manager has restarted, use the jco_spmd_status command to check the process statuses.
Logical host: After the command has executed, stop the JP1/IM-Manager_logical-host-name service by choosing Control Panel, Administrative Tools, and Services, and then restart JP1/IM - Manager. If you use cluster software to monitor the JP1/IM-Manager_logical-host-name service, use the cluster software to either restart the service or trigger failover.
-
In UNIX:
Physical host: After the command has executed, the selected processes are terminated forcibly. Use the jco_stop command to terminate all processes and then restart the processes with the jco_start command. After the processes have restarted, use the jco_spmd_status command to check the process statuses.
Logical host: After the command has executed, the selected processes are forcibly terminated. Use the jco_stop.cluster command to terminate all processes and then restart the processes with the jco_start.cluster command. If you use cluster software to monitor JP1/IM - Manager, use the cluster software to either restart the service or trigger failover.
Format
jcogencore [-h logical-host-name]
Execution permission
In Windows: Administrator permissions (if the Windows UAC feature is enabled, the command is executed from the administrator console)
In UNIX: Superuser permissions
Storage directory
- In Windows:
-
Console-path\bin\
- In UNIX:
-
/opt/jp1cons/bin/
Arguments
- -h logical-host-name
-
When you are operating in a cluster system, this option specifies the logical host name. The command outputs the thread or core dump of the JP1/IM - Manager processes at the specified logical host. If this option is omitted, the logical host name specified in the JP1_HOSTNAME environment variable is assumed. If the JP1_HOSTNAME environment variable is not specified, the physical host name is assumed. If you are not using a cluster system, specification of this option is not needed.
Notes
-
When you execute the jcogencore command in UNIX, the processes are terminated forcibly after the dump has been output. Execute this command only in the event of process hang-up. You can use health checking to detect process hang-ups.
Take precautions when you execute the jcogencore command during cluster system operation.
-
Before you execute this command in UNIX, first check the available disk space on your machine. If you output a core dump for five processes (jcfmain excluded), the total size of the core dump might be as much as 8,419 megabytes.
In addition, if you output a core dump for jcfmain, the total size of the core dump might be as much as of 560 + 230 × number-of-jcfallogtrap-processes megabytes.
-
If multiple processes have failed, execute the following commands in the order at which they are listed:
In Windows: Event Console Service (evtcon), Event Base Service (evflow)
In UNIX: Event Console Service (evtcon), Automatic Action Service (jcamain), Event Base Service (evflow)
You can execute the command on the correlation event generation function (evgen) and the IM configuration management service (jcfmain) in any order because there are no dependencies with other processes.
-
In UNIX, the jcogencore command might not generate core dump files if the operating system is configured to block core dump files from being generated.
For details about the settings for core dump files, see 2.17.4 Specifying settings for handling JP1/IM - Manager failures (for UNIX) in the JP1/Integrated Management - Manager Configuration Guide.
Return values
0 |
Normal termination |
1 |
Option or argument analysis error |
2 |
Process check error |
3 |
Logical host error |
4 |
Execution permission error (Windows only) |
5 |
Pipe creation error (Windows only) |
6 |
Thread dump output processing error |
10 |
Other error |
Example 1
Execute the command because a hang-up occurred in the Event Console Service (evtcon) process on the physical host in Windows:
jcogencore
Example 2
Execute the command because a hang-up occurred in the Event Console Service (evtcon) process on the logical host hostA in Windows:
jcogencore -h hostA
Example 3
Execute the command because a hang-up occurred in the Event Console Service (evtcon) process on the physical host in UNIX:
/opt/jp1cons/bin/jcogencore
Example output
When a hang-up occurred in the Event Console Service (evtcon) process on the physical host in UNIX, and core and thread dumps were output:
ProcessName PID [1] : evflow 1234 [2] : jcamain 94320 [3] : evtcon 3333 [4] : evgen 65654 [6] : jcfmain 3316 [7] : Exit KAVB8427-I When outputting dumps for the three processes evflow, jcamain, and evtcon at the same time, output the dumps in order of evtcon, jcamain, and evflow. KAVB8417-I Please enter a number for the process to output the core dump file [1-7]:3 KAVB8414-I The thread dump output request has been sent. KAVB8407-I When the core dump is output, evtcon will stop. Is this OK? (y/n):y KAVB8406-I The core dump file will be output. KAVB8416-I The core dump file has been output.