Top
Systemwalker Operation Manager Q & A Guide

2.6 Q: What Points Should I Note when Starting a Large Number of Jobs on Multiple Subsystems?

Applicable versions and levels

A

If you create multiple subsystem environments by using the createsubsystem command and you start a large number of jobs, performance may deteriorate immediately after you create the multi-subsystem environments because they are created on a single disk.

When operating a large number of jobs, preferably distribute the Jobscheduler database directory and the Job Execution Control spool directory to different disks for each subsystem.

The following description explains how to distribute the database directory and spool directory to the desired disks.

Directory storage location

When you create a subsystem environment, the database directory and spool directory are stored in the following locations:

Database directory

Windows versions

<installation directory>
\MpWalker.JM\mpjobsch\jobdbn

Solaris versions

/var/opt/FJSVJOBSC/JOBDBn

HP-UX versions

/opt/FHPJOBSCH/db/JOBDBn

Linux versions

/var/opt/FJSVJOBSC/JOBDBn

AIX versions

/usr/FAIXJOBSC/db/JOBDBn

Spool directory

Windows versions

<installation directory>
\MpWalker.JM\mpmjessv\mjesn\mjespool

Solaris versions

/var/opt/FJSVMJS/var/spool/mjes/mjesn

HP-UX versions

/opt/FHPMJS/var/spool/mjes/mjesn

Linux versions

/var/opt/FJSVMJS/var/spool/mjes/mjesn

AIX versions

/opt/FAIXMJS/var/spool/mjes/mjesn

In "JOBDBn", "jobdbn", and "mjesn" above, "n" corresponds to the subsystem number from 1 to 9. When the subsystem number is 0, the database directory and spool directory will be stored in the same directory as that used when multiple subsystems are not started.

Distributing directories [UNIX versions]

You can distribute the database directory and spool directory in a UNIX environment by moving them to any disk and creating symbolic links in the storage location used when the subsystem environment is created.

The following provides an example of moving the database directory to "/disk1" and the spool directory to "/disk2" in the Solaris environment.

1) Moving to disks

Move the database directory and spool directory to different disks.

# mv /var/opt/FJSVJOBSC/JOBDBn /disk1/JOBDBn
# mv /var/opt/FJSVMJS/var/spool/mjes/mjesn /disk2/mjesn

2) Creating symbolic links

Create symbolic links as in the following example:

# ln -s /disk1/JOBDBn /var/opt/FJSVJOBSC/JOBDBn
# ln -s /disk2/mjesn /var/spool/mjes/mjesn

Note: Replace "n" in the example above with the actual subsystem number.

Distributing directories [Windows versions]

To distribute the database directory and spool directory in the Windows environment, open the Systemwalker Operation Manager Environment Setup window. In the tab that is displayed, specify the database directory and spool directory of the distribution destination disks.

Database directory:

From the Systemwalker Operation Manager Environment Setup window, open the Define Jobscheduler Startup Parameters window and click the Database tab. In this window, change the default drive name to a drive name on another disk.

Spool directory:

From the Systemwalker Operation Manager Environment Setup window, open the Define Operating Information window.

On V10.0L21 or earlier, open the Define Operating Information and click the Options tab. For Spool directory, specify the directory of a different disk.

On V11.0L10 or later, open the Define Operating Information window and click the Cluster settings tab. For Spool directory, specify the directory of a different disk.