23.1.1 Examining the embedded-database operating environment and operating method

■ System common definition

The following describes the operands in the system common definition, the values of which can be changed by the embedded database system administrator.

The path name of the system common definition file is as shown below. You (the embedded database system administrator) can change the values of operands specified in this file.

• Windows: embedded-database-practical-directory\conf\pdsys

• UNIX: embedded-database-practical-directory/conf/pdsys

Note that the operands in the following table that are enclosed in square brackets ([ ]) are not written when the embedded database is set up. Add the operands if necessary.

Important

To ensure correct JP1/AJS3 operation, do not change the values of any operands that are not described here.

(1) pd_max_users = maximum-number-of-concurrently-connectable-users

~ <unsigned integer> ((1 to 2000))

This operand specifies the maximum number of users who can connect to the embedded database. If you want to change the maximum number of users who can connect to the embedded database, change the value of this operand. If connection to the embedded database is requested when the number of users who have already been connected to the embedded database has reached the maximum, the embedded database does not accept the request (the request results in an error). Note that in JP1/AJS3, if the backup enhancement function is disabled, the default setting is 128, and if the backup enhancement function is enabled, the default setting is 256.

Cautionary notes:

• For embedded-database operation commands, each command is counted as a single connection to an embedded database.

  The number of connections to the embedded database is counted as shown in the following table.

  JP1/AJS3 behavior	Number of connections
  Startup of JP1/AJS3 - Manager	2
  Startup of a scheduler service	4#1, #2, #3
  Execution of a command that performs an operation on a unit#4	1
  Connection from JP1/AJS3 - View	The number of viewed scheduler services
  Execution of a command that performs an operation on the embedded database	2
  Execution of a remote jobnet (execution host)	1
  Execution of a remote jobnet (target host)	1
  Connection from JP1/AJS3 - Web Console	Number of connected scheduler services

  #1

  If wait conditions are used, add 1 to this value. To determine whether wait conditions are used, check whether the procedure described in 6.1.8 Settings for using wait conditions has been performed.

  #2

  If the manager functions for queueless job execution have been set up, add the value specified in the QUEUELESSMULTIREQ environment setting parameter to this value.

  For details about the QUEUELESSMULTIREQ environment setting parameter, see 20.4.2(76) QUEUELESSMULTIREQ.

  #3

  If the backup enhancement function is enabled, add 1. You can check the setting of the backup enhancement function by using the ajsembdbstatus command. For details about the ajsembdbstatus command, see ajsembdbstatus in 3. Commands Used for Normal Operations in the manual JP1/Automatic Job Management System 3 Command Reference.

  #4

  Cases of linking to JP1/IM2 and referring to unit information are also applicable.

  Note that, when a JP1/AJS3 command is executed from JP1/AJS3 - View, 1 connection is used from JP1/AJS3 - View to the embedded database, and 1 connection is used for executing the command.

• If you set a value larger than 500 for pd_max_users, you need to expand the space for the sink point dump files. Execute the ajsembdbaddlog command three times with the -s 7 and -r spd options specified to expand the space for the sink point dump files.

  The following shows an example of specifying the ajsembdbaddlog command:

  ajsembdbaddlog -s 7 -r spd -id _JF0 -d directory-name

  For details about the option of the ajsembdbaddlog command, see ajsembdbaddlog in 3. Commands Used for Normal Operations in the manual JP1/Automatic Job Management System 3 Command Reference.

• In UNIX, if you increase the value of this operand, adjust system resources. For the system resources that require adjustment, see the Release Notes.

(2) pd_service_port = client-connection-port-number

~ <unsigned integer> ((5001 to 65535))

If there is a firewall on the embedded database server side, specify the client connection port number. Note that the port number you specify must satisfy the following condition:

• Port number that is unique on the host

Explanation:

• Specify this operand as necessary if you use a product that links to JP1/AJS3. You do not need to specify this operand in JP1/AJS3 - Manager because it is not available in JP1/AJS3 - Manager.

• If you start multiple servers or units on the same computer, specify a separate port number for each embedded database.

Cautionary note:

Make sure that the port number you specify is outside the range of port numbers assigned automatically by the OS. This differs depending on the OS. If another program is already using a port number in the range and you specify that port number, you will not be able to start the embedded database.

(3) pd_mode_conf = starting-the-embedded-database

Specify how the embedded database starts. You can set the values below for pd_mode_conf. The default is MANUAL2. You usually do not need to change the default.

• AUTO

  When the OS starts, the embedded database is started automatically. If the embedded database terminates abnormally, it is restarted automatically.

• MANUAL1

  When the OS starts, the embedded database is not started automatically. The embedded database is restarted automatically only if it terminates abnormally.

• MANUAL2

  When the OS starts, the embedded database is not started automatically. Unlike MANUAL1, if the embedded database terminates abnormally, it is not restarted automatically.

(4) pd_utl_exec_time = execution-monitoring-time-of-a-command-for-operating-the-embedded-database

~ <unsigned integer> ((0 to 35791394)) <<0>> (unit: second)

If you monitor the execution time of the following commands that operate the embedded database, specify the monitoring time in minutes.

• ajsembdbreclaim command

• ajsembdbrorg command

If the command does not complete execution within the time specified in this operand, a message (KAVS2117-E) is output to the standard error output and the command terminates abnormally.

Advantage:

If the command does not send any response during execution in a nighttime batch job because of a problem such as a communication error or disk error, you can allow the command to terminate abnormally at the specified time.

Guideline for specifying a value

The purpose of this operand is to handle a non-response error, not to monitor the maintenance processing time of the database. Therefore, specify a value for the operand that includes a margin for the maximum value of the actual performance value for the command execution time. For example, if the maximum execution time of the ajsembdbreclaim command is about 60 minutes and the maximum execution time of the ajsembdbrorg command is about 90 minutes, specify 120 for pd_utl_exec_time to provide the margin. The reason for this value is that a non-response error might occur if processing that always finishes in 90 minutes fails to send a response after an additional 30 minutes has elapsed. Note that the time of 30 minutes added to the maximum command execution time is based on the assumption that maintenance is being performed on a scheduler database and the agent management database. If a different type of database access is being performed (for example, access for execution of a job or unit operation command) during the maintenance processing, the maintenance processing waits one minute for the processing to be completed. Accordingly, as a guideline for the time to be added, specify a value that includes a margin obtained from the expression number-of-processing-target-tables x 2 x 1 (in minutes).

For details about the tables that are created in the embedded database, see 2.6.3 JP1/AJS3 database tables in the JP1/Automatic Job Management System 3 System Design (Configuration) Guide.

■ Single server definition

The following describes the operands in the single server definition, the values of which can be changed by the embedded database system administrator. The path name of the single server definition file is as shown below. You (the embedded database system administrator) can change the values of operands specified in this file.

• Windows: embedded-database-practical-directory\conf\ajs2

• UNIX: embedded-database-practical-directory/conf/ajs2

Important

To ensure correct JP1/AJS3 operation, do not change the values of any operands that are not described here.

(1) pd_lck_pool_size = size-of-pool-used-for-locking

~ <unsigned integer> (unit: KB) ((1 to 2000000))

Specify, in kilobytes, the size of the shared memory area in the embedded database.

In JP1/AJS3, the default differs depending on the scale specified by the ajsembdbbuild, jajs_setup, or jajs_setup_cluster command.

Scale	Command option	Lock pool size
Large	-l	60,000
Medium	-m	37,500
Small	-s	15,000

Operation method:

If units are to be operated by using commands or JP1/AJS3 - View, the size of memory area used for locking changes based on the number of units to be operated. For this reason, estimate and set the appropriate memory area size in advance.

If many units are operated at the same time, the memory area for locking might become insufficient, causing the following problems:

• The message KAVS0902-E is output to the integrated log, syslog, or Windows event log, and the message KFPA11912-E is output as maintenance information for the message KAVS0902-E. The operations being performed end abnormally.

If insufficient memory area for locking prevents scheduler services from normally accessing the database, the following problems might occur:

• The message KAVS0902-E is output, and jobs and jobnets cannot be executed normally.

• The message KAVS0230-E is output to the integrated log, syslog, or Windows event log, and scheduler services end abnormally.

The unit operations^#1 that use the memory area for locking are as follows:

1. The unit defined by using the ajsdefine command and all the units under control of the unit

2. The unit to be deleted by using the ajsdelete command or JP1/AJS3 - View and all the units under control of the unit^#2

3. The unit specified as the copy source for the unit copy operation by using the ajscopy command or JP1/AJS3 - View and all the units under control of the unit

4. The unit to be restored by using the restoring operation via the ajsrestore command or JP1/AJS3 - View and all the units under control of the unit

5. The root jobnet specified as the release source for the release operation by using the ajsrelease command or JP1/AJS3 - View and all the units under control of the root jobnet^#3

6. The root jobnet to be registered by using the execution registration operation via the ajsentry command or JP1/AJS3 - View and all the units under control of the root jobnet

7. The root jobnet to be re-executed by using the re-execution operation via the ajsrerun command or JP1/AJS3 - View and all the units under control of the root jobnet

8. The units that have been added to or deleted from the root jobnet to be released from suspended state by using the suspension cancellation operation via the ajssuspend command or JP1/AJS3 - View

9. The unit to be imported by using the ajsimport command or JP1/AJS3 - View and all the units under control of the unit

10. The units to be changed by using the ajsplan command or the temporary change operation in JP1/AJS3 - View

11. The root jobnet to which registered execution-schedule information is to be imported by using the ajsrgimport command and all the units under control of the root jobnet

12. The units to be checked by using the pre-check on definition contents via the ajschkdef command

13. Registration cancellation by using the ajsleave command or JP1/AJS3 - View^#4

To prevent shortages in the memory area for locking, compare the value set for pd_lck_pool_size and the value calculated by using the following expression^#5. If the value set for pd_lck_pool_size is smaller, change the value to the calculated value.

• In cases 1, 2, 3, 4, 5, 8, or 9 above

  (a + b + c + d + e) / 4 x 1.2

  a: (total-number-of-job-groups x 2) + total-number-of-years-in-the-calendar-defined-for-job-groups

  b: (total-number-of-jobnets x 2) + total-number-of-schedule-rules-defined-for-all-jobnets

  c: total-number-of-jobs x 2

  d: total-number-of-relations-defined-for-jobs-and-jobnets

  e: total-number-of-jobs-and-jobnets-with-wait-condition-defined + total-number-of-defined-units-to-be-waited-for

  If the UNITDEFDIVIDEUPDATE environment setting parameter is set to yes and the calculated value is 4,000 or greater, set 4,000.

• In cases 6, 7, 10, or 11 above

  (f + g) / 4 x 1.2

  f: total-number-of-jobs-and-jobnets x 3

  g: total-number-of-jobs-and-jobnets-with-wait-conditions-defined

• In cases 12 above

  h / 4 x 1.2

  h: total-number-of-jobs-and-jobnets-and-jobgroups

If the jobnets used in the calculation use the jobnet release function, perform the calculation by using the jobnet definition as the target of operation. Also when an operation equivalent to the above operation is performed by a product related to or linked with JP1/AJS3, the memory area for exclusive control is used in a similar way.

If you change the value set for pd_lck_pool_size, see 3. Estimates in the JP1/Automatic Job Management System 3 System Design (Configuration) Guide and the Release Notes, and then revise the memory requirements.

#1:

The memory area for locking is also used when operations equivalent to those described above are performed from products related to or linked with JP1/AJS3.

#2:

If the jobnet used for calculation uses the jobnet release function, definitions whose statuses are Release wait or Applied are subject to processing. Add the total number of units in those definitions. To estimate the number of units in a definition whose status is Applied, use the following expression: number-of-units-under-control-of-the-root-jobnet x number-of-release-registrations-performed-in-maximum-number-of-days-for-which-generations-are-retained

For details about the status of a jobnet definition, see 7.3.3 Jobnet status after registration of a jobnet-definition release in the JP1/Automatic Job Management System 3 Administration Guide.

#3:

If the jobnet used for calculation uses the jobnet release function, definitions of release-destination units whose statuses are Applied are subject to processing. Add the total number of units in those definitions. To estimate the number of units in a definition whose status is Applied, use the following expression: number-of-units-under-control-of-the-root-jobnet x number-of-release-registrations-performed-in-maximum-number-of-days-for-which-generations-are-retained

For details about the status of a jobnet definition, see 7.3.3 Jobnet status after registration of a jobnet-definition release in the JP1/Automatic Job Management System 3 Administration Guide.

#4:

It is necessary to estimate only when the backup enhancement function is enabled.

#5:

If multiple operations are performed at the same time, use the sum of the calculated values for comparison.

Cautionary notes:

• If you change the specified value of this operand to a larger value and then concurrently operate more units than the guideline for the database scale suggests, a message (KFPS01220-E) is output to the event log or to syslog and the embedded database might terminate abnormally. In such cases, reduce the number of units that are operated concurrently or consider configuring the embedded database on a larger scale.

  For details about the guideline for the number of units, see 23.2 Estimating the database area.

• If you increase the value of this operand, the amount of memory used by the embedded database also increases proportionally. Before you change the value, make sure that there is sufficient memory.

  In addition, if you change the specified value of this operand to a larger value in UNIX, adjust the system resources accordingly. For details about the system resources that need to be adjusted, see the Release Notes.

  If the value of this operand is too small, an attempt to access JP1/AJS3 might result in an error.

• The maximum value that can be specified for this operand is 2,000,000. However, if OS memory management cannot secure a continuous area, startup of the embedded database fails. If a continuous area cannot be secured, adjust the value so that a continuous area can be secured.

■ Changing the system definition

To change the system definition:

1. Stop the JP1/AJS3 services.

2. Check the status of the target embedded database.

   ajsembdbstatus -s ust -id _JF0#

3. If the target embedded database is running, stop it normally. If the target embedded database is not running, start it, and then stop it normally.

   Example of the command used to start an embedded database:

   ajsembdbstart -id _JF0#

   Example of the command used to stop an embedded database normally:

   ajsembdbstop -id _JF0#

4. Using a text editor, open the system common definition file or a single server definition file.

   Before performing this operation, create a backup of the file in case a problem occurs.

5. Change the value of the operand.

6. Save the change made in step 5, and then close the file.

7. In a cluster configuration, copy the changed system common definition file or the single server definition file to the standby host.

8. Start the JP1/AJS3 services that you stopped.

   The embedded database is also started at this time.

#:

For the -id option, specify the setup identifier of the target embedded database. To start the embedded database in Windows, start the JP1/AJS3 Database _JFn (n: 0 to 9 or A to Z) service in advance.

Important

If you use a multi-byte space or character, or add an unnecessary line-feed or space character in the system common definition or single server definition, startup of the JP1/AJS3 services fails.

■ Required disk space

For details about disk capacity, see the Release Notes.

Note that if you install an embedded database on an NTFS-formatted disk in Windows, you must not compress the files in the embedded database practical directory. If you do so, the embedded database will not operate normally.

■ Required memory

For the memory requirements, see the Release Notes.

In Windows, you must also check the virtual memory requirement. To do this, perform the following:

1. In Windows Control Panel, and then double-click System.

   The System dialog box appears.

2. Click Advanced system settings.

   The System Properties dialog box appears.

3. Click the Advanced tab.

4. Click the Settings button of Performance.

   The Performance Options dialog box appears.

5. Click the Advanced tab.

6. Click the Change button of Virtual memory.

The virtual memory required for one instance of an embedded database is as follows:

Paging file size = 130 MB

To determine the size you actually specify, add the size required by Windows and other programs. If you change the size of the virtual memory, you must restart Windows.

Cautionary note

The size may increase depending on the number of JP1/AJS3 - Views that are connected or the number of scheduler services that are set.

For the initial size, specify a quantity of virtual memory that can be allocated in a continuous area on one drive. If a contiguous area cannot be used for virtual memory, the embedded database might terminate abnormally because memory is insufficient.

■ Checking kernel parameters

In UNIX, you must change the kernel parameters as necessary. For details about estimating kernel parameters, see 3.2.5 Estimating the values for kernel parameters in the JP1/Automatic Job Management System 3 System Design (Configuration) Guide.