Browse Prior Art Database

Utility Restart

IP.com Disclosure Number: IPCOM000050718D
Original Publication Date: 1982-Dec-01
Included in the Prior Art Database: 2005-Feb-10
Document File: 3 page(s) / 16K

Publishing Venue

IBM

Related People

Mortenson, JA: AUTHOR

Abstract

Herein is described a method for maintaining synchronization between sequential data sets and a data base, comprising the steps of: maintaining in a data base table, a record for each active utility which specifies the current position of that utility in both the data base and the sequential data set; when creating a sequential data set, committing only after both updating the data base table record and assuring completion of writes to the sequential data set; and responsive to a utility restart, positioning the data base and the sequential data set to the locations specified in the corresponding data base table record.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 45% of the total text.

Page 1 of 3

Utility Restart

Herein is described a method for maintaining synchronization between sequential data sets and a data base, comprising the steps of: maintaining in a data base table, a record for each active utility which specifies the current position of that utility in both the data base and the sequential data set; when creating a sequential data set, committing only after both updating the data base table record and assuring completion of writes to the sequential data set; and responsive to a utility restart, positioning the data base and the sequential data set to the locations specified in the corresponding data base table record.

A data base system provides mechanisms to ensure the consistency of the data it manages (i.e., only 'committed' data is viewed). However, utilities which operate on this data also make use of external storage which is not managed by the data base system (e.g., tape backup copies). In order to provide a utility checkpoint/restart mechanism, there must be some means to synchronize both types of data to the same point of consistency.

Data bases tend to contain very large amounts of data. In addition this data is related to many complicated ways. Any program which must access all or most of the data in one of these data bases will take a very long time to execute. Data base utilities are prime examples of such program.

The ability to restart a utility allows execution to begin from some prior (hopefully, not too prior) point of consistency. Without this ability the entire program would have to be reexecuted (this can be very frustrating if a 20-hour job fails after 19.5 hours).

For a data base system to support large data bases, the ability to restart utilities is a prime requirement.

The IBM MVS operating system (as well as most other operating systems) provides programs known as 'access methods' which are invoked by other programs to access data. One set of access methods provides access to sequential data sets.

In order to read or write data into a sequential data set there first must be established a 'position' from which to start. In most cases this is done automatically by the access method (e.g., when a data set is OPENed for input, the position indicator is automatically set to the first record). After a record is read/written, the position indicator is automatically incremented to the next position.

However, the OS/VS BSAM (Basic Sequential Access Method) provides two special operators to allow explicit control of positioning: . NOTE - return the current position indicator. . POINT - sets the current position indicator to a program supplied value. By supplying the appropriate position

information on the POINT operator, the data set is positioned

at that location. Subsequent read or write activity takes

place from the specified location.

1

Page 2 of 3

Herein is described a data manager system (DMS) protocol that provides functions which are equivalent to the BSAM NOTE and POINT operators. This p...