Hitachi

Hitachi Advanced Database Setup and Operation Guide


7.2.1 Operands related to system configuration (set format)

[1] adb_db_path = DB-directory-name

~<path name> ((2 to 510 bytes))

Specify an absolute path name for the DB directory name. This operand must be specified. The DB directory name specified here must match the DB directory name specified using the adbinit command.

The following rules apply to the path name specified in this operand:

  • The root directory cannot be specified.

  • Specify a directory that is different from the server directory (including subdirectories).

  • Specify a path name consisting of only alphanumeric characters, underscore (_), hyphen (-), hash mark (#), at sign (@), and forward slash (/).

[2] adb_rpc_port = HADB-server-port-number

~<integer> ((5,001 to 65,535)) <<23,650>>

Specify the HADB server port number that will be used to communicate between the HADB server and an HADB client.

Multi-node function

If you are using the multi-node function, specify the same value for this operand for all of the HADB servers on all nodes.

[3] adb_sys_max_users = maximum-number-of-concurrent-connections

~<integer> ((1 to 1,024)) <<10>>

Specify the maximum number of HADB clients that can concurrently connect to the HADB server.

If you will be using the client-group facility, set this operand to a value that is greater than the sum of the guaranteed minimum numbers of concurrent connections for all groups. If the sum of the guaranteed minimum numbers of concurrent connections for all groups is greater than the value set for this operand, an error will occur when the HADB server starts.

For details about the guaranteed minimum number of concurrent connections for each group, see the following explanations of the -u option of the adbcltgrp operand in 7.2.12 Operands and options related to the client-group facility (command format):

Multi-node function
  • If you are using the multi-node function, specify the same value for this operand for all of the HADB servers on all nodes.

  • In a multi-node configuration, the value specified for this operand becomes the maximum number of concurrent connections allowed for the HADB servers.

[4] adb_core_path = name-of-output-destination-directory-for-error-information-(core-files)

~<path name> ((2 to 510 bytes))

Specify an absolute path for the output destination directory for error information (core files). Error information (core files) is output to the directory specified by this operand.

Normally, this operand does not need to be specified. When this operand is omitted, error information (core files) is output to the $DBDIR/SPOOL directory.

You specify this operand to output error information (core files) to a directory other than the $DBDIR/SPOOL directory. Do not specify the conf directory ($ADBDIR/conf) under the server directory or its subdirectories.

Important

When you specify this operand, specify a directory for which the HADB administrator has access privileges (read, write, and execution privileges).

[5] adb_blk_path_wrk = path-name-of-block-special-file-of-work-table-DB-area-file

~<path name> ((2 to 255 bytes))

Specify a path name for the block special file to be allocated to the work table DB area file when the HADB server starts.

The work table DB area file is changed according to the value specified for this operand and the status of the existing work table DB area file (ADBWRK).

If you change the existing work table DB area file, after making the change, initialize the file. If you are not changing the existing work table DB area file, check its content, and initialize it if it has not been initialized. The following table shows the details.

Table 7‒1: Relationship between the adb_blk_path_wrk operand and the work table DB area file

Specification of the adb_blk_path_wrk operand in the server definition

Existing work table DB area file

None

Regular file

Block special file

Specification is omitted.

A regular file is created.

Unchanged (The existing regular file remains as is.)

Unchanged (The existing block special file remains as is.)

A path name is specified.

Uses the block special file specified in the adb_blk_path_wrk operand.

Uses the block special file specified in the adb_blk_path_wrk operand. If the same name as an existing block special file is specified, no change is made.

Multi-node function

If you are using the multi-node function and if the work table DB area file is not unique to one of the nodes, take one of the following steps:

  • For this operand, specify a block special file that is not shared with other nodes.

  • Alternatively, delete the existing work table DB area file and omit this operand.

[6] adb_dbarea_wrk_page_size = page-size-of-work-table-DB-area

~<integer> ((32 to 32,768)) (kilobytes)

When you want to change the page size of the work table DB area that was specified when the adbinit command was executed, specify the desired page size. If you specify a value that is not divisible by 32, the specified value is rounded up to the next multiple of 32 kilobytes.

Normally, this operand does not need to be specified. If this operand is omitted, the page size that was specified when the adbinit command was executed is applied.

Increasing the page size might cause memory to be insufficient. Therefore, before you increase the page size by using this operand, review the memory requirements. For details about how to estimate the memory requirements, see 6.3.3 Determining the memory requirement for starting the HADB server and 6.3.4 Determining the memory requirement during normal operation.

You can use the following commands to check the page size that is currently applied on the active HADB server:

  • adbls -d gbuf command

    For details about this command, see adbls -d gbuf (Display Global Buffer Information) in the manual HADB Command Reference.

  • adbls -d lbuf command

    For details about this command, see adbls -d lbuf (Display Local Work Table Buffer Information) in the manual HADB Command Reference.

The page size of the work table DB area that was specified when the adbinit command was executed can be checked by searching the dictionary table. For details about searching a dictionary table, see (37) Checking the page size of the work table DB area specified when the adbinit command was executed in B.22 Searching a dictionary table.

Multi-node function

If you are using the multi-node function, specify the same value for this operand for all of the HADB servers on all nodes.