|
|
Generating the Mimer SQL System Databanks
The Mimer SQL system databanks SYSDB, TRANSDB, LOGDB and SQLDB are generated by the SDBGEN program.
SDBGEN will load the system tables, see Data Dictionary Tables, and defines the data dictionary views detailed in the Mimer SQL Reference Manual.
Note: A databank created for one SYSDB cannot be accessed by using a different SYSDB even if identical data dictionary definitions are created in it.
Setting the Initial Size
The initial size for each of the Mimer SQL system databanks can be specified.
The size for the databanks is specified in Mimer SQL pages. The size of a Mimer SQL page is 2 kilobytes.
Creating SYSADM
The database administration ident SYSADM is also created and a password (passwords are case-sensitive) must be specified for this ident.
For security reasons, the password specified for SYSADM is not echoed on the screen when it is entered. It should be chosen carefully and changed at appropriate intervals using Mimer SQL with the ALTER IDENT statement.
UNIX: A local database is set up on a UNIX node by running the dbinstall command (see the dbinstall man-page) and SDBGEN is started automatically when required.
The databank files are by default created in the database home directory which is not the ideal arrangement from a security and performance point of view, see Re-creating TRANSDB, LOGDB and SQLDB for guidelines relating to placement and organization.
If there is more than one disk available on the system, it is recommended that directories be created on those disks specifically for locating databanks. When created, the LOCAL definition for the database should be updated in the /etc/sqlhosts file by changing the single home directory path to a directory path list that includes these directories. The list is colon separated as can be seen in the following example:
VMS: In order to create the Mimer SQL system databanks for a local database on a VMS node, an entry for the database must be specified in the SQLHOSTS file. SDBGEN should then be executed to create the actual database. See SDBGEN Syntax for information on how to run SDBGEN.
SDBGEN Syntax
The SDBGEN command has two purposes. Either to create a new set of system databank files, or to upgrade database files created in an earlier version of Mimer SQL to version 8.2. (Upgrade can be done for databank files created by Mimer SQL version 7.1 and later.)
The SDBGEN program is controlled by flagged information specified on the command-line.
The syntax (expressed in VMS-style) for creating databank files is as follows:
SDBGEN [/PASSWORD=passw] [dbase] [syssz] [tfn] [tsz] [lfn] [lsz] [sfn] [ssz]sdbgen Command-line Arguments
When creating databanks files:
If the password parameter is omitted, the SDBGEN command will prompt for all parameters that are missing, including the password for the SYSADM user.
If the password parameter is given, the SDBGEN command will not prompt for any missing parameters, but use default values.
If the dbase parameter is missing, the environment variable MIMER_DATABASE is used to determine which database the databank files should be created for.
The SDBGEN program syntax (expressed in VMS-style) for upgrading databank files to Mimer SQL version 8.2 is as follows:
SDBGEN /UPGRADE [dbase]sdbgen Command-line Arguments - Upgrading
UNIX-style
VMS-style
Function
-u /UPGRADE Upgrade existing database to Mimer SQL version 8.2 dbase dbase Database name
Establishing the Ident and Data Structure
Once the local database environment has been created for a Mimer SQL database (database name, server parameters, system databanks, the SYSADM ident plus the system tables and views), the data structure for the database (idents, user databanks, tables, and so on) can be created using Mimer SQL data definition statements.
BSQL allows the execution of a sequential file which can then be used as a permanent record of the CREATE statements used to create the database objects, see the Mimer SQL User's Manual.
BSQL also supports the saving of input and/or output to a log file (using the LOG command), so this facility could be used to create a permanent record of an interactive BSQL session which could be run again at a later date. BSQL, however, only has limited support for error handling.
An application program using embedded SQL, JDBC or ODBC can also be used, but this requires more work on the part of the programmer and it provides a less concise record of the ident and data structure in the database.
Third party SQL tools are also available which may be used to create the database data structure.
Note: A sequential file intended for non-interactive execution through the BSQL facility may include username and password information relating to any CONNECT statements used. For security reasons, such a file should be well protected in the operating system, preferably with any username and password edited out of any permanent copy of the file.
|
Upright Database Technology AB Voice: +46 18 780 92 00 Fax: +46 18 780 92 40 dbtechnology@upright.se |
|
|