2.8.1 Defining Startup Parameters

The following explains how to define startup parameters of Jobscheduler services and daemons.

If you have changed the startup parameter definition, you must restart Jobscheduler services and daemons to make your change valid.

Outline

The following seven (7) sheets are included in the Define Jobscheduler Startup Parameters window to define startup parameters of Jobscheduler services and daemons.

Windows used for startup parameter definition:

Definition procedure

Display the Define Jobscheduler Startup Parameters window.
Click Startup parameter in the Systemwalker Operation Manager Environment Setup window, and the Define Jobscheduler Startup Parameters window will appear.
Select the desired sheet (Database, Use Function1, Use Function2, Event output, Mail, Output file, or Test mode sheet).
To define the database information (for database directory or log file setup), select the Database sheet.
To enable the Job net startup API, Message waiting, Verification at error occurrence, and Status option functions, select the Use Function1 sheet.
To set the Halt recreation of startup days, warn at job net registration count, and Limit the number of connected clients options, select the Use Function2 sheet.
To set up information relating to event output (the output conditions for job net execution histories, links to Systemwalker Centric Manager and delay monitoring management for jobs and job nets), select the Event output sheet.
To set the mail waiting conditions, select the Mail sheet.
To set output file options of jobs, select the Output file sheet.
To use the Test Mode operation (in virtual time), select the Test mode sheet.
Register the startup parameter information.
After you have set the startup parameters on each sheet, click OK to store them.

Database sheet in the Define Jobscheduler Startup Parameters window

Directory:

[Windows]

Specify the directory to store the database that the Jobscheduler services use in the format of "drive-name directory-name" in no more than 254 bytes. The "drive-name" can be a fixed disk drive only. The directory cannot be specified during operation of a Jobscheduler service. This parameter is always required. The following initial values are set during system installation.

Systemwalker Operation Manager installation directory \MpWalker.JM\mpjobsch\jobdb

The standard output (stdout) and standard error output (stderr) for each job registered with the Jobscheduler will be stored in a file in the directory specified here, with the file name "<project name>_<job net name>_cccccccccccccccc.log", where "cccccccccccccccc" is an arbitrary string.

When connected to the server for multiple-subsystem operations, the initial values of database storage directory are stored in the following file.

Systemwalker Operation Manager installation directory\MpWalker.JM\mpjobsch\jobdbn

Where, "n" of "jobdbn" is a subsystem number (1 to 9). If subsystem 0 is specified, the information is stored in the same directory as that is used when multi-subsystems are NOT started.

[UNIX]

You cannot specify a database directory. The database used by Jobscheduler daemons is always stored in the following location. You must create a symbolic link to change it.

Solaris	/var/opt/FJSVJOBSC
HP-UX version	/opt/FHPJOBSCH/db
AIX version	/usr/FAIXJOBSC/db
Linux	/var/opt/FJSVJOBSC

When connected to the server for multiple-subsystem operations, the database is always stored in the following directory independent from destination subsystems.

Solaris	/var/opt/FJSVJOBSC/JOBDBn
HP-UX version	/opt/FHPJOBSCH/db/JOBDBn
AIX version	/usr/FAIXJOBSC/db/JOBDBn
Linux	/var/opt/FJSVJOBSC/JOBDBn

where, "n" of "JOBDBn" is a subsystem number (1 to 9). If subsystem 0 is specified, the information is stored in the same directory as that is used when multi-subsystems are NOT started.

The standard output (stdout) and standard error output (stderr) for each job registered with the Jobscheduler will be stored in a file in the fixed database directory shown above, with the file name "<project name>_<job net name>_nn.log", where nn is an arbitrary string.

Change option [Windows]:

This option allows you to change the directory which stores the database used by Jobscheduler services. This option cannot be specified when the Jobscheduler service is running. These options are explained below.

Delete destination database:: Deletes a database if it exists in the specified directory.

Copy current database to destination:: Copies the current database to the specified directory. If the database already exists in the specified directory, it is overwritten.

Delete current database:: Deletes the current database.

Log file size:: Specify the file size (from 1M to 99M bytes) to be used when the Jobscheduler log file is switched. The initial value is 3M bytes. If the size of Jobscheduler log file exceeds the specified value, three files (jobdb1.log, jobdb2.log, and jobdb3.log) are switched in order.

Note

Intensified log file access

The log file is accessed by the following Systemwalker Operation Manager operations:

Display of the Job History window or the Job Net History window
Startup of a job net or job

If startup of a job net or job is triggered by the display of the Job Net History window or the Job History window, access to the log file may become intensified, which may result in slower display performance or startup performance. Deterioration in startup performance may lead to the job net or job not starting at the scheduled time.

Therefore, refrain from displaying the Job Net History window or the Job History window while a job net or job is running.

Information

Setting the log file size

Estimate the total log file size by referring to the following formula. As a guide, if the log file size estimated exceeds 20MB (Note), set the estimated size after thoroughly examining the time required to display the Job History window and impacts on responses to other Operation Manager clients.

Note)

In addition to the operations listed in the "Concentrated access of the log file" section above, responses to other Operation Manager clients may also be affected by the log file size.

If the estimated log file size exceeds the guideline of 20 MB, carefully consider any possible impact when setting this value.

The correct value of the log file size varies depending on computer performance or network performance. The log file size shown here should be considered as just a reference value.

(Size of daily job net start log + Size of daily job start log + Size of daily job termination log + Size of daily job net termination log) x Number of keepdays / 2

Note that immediately after the log file is switched, etc, only 1 file can be referenced by the Operation Manager client and jobschprint command. In cases where there is a requirement to constantly reference the file for the number of keepdays, make your estimate without including the '/ 2' part of the estimation formula cited above.

Each log size can be estimated by referring to the following.

When the number of job net starts varies every day, it is recommended to estimate the job net log size based on the largest number of job nets started a day. The lengths of each name or comment should be the maximum value among the actually registered values. If any value is unknown, use the possible maximum value provided inside ( ) below.

Job net start log	50 + S + P + J + N
Job start log	50 + S + P + J + N + j + n
Job termination log	75 + S + P + J + N + j + n
Job net termination log	75 + S + P + J + N

S: Local server name length
P: ProjectName length (50 bytes)
J: JobNetName length (50 bytes)
N: JobNetComment length (100 bytes)
j: JobName length (Job command length if JobName is not specified(300 bytes)) (64 bytes)
n: Comment length (64 bytes)

[Estimation examples]

Examples of size estimations will be shown assuming the following environment.

Number of job nets started and terminated a day: 1000
Number of jobs started and terminated a day: 4000
Server name: 6 bytes
ProjectName: Max. 12 bytes
JobNetName: Max. 8 bytes
JobNetComment: Max. 16 bytes
JobName: Max. 10 bytes
Comment: Max. 20 bytes
Keepdays: 30 days

Job net start log: 50 + 6(S) + 12(P) + 8(J) + 16(N) = 92

Job start log: 50 + 6(S) + 12(P) + 8(J) + 16(N) + 10(j) + 20(n) = 122

Job termination log: 75 + 6(S) + 12(P) + 8(J) + 16(N) + 10(j) + 20(n) = 147

Job net termination log: 75 + 6(S) + 12(P) + 8(J) + 16(N) = 117

Log file size: (92 x 1,000 + 122 x 4,000 + 147 x 4,000 + 117 x 1,000) x 30 / 2 = 19,275,000 (Approx. 18.4 MB)

In this example, round up 18.4 and register 19 MB as the total log size.

Use function1 sheet in the Define Jobscheduler Startup Parameters window

Job net startup API [Windows]:

Select Use job net startup API to start job nets using a job net startup API.

This option is not selected by default. If omitted, you cannot start up a job net using the job net startup API. However, you can use the job net startup time modification API, the job net operation API and group operation API independent from this selection.

Message waiting [Windows]:

Select Message waiting to start job nets when a specific message (an event) is output to the event log. This option is enabled by default. If this option is unselected, no message is waited even if a message has been registered on the message table.

There are three types of event logs (applications, security and system event logs). When you specify the Message waiting option, you can select one or more of these logs to be monitored. If omitted, Application is only selected. To generate a message event using the message event generation command (jobschmsgevent), select Application. This option is not selected by default. The next operation starts regardless of the result of the previous execution.

Verification at error occurrence:: Specify this option to suppress the execution of the next job when a job net or group has terminated abnormally or been terminated forcibly. Details relating to this setting can be specified via the Advanced Settings for Verification window, which is displayed by clicking the Advanced Settings button.

Status Option:: Select this option to indicate the normal termination or Pseudo-normal termination of each job, job net and group status display. This option is NOT selected by default. If not selected, the "Pseudo-normal termination" is indicated as "normal termination."

Advanced Settings for Verification window

This window is displayed when Validate confirmation operation is specified in the Use function tab of the Define Jobscheduler Startup Parameters window and then the Advanced Settings button is clicked.

Advanced Settings for Verification:

Use this option to suppress the next startup until you confirm after an abnormal or forcible termination of a job net or group. You can select Check cancelled only when you have specified at least one of other two options.

All of these options are selected by default.

Use function2 sheet in the Define Jobscheduler Startup Parameters window

Halt recreation of startup days:: Select this option NOT to recreate a job net execution day if you have changed the holiday calendar or the schedule pattern. This option is NOT selected by default.

Warn at job net registration count:

Select this option to output an alarm message if the number of a job net to be registered for a single project exceeds the limit. In the Number of registrations per project to trigger warning, you can specify the limit number of job net. If the user specified job net count exceeds this limit, the following alarm is output.

The job net count exceeded the limit.

The registration limit can be an integer of 1 to 99999. No alarm is output by default. If the message output is specified in the Systemwalker Operation Manager SE version, the Number of registrations per project to trigger warning is set to 255 by default. In the Systemwalker Operation Manager EE version, it is set to 4000 by default.

Limit the number of connected clients:

Specifies the maximum number of clients that can be connected to the Jobscheduler server. The servers to be monitored by the Multi-server monitoring client and the Print Jobscheduler Info clients are also included. If the clients exceeding the limit attempt to connect to the Jobscheduler server, such connection fails and the following error message appears when the Jobscheduler is selected during job selection from the Systemwalker Operation Manager client.

Too many clients attempt to connect to the Jobscheduler server.

One to 62 (Note) clients can be connected to the Jobscheduler server. The default is unlimited.

If the number of connected clients exceeds the limit during connection by multi-server monitoring client, the client displays the "Access denied" server status.

Note) The maximum value may vary depending on the operating system you use.

Event output sheet in the Define Jobscheduler Startup Parameters window

Event output:

To output error notifications to the event log or syslog, select the Output error notifications to syslog/event log check box.

Details relating to output are specified via the Advanced Settings for Event Output Details window, which is displayed by clicking the Advanced Settings button.

Link to Systemwalker Centric Manager:

Select this option to link to Systemwalker Centric Manager.

If you check the Automatic notification/handling box, the abnormal events are displayed in the Monitor window of Systemwalker Centric Manager when job net of Systemwalker Operation Manager have terminated abnormally. Also, when a user restarts the abnormally terminated job net, those events are set to the Resolved in the Monitor window. This box is UNCHECKED and the "Automatic notification/handling" is NOT executed by default.

Console Option [UNIX]:

If checked, a message is also output to the console or SYSLOG when it is output to the log file. The message facility level is user.info.

This is unchecked by default. No message is output. Consider that output to SYSLOG increases if you specify this option.

Advanced Settings for Event Output window

This window is displayed when the Output error notifications to syslog/event log check box is selected in the Event output tab of the Define Jobscheduler Startup Parameters window and then the Advanced Settings button is clicked.

[Windows]

Job net status:

Specify the output condition (Job Net Status) to output the job net execution history to the event log. You can select one or more of eight job net states (Started, Completed, Abended, Canceled, Closed, Refused, Skipped and Pseudo-normal). For example, select Abended to output an event that signifies that a job net terminated abnormally to the event log. If you want to output an event indicating abnormal termination resulting from execution suspension to the event log, select Closed instead of Abended.

The following are selected by default:

Abended:
Skipped:
Canceled:
Closed:
Refused:

The following are event ID and text details that can be output to the event log depending on the job net status. Using this event log, you can monitor batch jobs in Systemwalker Centric Manager.

Started:: 9000/The job net has started. JobNetComment=a JobNetName=b ProjectNam=c

Completed:: 9001/The job net has normal ended. JobNetComment=a JobNetName=b Code=c ProjectName=d

Abended:: 9002/The job net has abended JobNetComment=a JobNetName=b Code=c ProjectName=d

Canceled:: 9002/The job net has abended JobNetComment=a JobNetName=b Code=256 ProjectName=c

Closed:: 9002/The job net has abnormal ended. JobNetComment=a JobNetName=b Code=239 ProjectName=c

Refused:: 9003/The job net has refused. JobNetComment=a JobNetName=b ProjectName=c

Skipped:: 9004/The job net has skipped. JobNetComment=a JobNetName=b ProjectName=c

Pseudo-normal:: 9005/The job net has pseudo-normal ended. JobNetComment=a JobNetName=b Code=c ProjectName=d

The followings are the facility levels of the messages that are output to the SYSLOG according to the job net status.

Started:: Information

Completed:: Information

Abended:: Error

Canceled:: Error

Closed:: Error

Refused:: Warning

Skipped:: Warning

Pseudo-normal:: Information

[UNIX]

Job net status:

Specifies the output conditions (job net status) to be used when the job net history is output to the syslog. You can select one or more of eight job net states (Started, Completed, Abended, Canceled, Closed, Refused, Skipped and Pseudo-normal). For example, you can select Abended to output an abnormal termination of job nets to the syslog. If you want to output an event indicating abnormal termination resulting from execution suspension to the event log, select Closed instead of Abended.

The following are selected by default:

Abended
Skipped
Canceled:
Closed:
Refused:

The ID and text output to syslog according to the status of a job net are shown below. Note that this syslog can be used by Systemwalker Centric Manager to monitor batch business.

Started:: 328/The job net has started. JobNetComment=a JobNetName=b ProjectName=c

Completed:: 329/The job net has normal ended. JobNetComment=a JobNetName=b Code=c ProjectName=d

Abended:: 330/The job net has abnormal ended. JobNetComment=a JobNetName=b Code=c ProjectName=d

Canceled:: 330/The job net has abnormal ended. JobNetComment=a JobNetName=b Code=256 ProjectName=c

Closed:: 330/The job net has abnormal ended. JobNetComment=a JobNetName=b Code=239 ProjectName=c

Refused:: 331/The job net has refused. JobNetComment=a JobNetName=b ProjectName=c

Skipped:: 332/The job net has skipped. JobNetComment=a JobNetName=b ProjectName=c

Pseudo-normal:: 333/The job net has pseudo-normal ended. JobNetComment=a JobNetName=b Code=c ProjectName=d

The followings are the facility levels of the messages that are output to the SYSLOG according to the job net status.

Started:: user.info

Completed:: user.info

Abended:: user.er

Canceled:: user.err

Closed:: user.err

Refused:: user.warning

Skipped:: user.warning

Pseudo-normal:: user.info

For details on the job net status, refer to "Job net status" in the Systemwalker Operation Manager User's Guide.

Watching delay of job/job net:

Use this option to monitor the scheduled job net startup time, the job's estimated processing time, and the scheduled job net end time. These options are explained below.

Notify when job net is not started even after the specified start time is lapsed:

Check this option to output a notification event or a message to the event log or syslog if a job is not executed even after the job net startup time has elapsed.

This option is selected by default.

Notify when job is not terminated even after the specified time is lapsed:

Select this option to output a notification event or message to the event log or syslog if a job fails to end after the specified time has elapsed.

This option is selected by default.

Notify when job net is not ended even after the specified end time is lapsed:

Check this option to output a notification event or message to the event log or syslog if a job net does not end even though the scheduled termination time has elapsed (the status of the job net is "Executing Jobs", "Warning" or "Start delayed").

This option is selected by default.

The following explains the event IDs, message IDs and message text that are output to the event log or syslog in these states.

All of the messages output to the event log become the "Warning" messages.

Facility levels of messages to syslog are as follows. [UNIX]

If a job net is not started even after the specified startup time has elapsed or a job net is not ended after the specified end time has elapsed:: user.warning

If a job does not end even after the job's estimated processing time has elapsed:: daemon.warning

Using these event logs or syslog, you can monitor batch jobs in Systemwalker Centric Manager.

If a job net is not executed even after the scheduled startup time has elapsed:

[Windows]

4305/XXXX YYYY did not start at scheduled time.

[UNIX]

310/XXXX YYYY did not start at scheduled time.

XXXX: Project name

YYYY: Job net name

If a job does not end even after the job's estimated processing time has elapsed:

[Windows]

9006/It doesn't end even if job name [jobname] lapses by end plan time [time] minutes.

[UNIX]

10114/It doesn't end even if job name (jobname) lapses by end plan time (time) seconds.(Project Name=project name, Job Net=jobnet name)

job name: Job name

time: The estimated processing time that is defined

project name: Project name

jobnet name: Job net name

If a job net does not end even though the scheduled end time has elapsed:

[Windows]

4306/XXXX YYYY did not finish before Estimated end time.

[UNIX]

311/XXXX YYYY did not finish before Estimated end time.

XXXX: Project name

YYYY: Job net name

Note

Notification events or messages when a job net is not executed even after the scheduled startup time has elapsed

No notification event or message is output for a job net in a group that has an invalid startup time.
No notification event or message is output for a job net that is paused or disabled.
Validate job net confirmation is specified in the Advanced Settings for Verification window and no notification event or message is output for a job net that terminated abnormally
If job nets belonging to a group have valid startup time, notification events or messages are output at the startup time specified in the Job Net Properties window rather than the "Scheduled Day and Time" of job nets shown in the Job Net Lists window or Job Lists window.

Mail sheet in the Define Jobscheduler Startup Parameters window

This window is for setting up the mail waiting conditions when email is used.

User ID:: For the mail waiting condition, specify the user ID of the user who will receive the mail, using no more than 20 bytes. Mail sent to the specified user will be queued.
For systems operating multiple subsystems, the user ID must be unique in each subsystem.

Password:: Specify the password for the user ID that was specified in the User ID field.

Mail server name:: Specifies a host name of the mail server (POP3 server) that will receive e-mails, using up to 128 characters.

Definition procedure

To start job nets triggered by email reception, specify the user ID, password, and name of the mail server as mail waiting conditions.
On the mail server, create a user with the user ID specified in Step 1, and specify a password for the user ID.

Output file sheet in the Define Jobscheduler Startup Parameters window

[Windows]

Deleting the job output file:

Use this option to delete the job output file created during job execution.

This is checked by default, and the job output file is deleted.

[UNIX]

Deleting the job output file:

Use this option to delete the job output file which is created when a job stored in the job net having "Job Execution Control" attributes is executed.

This is checked by default, and the job output file is deleted.

The following lists the "job output files" which are created by the Job Execution Control functions

jobname.o_job-number (Standard output files)
jobname.e_job-number (Standard error output files)
jobname.1_job-number (Job list files)

However, if the schedule server is a Windows server, Delete job output file has not been selected, and the names of the standard output file and the standard error output file for schedule jobs have not been specified, then the standard error output file (jobname.e_job-number) will not be created and both the standard output and the standard error output will be sent to the standard output file (jobname.o_job-number).

The "job output files" can be output in the following format if a subsystem other than subsystem 0 is used.

jobname.s_subsystem-number.o_job-number (Standard output files)
jobname.s_subsystem-number.e_job-number (Standard error output files)
jobname.s_subsystem-number.1_job-number (Job list files)

Note

If you have UNCHECKED the Deleting the job output file checkbox, you must delete job output files periodically. If not deleted, the system performance may drop or the disk resources may be exhausted.
Job list files are not deleted for jobs for which you specified Standard output file in the Detail information tab of the Add/Change - Job window, even if you selected the Deleting the job output file check box. Periodically delete the job list files. Otherwise, performance may drop or disk resources may be depleted.

Job output files

[Windows]

If Delete job output file is unchecked, the output files are stored in the "work" directory locating under the database directory.

If Delete job output file is checked, the files stored in the "work" directory under the database directory are deleted when the job ends or when the Jobscheduler service starts.

If Delete job output file has not been selected, and if the names of the standard output file and the standard error output file for schedule jobs have not been specified, the standard error output file (jobnname.e_job-number) will not be created and both the standard output and the standard error output will be sent to the standard output file (jobname.o_job-number).

[UNIX]

With a local job
If Delete job output file is unchecked, the output files are stored in the directory you have specified in the Standard information of the job. If no directory is specified, the output files are stored in the home directory of the project owner registering the job (or the user if the user name is specified in Effective user name in the Detail Information sheet in the Add/Change/Monitor - Job window).
If Delete job output file is checked, the output files are stored in the "work" directory locating under the database directory and they are deleted later.
With a network job
- If Before job execution, change to directory specified at job registration is selected:
  The following directory of the schedule server becomes the output destination.
  - If Delete job output file is not selected:
    The output files are stored in the home directory of the project owner (or the user if the user name is specified in Effective user name in the Detail Information sheet in the Add/Change/Monitor - Job window).
  - If Delete job output file is selected:
    The output files are created in the work directory under the database directory but later deleted.
- If Before job execution, change to directory specified at job registration is not selected:
  In the schedule server, the output files are stored in the directory specified in above "a. With a local job."

[Common]

If you have specified the "-o" or "-e" option in the "command name" of jobs, the job files are output to a directory OTHER THAN the "work" directory under the database directory. In such case, the files are output to the specified directory and NOT deleted even if you have checked the Delete job output file box.

Also, for on-demand jobs, output files will not be deleted regardless of the settings for the Delete job output file checkbox.

The system administrator must delete those output files at regular intervals.

Test mode sheet in the Define Jobscheduler Startup Parameters window

You can set a virtual time on a subsystem to be operated in the Test mode. After you have set the virtual time, you must restart the Jobscheduler services or daemons.

Test mode:: Indicates a time difference from the system if the virtual time has been set.

Change the virtual time:

When you set a virtual time, select the Set the virtual time option and set its date and time. It can be up to three (3) years from the present year (this year) to the year after next. (You can set a virtual time before December of 2014 when this year is 2012.)

You cannot set the past date.

To release the virtual time setup, select the Release the virtual time option.

Refer to "Operating in the Test Mode" in the Systemwalker Operation Manager User's Guide for details.