Job Management Partner 1/Automatic Job Management System 3 Command Reference 1

jajs_killall.cluster (UNIX only)

Format

jajs_killall.cluster
     [logical-host-name]

Description

Forcibly terminates the JP1/AJS3 process during operation on a cluster system.

An attempt to execute the jajs_stop.cluster command at a failover may fail to stop the process, sometimes resulting in an unsuccessful failover. This command should be executed only when you want to forcibly stop a process that will not stop.

Execution privileges

Superuser privilege

Arguments

logical-host-name

Specify the name of the logical host that is set in JP1/Base.

You can specify a character string of 1 to 15 bytes.

By default, the system assumes the host name specified in the JP1_HOSTNAME environment variable. (The first 15 bytes of the logical host name are valid.) With this option omitted, the command terminate abnormally if the JP1_HOSTNAME environment variable contains nothing.

Notes

This command applies only to UNIX. You cannot use this command in Windows.

This command should be executed only when an attempt to execute the jajs_stop.cluster command fails to stop the process.
When a request to forcibly stop the JP1/AJS3 process is issued, the shared disk might be disconnected from JP1/AJS3. If disconnected, the shared disk becomes inaccessible. If this situation occurs, you need to modify the shell script for forcibly stopping JP1/AJS3 (jajs_killall.cluster) so that the log output destination is a physical disk. The following shows how you can modify the shell script.
Before modification:
LOGDIR=`jbsgetcnf -h "$JP1_HOSTNAME" | grep '^"JP1AJS2_LOGDIR"=' | \
        sed -e 's/^[^=]*=//' -e 's/^"//' -e 's/"$//'`
COMNAME=`basename "$0"`
 
After modification:
LOGDIR="/var/opt/jp1ajs2/log/schedule"  # path on a physical disk
COMNAME=`basename "$0"`

Return values

0 Normal end

Values other than 0 Abnormal end

Additional informations

This command is stored in the /etc/opt/jp1ajs2/ path.

This command forcibly terminates the process on a logical host basis. Of the JP1/AJS3 processes, however, those listed below are out of the target of forced termination.

Process name Description

jp1mqsup MQ monitor process at the time of TP1/Message Queue linkage
SUP under TP1 control

jpocwtmqmaii MQ monitor process at the time of MQSeries linkage

The processes in the table do not need a failover together with the JP1/AJS3 service. You can terminate a process, as necessary.

Log information is output to files. The target is log files of five generations.
The log file's directory and name are given below.

Log file storing directory:

shared-directory/jp1ajs2/log/schedule

(If multiple scheduler services are set up, the directory name is as follows: shared-directory/jp1ajs2/log/schedule/scheduler-service-name)

Log file name:

jajs_killall.cluster_logical-host-name.{1|2|3|4|5}.log

Process name	Description
jp1mqsup	MQ monitor process at the time of TP1/Message Queue linkage SUP under TP1 control
jpocwtmqmaii	MQ monitor process at the time of MQSeries linkage

[Contents][Back][Next]

[Trademarks]