Browse Prior Art Database

Tracing, Formatting and Storage Referencing in an MVS Multitasking Online System

IP.com Disclosure Number: IPCOM000061540D
Original Publication Date: 1986-Aug-01
Included in the Prior Art Database: 2005-Mar-09
Document File: 3 page(s) / 42K

Publishing Venue

IBM

Related People

Frantz, BJ: AUTHOR

Abstract

This invention relates to a method for tracing, formatting, and storage referencing in an MVS multitasking online system. The steps include (a) specifying the contents of systems to be collected and correlating the systems' contents, hierarchical level, and failure type; (b) upon occurrence of a failure within a particular type, specifying the level of collected data to be dumped; and (c) identifying the subset of the dumped data to be displayed. Generally, dump routines utilize no logic in their attempt to produce the data needed to correct the problem. Rather, they capture all available data that may be remotely associated with the address space of the failing subsystem and leave it to the problem solver to sift through it and determine its usefulness.

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 43% of the total text.

Page 1 of 3

Tracing, Formatting and Storage Referencing in an MVS Multitasking Online System

This invention relates to a method for tracing, formatting, and storage referencing in an MVS multitasking online system. The steps include (a) specifying the contents of systems to be collected and correlating the systems' contents, hierarchical level, and failure type; (b) upon occurrence of a failure within a particular type, specifying the level of collected data to be dumped; and
(c) identifying the subset of the dumped data to be displayed. Generally, dump routines utilize no logic in their attempt to produce the data needed to correct the problem. Rather, they capture all available data that may be remotely associated with the address space of the failing subsystem and leave it to the problem solver to sift through it and determine its usefulness. This approach, which has been relatively unchanged through several computer generations, had begun to become unwieldly for large subsystems, even before the advent of the MVS/XA system. The method of this invention allows the analysis and correction of subsystem problems with a minimum amount of captured data. It is necessary to recognize which storage areas are in use by the currently dispatched subtask which caused the abnormal termination. Using system-wide standards and code identification techniques, along with internal service aids, the data required to solve the problem can be isolated through programmed logic. Since a single subtask uses a very small subset of the total storage allocated to a large online subsystem, the data captured using this process is reduced considerably. Each subtask within the system establishes either a first or second state at initiation. Thus, whenever an abnormal termination occurs, control is given to the subsystem-supplied dump capture routine. Since it is highly probable that in a busy multitasking system a failure may cause more than one subtask to terminate, the recovery routine is designed to capture only the initial termination, thus avoiding the production of multiple dumps which generally cause more confusion than help in problem determination. A dump data set is opened during system initialization to provide a collection medium for the dump routine. A dump routine, once entered due to an abnormal termination, provides abend recursion protection by establishing its own functional recovery routine immediately upon entry. This routine is designed to provide two functions. If, while recording information, the dump routine abends attempting to access storage with an invalid address, recovery takes place and the dump continues after recording the invalid address as such. If a failure occurs due to a dump routine logic error or unforeseen condition, the recovery routine reverts to a special dump routine called SDUMP to collect sufficient data to correct the dump routine failure. An ABEND condition can normally be traced to some flaw in the code of the currently a...