Hitachi

JP1 Version 12 JP1/Automatic Job Management System 3 System Design (Configuration) Guide


3.4.1 Estimating the size of scheduler log files

This subsection explains how to estimate the size of scheduler log files. The execution logs and operation logs of jobnets and jobs are output to the scheduler log, and this information is used mainly for monitoring the execution statuses of applications. The required size for scheduler log files differs greatly depending on the following considerations.

You should therefore estimate an appropriate size for the scheduler log files and change the default disk capacity as required. For details about the default disk capacity and the maximum disk area used for the scheduler log, see 1.2.5 List of log files and directories in the manual JP1/Automatic Job Management System 3 Troubleshooting.

The flow for estimating scheduler log file size is shown below.

Figure 3‒1: Flow for estimating scheduler log file size

[Figure]

Organization of this subsection

(1) Formula for estimating scheduler log file size

The number of formulas that are required for estimating the size of the scheduler log file by scheduler service differs depending on the values set for environment setting parameters. Select the required formulae based on the conditions listed in the following table.

Table 3‒14: Formulas for estimating scheduler log file size

Settings for scheduler service configuration definitions

Estimation formula (size: kilobytes)

When all is set for the environment setting parameter AJSLOG

A =

80 x (scheduler-service-start-count + scheduler-service-end-count)

/ 1,024

When a setting other than none is set for the environment setting parameter NETLOG

B =

{ (90 + maximum-length-of-jobnet-name)

x (number-of-jobnets-started-per-day + number-of-jobnets-ended-per-day + number-of-start-condition-monitoring-processes-started + number-of-start-condition-monitoring-processes-ended + number-of-jobnets-with-hold-setting + number-of-jobnets-being-monitored-for-start-and-end-delays) }

/ 1,024

When a setting other than none is set for the environment setting parameter JOBLOG

C =

{ (120 + maximum-job-name-length + maximum-length-of-the-job-execution-agent-name)

x (number-of-jobs-under-the-root-jobnet-started-per-day + number-of-jobs-under-the-root-jobnet-ended-per-day + number-of-jobs-with-hold-setting) }

/ 1,024

When a setting other than none is set for the environment setting parameter OPELOG or REFLOG

D =

{ (100 + maximum-length-of-JP1-user-names-that-execute-commands + maximum-length-of-logical-host-names + maximum-length-of-options-specified-for-commands-that-operate-units-and-are-specified-for-output-to-a-log)

x (total-execution-count-of-commands-that-do-not-operate-units-and-have-been-specified-to-output-a-log + total-number-of-units-subject-to-commands-that-operate-units-and-have-been-specified-to-output-a-log) }

/ 1,024

When a setting other than none is set for the environment setting parameter INFOLOG

E =

{ (100+ maximum-length-of-JP1-user-names-that-execute-commands + maximum-length-of-logical-host-names + maximum-length-of-options-specified-for-commands)

x (total-execution-count-of-commands-that-do-not-operate-units-and-have-been-specified-to-output-a-log + total-number-of-units-subject-to-commands-that-operate-units-and-have-been-specified-to-output-a-log x 2) }

/ 1,024

After calculating values with the formulae in this table, find the total of the values obtained.

When log data is to be output for each host, calculate the above formula for each scheduler service, and then add up the results to find the total daily volume of log data. If necessary, multiply this by the number of days that you want to keep the log. Set a value a little larger than the estimate.

(2) Example scheduler log file estimate

In this example the scheduler logs for five days are saved in a single file.

Table 3‒15: Example scheduler log file estimate

Condition

Example estimation formula

Setting for the environment setting parameter AJSLOG

all

Scheduler service start count

1

Scheduler service end count

1

A (kilobytes)=

80 x (1 + 1) / 1,024 = approx. 0.16

Setting for the environment setting parameter NETLOG

all

Maximum length of jobnet name

40 bytes

Start count for jobnets executed during one day

1,000

End count for jobnets executed during one day

1,000

Number of jobnets with a hold setting

10

Number of jobnets for which start or end delay monitoring is set

20

B (kilobytes) =

{ (90 + 40) x (1,000 + 1,000 + 10 + 20) } / 1,024 = approx. 258

Setting for the environment setting parameter JOBLOG

all

Maximum length of job name

50 bytes

Maximum length of job execution destination host name

10 bytes

Start count for the number of jobs executed during one day

3,000

End count for the number of jobs executed during one day

3,000

Number of jobs with a hold setting

30

C (kilobytes) =

{ (120 + 50 + 10) x (3,000 + 3,000 + 30) } / 1,024 = approx. 1,060

Setting for the environment setting parameter OPELOG or REFLOG

all

Maximum length of names of JP1 users that execute commands

10 bytes

Length of logical host name

11 bytes

Length of options specified for commands that operate units (ajschange, ajsplan, etc.) and are output to a log

300 bytes

Total execution count of commands that operate units and are specified for output to a log

200

Total number of units subject to commands that operate units and are specified for output to a log

50

D (kilobytes) =

{ (100 + 10 + 11 + 300) x (200 + 50) } / 1,024 = approx. 103

Setting for the environment setting parameter INFOLOG

all

Maximum length of names of JP1 users that execute commands

10 bytes

Length of logical host name

11 bytes

Length of options specified for commands that operate units (ajschange, ajsplan, etc.) and are output to a log

300 bytes

Total execution count of commands that operate units and are specified for output to a log

200

Total number of units subject to commands that operate units and are specified for output to a log

50

E (kilobytes) =

{ (100 + 10 + 11 + 300) x (200 + 50 x 2) } / 1,024= approx. 123

Scheduler log file capacity for one day (units: kilobytes) =

A + B + C + D + E = 1,544 kilobytes

Size of scheduler log file capacity for 6 days (units: kilobytes) =

1,544 x 6 days = approximately 9,264 kilobytes

The initial size of a scheduler log file is 10,240 kilobytes. Two files are stored. When one week is six business days and the size of a scheduler log file is the initial size, for operation using the example estimates, one file can hold a week's worth of log data.

If the unit you have set for outputting log data is hosts, estimate the sizes of the scheduler log files for all the configured scheduler services for each physical host or logical host, and total the estimates. The result is the estimate for one day.

(3) Expanding the scheduler log file size

To expand the size of the scheduler log file, use the jajs_config command. In the environment setting parameter LOGSIZE (if outputting log files per scheduler service) or HOSTLOGSIZE (if outputting log files per host), set the value you calculated as described in (1) Formula for estimating scheduler log file size.

For details about the environment setting parameters, see 20.4 Setting up the scheduler service environment in the JP1/Automatic Job Management System 3 Configuration Guide.