This glossary includes terms that are unique to Databridge, as well as terms that are standard for DMSII databases. Complete, detailed definitions for Unisys MCP-hosted mainframes and DMSII terms can be found in Unisys documentation. The purpose of this glossary is to explain how these terms relate to Databridge.
An Accessory is any program (including user-written programs) that links to a support library, such as DBEngine, DBSupport, or a user-written library.
The Accessroutines program is a DMSII utility that controls access to the database, reads and writes records, and creates the audit trail.
The audit file is created by DMSII and contains the raw format of changes made to the DMSII database by update programs. Audit file records contain the deletes, adds, and modifies that were made to the various structures. It can contain, for example, hours, days, or weeks worth of information.
When an audit file is closed, DMSII creates the next one in the series. Audit files are closed for several reasons, including the following:
- An operator closes the audit file with the mixnumber SM AUDIT CLOSE command
- The audit file reaches the file size set in its DASDL
- There is an I/O error on the audit file
- There is not enough disk space for this audit file
- The database update level changes due to database definition changes
- The current audit file could not be found
- A file reorganization was executed to modify a DMSII structure
Databridge uses the audit file for the raw data of each database change to exactly replicate the primary database. Databridge records the audit location (AFN, ABSN, SEG, IDX) between runs, so it can restart without losing any records.
If you set READ ACTIVE AUDIT to true in the Engine control file Databridge can access up to and including the current audit file. If you set READ ACTIVE AUDIT to false in the Engine control file, Databridge can access audit information only up to and including the current audit file minus one. Additionally, the audit file contains the update level at the time the audit file was created. The update level in the audit file and the update level in the DESCRIPTION file used by Databridge must match before Databridge will update a replicated database.
The audit trail consists of all of the audit files generated for a database. It can contain recovery records, which indicate that there was a failure such as a HALT/LOAD that caused the Accessroutines to rollback the DMSII database to a quiet point. If DBEngine encounters one of these recovery records, it notifies the caller to rollback the replicated data accordingly.
The audit trail consists of the audit files named as follows:
where databasename is the name of the DMSII database, AUDIT is a literal, and nnnn is the AFN (Audit File Number), a number between 1 and 9999. For example, if you have a database named BANKDB, an audit file would be named similar to the following:
The client is the computer system that will receive DMSII records from the primary database. The client could be a PC, a UNIX computer, or a mainframe. The client can have a relational or a DMSII database.
Cloning is the process of generating a complete snapshot of a data set to another file. Cloning creates a static picture of a dynamic database. Databridge uses the DMSII data sets and the audit trail to ensure that the cloned data represents a snapshot of the data sets at a quiet point, even though other programs may be updating the database concurrently. Databridge clones only those data sets you specify.
Cloning is one phase of the database replication process. The other phase is tracking (or updating), which is the integration of database changes since the cloning. For more details, see the definition for tracking.
Databridge Accessories are available for cloning, as follows:
- DBSnapshot Accessory uses a batch method that provides a one-time snapshot only.
- DBSpan Accessory uses a dynamic method that provides a one-time extraction and fixup followed by ongoing tracking.
- Databridge Clients perform an initial clone of a DMSII database and then subsequent tracking of the changes made to the DMSII database. Databridge Administrative Console connect to Databridge Server or Databridge Enterprise to get the DMSII data.
An elementary item that could be altered into a GROUP item containing multiple elementary items.
For instance, assume that CUST-NAME ALPHA (30) has a 20-character last name, followed by a 9- character first name, and a 1-character middle initial. CUST-NAME could be altered to be a GROUP containing CUST-LAST-NAME ALPHA (20), CUST-FIRST-NAME ALPHA (9), and CUST-MID-INITIAL ALPHA (1).
A file created by DBSpan that contains records for all selected data sets.
The DMSII CONTROL file is the run-time analog of the DESCRIPTION file. The DESCRIPTION file is updated only when you compile a modified DASDL. The CONTROL file controls database interlock. It stores audit control information and verifies that all database data files are compatible by checking the database timestamp, version timestamp, and update level. The CONTROL file is updated each time anyone opens the database for updates. The CONTROL file contains timestamps for each data set (when the data set was defined, when the data set was updated). It contains parameters such as how much memory the Accessroutines can use and titles of software such as the DMSUPPORT library (DMSUPPORT/databasename).
Databridge uses the CONTROL file for the following information:
- INDEPENDENTRANS option
- AFN for the current audit file and ABSN for the current audit block
- Data set packnames
- Audit file packname
- Database usercode
Data And Structure Definition Language—This is the language that defines DMSII databases. The DASDL must be compiled to create a DESCRIPTION file.
A file (structure) in DMSII in which records are stored. It is similar to a table in a relational database. You can select the data sets you want to store in your replicated database.
The DESCRIPTION file contains the structural characteristics of a database (physically and logically). It is created from the DASDL source by the DASDL compiler and contains the layout (physical description), timestamp, audit file size, update level, logical database definition, and any static information about the database. It contains information about the database, not the data itself.
There is only one current DESCRIPTION file for each DMSII database. Databridge must have access to the DESCRIPTION file before it can replicate a database. Additionally, Databridge uses the DESCRIPTION file information for consistency checks between the primary database and the secondary or replicated database.
The DESCRIPTION file corresponds to the schema in a relational database.
The process of reading through a data set sequentially and writing those records to a file (either a secondary database or flat file).
file format conversion
A DMSII file format conversion affects file size values (for example, AREASIZE, BLOCKSIZE, or TABLESIZE), but it does not change the layout of the DMSII database.
Changing an occurring item into a series of individual items.
A procedure residing either directly or indirectly in DBSupport that contains routines for formatting the data items of a data set record. DBGenFormat generates formatting procedures such as COMMAFORMAT and BINARYFORMAT based on the setting of certain options in the DBGenFormat parameter file. You can write custom formatting procedures in ALGOL or COBOL to satisfy specific formatting requirements.
A section of code in a formatting procedure that formats a specific type of data item, such as an ALPHA.
garbage collection reorganization
A garbage collection reorganization moves records around, but it doesn't change the layout of the DMSII database. Its primary function is to improve disk and/or I/O efficiency by eliminating the space occupied by deleted records. Optionally, a garbage collection reorganization reorders the remaining records in the same sequence as one of the sets.
The value defined in the DASDL to be NULL for that ALPHA data item. If the DASDL does not explicitly specify a NULL value for a data item, the NULL value is all bits turned on.
This is the original DMSII database that resides on the ClearPathNX/LX or A Series host. Databridge replicates from the primary database to one or more client databases. The client databases can be another DMSII database or one of several relational databases. Compare this to the replicated (or secondary) database.
quiet point (QPT)
A point in time when no program is in transaction state. This can occur naturally, or it can be forced by a DMSII syncpoint. The quiet point is a point in time in the audit trail that Databridge uses as a reference point to help synchronize cloning or tracking of the DMSII database. Databridge uses the quiet points to ensure an accurate snapshot of the data. Audit addresses of these quiet points are stored in the replicated database for database synchronization purposes.
An ALGOL procedure that allows you to alter or convert data items to different layouts using custom written reformatting routines. The ALTER declaration in the DBGenFormat parameter file indicates which data items will be converted by the reformatting procedure. The procedure itself must call individual reformatting routines to convert the data items.
An ALGOL conversion routine that alters the layout of a data item.
This is the database that resides on the client (also called the client or secondary database) and that contains all of the records cloned from the DMSII database you specified for cloning. The replicated database is updated periodically with changes made to the primary (original) DMSII database. The periodic update (or tracking process) is explained later in this section. Compare this to the primary database.
The ongoing process of cloning and tracking a DMSII database. With the DBSnapshot Accessory, you can clone a database as a one time snapshot to flat files. With the DBSpan Accessory, however, you can extract the database to flat files and then subsequently update it by tracking. The DBSpan Accessory performs extraction as well as tracking.
See replicated database.
An index into a data set.
A data set, set, subset, access, or remap. Each structure has a unique number called the structure number.
A data structure in the client database corresponding to a data set or remap in the host DMSII database.
Retrieving only the changes from the audit trail to apply to the replicated database. Tracking is an ongoing process for propagating changes made to records in the DMSII primary database to the replicated database. The DBSpan and DBServer Accessories perform extraction as well as tracking.
Tracking is one phase of the database replication process. The other phase is cloning.
A NUMBER data item containing values from 10 to 15. The NUMBER data item should contain values from 0 to 9; however, it is possible for NUMBER data item to contain values 0 to 15. Because values 10 to 15 are not valid digit values, NUMBER data items containing values from 10 to 15 are called undigits.