jbs_start.cluster (UNIX only)
- Organization of this page
Function
In a cluster system, the jbs_start.cluster command starts JP1/Base (the event service, process management functions including the user management function, and the log-file trap management daemon). To execute this command, you must first register it in your cluster software.
The following commands are executed within this command:
-
jevstart logical-host-name
-
jbs_spmd -h logical-host-name
Format
jbs_start.cluster logical-host-name
Required execution permission
Superuser or JP1/Base administrator permission
Command directory
/etc/opt/jp1base/
Arguments
logical-host-name
Specify the logical host for which you want to execute this command.
Notes
-
After issuing the startup request to the JP1/Base processes, this command ends with the return value 0. To verify the proper state of the processes, after the jbs_start.cluster command has finished, use the jbs_spmd_status command.
-
In Linux, the maximum size for core file dumps is sometimes set to 0 by default. In this case, the system will not output a core dump file. To avoid this issue, the jbs_start and jbs_start.cluster scripts contain the following standard setting:
if [ 'uname' = Linux ]; then ulimit -c unlimited fi
If this setting contravenes the security policy of the system on which the script is executed, comment it out by placing a hash mark (#) at the beginning of each line.
#if [ 'uname' = Linux ]; then #ulimit -c unlimited #fi
This invalidates the setting. However, it also means that the system will not create a core dump file when an event such as a segmentation fault or bus error that would usually trigger a core dump occurs in the JP1/Base process, denying you information that could be used to investigate the cause.
-
When you execute the jbs_start.cluster command, a message might appear indicating that the log file trap management daemon is already running. You can prevent this message from appearing by modifying the jbs_start.cluster script. Note that you do not need to modify the script if you performed a new installation of JP1/Base version 10-00 or later.
The jbs_start.cluster script contains the following setting:
## Start services echo "Please wait a minutes, now starting JP1/Base..." if [ "$LHHOST" ]; then /opt/jp1base/bin/jevstart ${LHHOST} /opt/jp1base/bin/jevlogdstart else /opt/jp1base/bin/jevstart /opt/jp1base/bin/jevlogdstart fi
Change the script as follows:
## Start services echo "Please wait a minutes, now starting JP1/Base..." if [ "$LHHOST" ]; then /opt/jp1base/bin/jevstart ${LHHOST} /opt/jp1base/bin/jevlogdstat >/dev/null 2>/dev/null if [ $? -ne 0 ]; then /opt/jp1base/bin/jevlogdstart fi else /opt/jp1base/bin/jevstart /opt/jp1base/bin/jevlogdstart fi
-
If you execute the jbs_start.cluster command as a remote shell command, you must terminate the standard input, standard output, and standard error output by assigning /dev/null to those beforehand. The remote shell command might not terminate after JP/Base processes started.
-
In the following cases, you must specify Japanese as the language of the LANG environment variable in the automatic start script:
-
Japanese is specified in the event filter of the forwarding settings file (forward).
-
Japanese is specified in the lpszFilter parameter of the JP1 event acquisition function (JevGetOpen) in the user program.
-
Japanese is specified in various JP1/IM filters in JP1/IM#.
#: For detailed conditions of servers that require language specification, see the JP1/Integrated Management 2 - Manager Overview and System Design Guide.
-
The automatic start script for logical hosts (jbs_start.cluster) performs an unsetup of the locale settings of LC_ALL, LC_COLLATE, LC_CTYPE, LC_MONETARY, LC_NUMERIC, LC_TIME, and LC_MESSAGES by default.
-
Return values
0 |
Normal end |
1 |
More than one argument was specified. |