Browse Prior Art Database

Recovery Control System

IP.com Disclosure Number: IPCOM000075127D
Original Publication Date: 1971-Aug-01
Included in the Prior Art Database: 2005-Feb-24
Document File: 3 page(s) / 39K

Publishing Venue

IBM

Related People

Harrison, WH: AUTHOR

Abstract

When a programming system has reached a state in which it can no longer proceed with useful work, it is said to have "crashed". After such a crash, it is customary to produce a "dump" in which the contents of the main storage are printed. Often this dump is written on tape (or other recording medium) so that the printing may be performed later after the system has been reloaded.

This text was extracted from a PDF file.
At least one non-text object (such as an image or picture) has been suppressed.
This is the abbreviated version, containing approximately 53% of the total text.

Page 1 of 3

Recovery Control System

When a programming system has reached a state in which it can no longer proceed with useful work, it is said to have "crashed".

After such a crash, it is customary to produce a "dump" in which the contents of the main storage are printed. Often this dump is written on tape (or other recording medium) so that the printing may be performed later after the system has been reloaded.

A system crash often leaves many portions of the system in a state which is inconsistent or in which information has been lost. For example, if direct access storage (i.e. disk or drum) was being allocated, it may be lost forever; the time used by jobs or by user sessions in a time-sharing system may be unaccounted for even though useful work had been performed; or information gathered from a terminal or teleprocessing networks and not yet recorded on secondary storage may be lost. In general, however, this information is not really "lost", but it is merely buried on the dump tape in such a fashion that it cannot be retrieved. The system described here provides for the retrieval of this information, in order to remove the inconsistencies and information gaps induced by the system crash.

The recovery control system is comprised of three elements, one of which is supplied for each particular inconsistency or information gap to be removed. The relation between these components is illustrated in the Figure.

Operation of the system proceeds as follows: During the normal running of the system, a problem program P1 is entered which it is known might leave the system in an undesirable state in the event of a crash. This program, therefore, makes a request 2 to the supervisor program 1 indicating that program P1 has a recovery coprogram P2. The portion of supervisor 1 which processes this request is called the Recovery Setup Routine (RSR).

Any operating system maintains information concerning the programs which are currently active, their callers, their status, etc. The area wherein this information is recorded is referred to here as the "system tables". The RSR, upon receiving the request 2 records in...