Browse Prior Art Database

Transaction Recovery Scheme for VM

IP.com Disclosure Number: IPCOM000048344D
Original Publication Date: 1982-Jan-01
Included in the Prior Art Database: 2005-Feb-08
Document File: 1 page(s) / 12K

Publishing Venue

IBM

Related People

Larson, LE: AUTHOR

Abstract

This disclosure uses a combination of VM/VCS (Virtual Machine/ Virtual Control STore) and a modification to the directory tables of VM to provide to VM a transaction recovery capability. The present situation in VM is that if two users wish to share information on disk, each user is given a copy of the disk directory. Because changes to the disk are reflected in the directory, user changes to the disk files can cause the users of the disks to experience unpredictable results.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 54% of the total text.

Page 1 of 1

Transaction Recovery Scheme for VM

This disclosure uses a combination of VM/VCS (Virtual

Machine/ Virtual Control STore) and a modification to the directory tables of VM to provide to VM a transaction recovery capability. The present situation in VM is that if two users wish to share information on disk, each user is given a copy of the disk directory.

Because changes to the disk are reflected in the directory, user changes to the disk files can cause the users of the disks to experience unpredictable results.

Attempts to use VCS to correct the problems of unpredictable output result in the other extreme. If two users are sharing a disk (file) and one user issues a SAVE, the changes of all users sharing the disk are saved. Thus while the results are predictable from a programming sense, the results are still unpredictable because it is never clear which modifications are made to the disk (file).

The control of which updates are made to the disk as well as the ability to recover from partially completed updates to the disk files is fundamental to the operation of a controlled data access environment (database). This article solves the problems by using VCS to control the directories and using a modification to the directory maintenance logic to ensure that the data recorded on the disk media is always in the committed state. Because the data on the disk media is always in the committed state, the need to develop log tapes purely for media recovery is eliminated. T...