18.4.2 How to collect data (in UNIX)
- Organization of this subsection
(1) Execute the data collection tool
Execute the data collection tool (jbs_log.sh).
By executing jbs_log.sh, you can collect the data needed for looking into a JP1/Base error on that host.
The amount of data collected varies greatly depending on your operating environment. Before executing the data collection tool, estimate the amount of data as follows, and make sure you have sufficient disk space.
- Data size when a physical host is specified in jbs_log.sh
-
If you specify a physical host (by omitting the -h option) in the jbs_log.sh command, use the following equation to estimate how much data will be collected about JP1/Base and the computer environment:
Data size = 3 + a + b + (60 x c) MB
- a
-
Size of all the files in /var/opt/jp1base/ (maximum 508 MB#1, #2)
- b
-
Size of the core files under / (only if output)
- c
-
Number of core files under /, and in the /var/opt/jp1base/ or /opt/jp1base/ directory
#1: An extra 254 MB is required if you are running JP1/IM - Manager on the same host.
#2: This is the default value. This value increases if you change the size of the event database. For details on estimating the maximum size, see the Release Notes.
- Data size when a logical host is specified in jbs_log.sh
-
If you specify a logical host in the jbs_log.sh command, use the following equation to estimate how much data will be collected about JP1/Base and the computer environment:
Data size = 3 + a + b + (60 x c) + d + e (MB)
- a
-
Size of all the files in /var/opt/jp1base/ (maximum 83 MB#1, #2)
- b
-
Size of the core files under / (only if output)
- c
-
Number of core files under /, and in the /var/opt/jp1base/ or /opt/jp1base/ directory
- d
-
Data size of shared-directory/jp1base/log/ (maximum 45 MB#1)
- e
-
Data size of shared-directory/event/ (maximum 55 MB#2)
#1: An extra 142 MB is required if you are running JP1/IM - Manager on the same host.
#2: This is the default value. This value increases if you change the size of the event database. For details on estimating the maximum size, see the Release Notes.
You can check the size of each folder by executing the du command.
For details on estimating the maximum disk space requirements of each folder, see the Release Notes.
An example of jbs_log.sh execution is shown below.
jbs_log.sh -f output-file-name
The jbs_log.sh command provides options for excluding specific files, such as command execution logs (ISAM) and event database files. For details, see jbs_log.sh (UNIX only) in 15. Commands.
(2) Check the status of the process
The following table lists the processes displayed when you execute the ps command. In UNIX, by executing the data collection tool (jbs_log.sh), you can collect ps command execution results in addition to the other data.
The value in parentheses in the table indicates the number of processes that can be executed simultaneously.
|
Parent process name |
Function |
Child process name |
Function |
|---|---|---|---|
|
hntr2mon (1) |
Hitachi Network Objectplaza Trace Library (HNTRLib2) |
-- |
-- |
|
jbs_spmd (1) |
Process management#1 |
jbssessionmgr (1) |
Authentication server#1, #4 This process exists only on the host that is set as the authentication server. The displayed name is jbssessionmgr when the jbs_spmd_status command is executed. |
|
jbsroute (1 to 9) |
Configuration management#1, #4 The displayed name is jbsroute when the jbs_spmd_status command is executed. |
||
|
jcocmd (1) jcocmdexe (1) jcocmdapi (number of windows where commands are executed #2 + 1 (when JP1/IM - Manager has been installed)) jcocmdcmc (0 to the number of commands#3) |
Command execution#1, #4 The displayed name is jcocmd when the jbs_spmd_status command is executed. |
||
|
jbsplugind#5 |
Plug-in service#1, #4 The displayed name is jbsplugin when the jbs_spmd_status command is executed. |
||
|
jbshcd (1) |
Health check (for local host monitoring)#1, #4 The displayed name is jbshcd when the jbs_spmd_status command is executed. |
||
|
jbshchostd (1) |
Health check (for remote host monitoring)#1, #4 The displayed name is jbshchostd when the jbs_spmd_status command is executed. |
||
|
jbssrvmgr (1) |
Service management control function#1,#4 The displayed name is jbssrvmgr when the jbs_spmd_status command is executed. |
||
|
jbslcact (1) |
Local action function#1,#4 The displayed name is jbslcact when the jbs_spmd_status command is executed. |
||
|
jbscomd (1) jbscomd_api (1) jbscomd_ses (1) jbscomd_snd (1) jbscomd_rcv (1) |
Inter-process communication#1,#4 The displayed name is jbscomd when the jbs_spmd_status command is executed. |
||
|
jevservice (1) |
Event service#1, #6 |
jevservice (5 to 9,999) |
Event service |
|
jevlogd (1 to 2) |
Log file trap |
jelparentim (0 to the number of times the jevlogstart command is executed) |
Log file trap The jelchildim process is generated for each file to be monitored for each jelparentim. When the jevlogstop command is executed, the jelparentim process disappears. |
- Legend:
-
--: None
#1: The maximum number of processes that can be executed simultaneously with the indicated process is the calculation result of the following format when multiple logical hosts operate on one physical host in the cluster system or when one logical host and one physical host are started at the same time: (number-of-logical-hosts + 1) x number-of-processes
#2: The number of Execute Command windows opened by the connected JP1/IM - View. The number of processes increases as the number of open windows increases. When you close an Execute Command window, the corresponding process disappears.
#3: The number of the remote commands and automated actions executed using JP1/IM - Manager. A process is generated for each command. When processing finishes, the process disappears. If you execute commands successively, multiple processes might be generated.
#4: You can use the jbs_spmd_status command to check the status of these processes. If the processes are running normally, the jbs_spmd_status command returns the following information.
-
If an authentication server has been set:
jbssessionmgr
jbsroute
jcocmd
jbsplugin
jbshcd
jbshchostd
jbssrvmgr
jbslcact
jbscomd
-
If an authentication server has not been set:
jbsroute
jcocmd
jbsplugin
jbshcd
jbshchostd
jbssrvmgr
jbslcact
jbscomd
#5: The process name displayed by the ps-el command is jbsplugin.
#6: The status of these processes can be checked with the jevstat command. Executing the jevstat command when the processes are running normally displays the following string:
jevservice
(3) Check the operation data
If a problem occurs, check the operation data and record it. You need to check the following information:
-
Details of the operation
-
Time the error occurred
-
Machine configuration (version of each OS, host name, configuration of JP1/IM - Manager)
-
Whether the error occurs repeatedly under the same conditions
-
User name used to log in from JP1/IM - View
(4) Collect the error information on the screen
If an error is displayed on the screen, also collect that information. Collect hard copies of the following:
-
The error dialog boxes
In addition, copy the contents of the details if the dialog box contains a Details button.