System for performing dump analysis

• US 5,111,384 A
• Filed: 02/16/1990
• Issued: 05/05/1992
• Est. Priority Date: 02/16/1990
• Status: Expired due to Term

• Current Assignee: Bull HN Information Systems Incorporated (Atos SE)
• Original Assignee: Bull HN Information Systems Incorporated (Atos SE)
• Inventors: Aslanian, Edouard, Pitts, Susan L., Sztucinski, Leon A.
• Primary Examiner(s): Smith, Jerry
• Assistant Examiner(s): Cass, Russell E.
• Application Number:
  US07/481,143
• Time in Patent Office: 809 Days
• Field of Search: 371/15.1, 364/513, 364/274.2, 364/274.3, 364/275.7, 395/908, 395/50
• US Class Current: 714/26
• CPC Class Codes: G06F 11/2257 using expert systems
• CPC Class Codes: G06F 11/3476 Data logging G06F11/14, G06F11/2205 take precedence
• CPC Class Codes: Y10S 706/916 Electronic or computer, internal or network, circuit

Abstract

A system for automating the dump analysis process includes a remotely located host computer system which, in response to requests from a local expert computer system, retrieves only relevant values from one or more physical memory dumps. The expert system reconstructs from these values the operating system control structures represented in the dump, and applies expert knowledge on these control structures to determine the symptom of the problem occurring on the computer system which stopped operating and generated the dump.

Claims

1. A system for automatically analyzing memory dumps, said system comprising:

retrieval means;

a number of memory dumps, each memory dump generated as a result of a problem occurring on a given computer system operated- under the control of an operating system, said each dump containing information pertaining to a number of operating system control structures containing information used by the operating system and user programs at the time the dump was generated; and

an expert system comprising;

control means including;

a memory for storing knowledge in the form of objects, said objects including information representative of a plurality of operating system control structures organized into frames and rulesets linked to said frames so as to create decisions using a premise-conclusion if-then form of reasoning; and

,

an inference engine coupled to said memory for interpreting said rulesets using said if-then reasoning; and

,

knowledge representation means coupled to said control means and to said retrieval module, said knowledge representation means for providing dump control structures in response to requests received from said control means, said control means being operative to diagnose said problem causing said memory dump by analyzing the contents of said control structures contained in said memory dump in a predetermined sequence, said knowledge representation means in response to each request causing said retrieval module to transfer only that portion of said memory dump necessary for generating an operating system control structure requested by said control means.

2. The system of claim 1 wherein said control means including validation means, said validation means performing checks on said operating system control structures used by said control means in performing diagnosis.

3. The system of claim 1 wherein said control means include information means, said information means displaying said operating system structure information necessary to explain the analysis being performed by said control means.

4. The system of claim 1 wherein said control means include diagnosis means, said diagnosis means determining the symptom of the problem under consideration and retrieving each specific instance of said operating system structures in which the cause of said problem is apparent.

5. The system of claim 1 wherein said control means include search means, said search means sequencing through all of said operating system control structures contained in said memory dump when said diagnosis means did not generate a conclusion about said problem.

6. The system of claim 1 wherein said control means include known problem file means, said known problem file means storing descriptions of problems previously diagnosed, identified, and resolved, said known problem file means further including problem matching means for comparing the problem currently under consideration with the problems stored in said known problem file.

7. The system of claim 1 wherein said knowledge representation means is organized in a hierarchical data structure of objects in which said objects possess a number of related slots which give specific properties to said objects, a number of descriptors which represent specific constraints of said slots that can be used to modify their behavior, and a number of handlers and daemons associated with said objects to enable communications between said objects.

8. The system of claim 1 wherein said control means include batch control means for enabling the analysis of said number of dumps sequentially without interruption according to specific criteria.

9. The system of claim 8 wherein said batch control means include means for indicating how each of said number of dumps are to be loaded, saved and cleared.

10. The system of claim 1 wherein said system includes at least one host system on which said retrieval means and said number of memory dumps are resident and a local computer system on which said expert system is resident, said systems being interconnected together through a communications link, said retrieval means including means for transmitting data strings back and forth over said communications link between said expert computer system and said host system thus enabling said expert system to access information contained in said memory dump.

11. The system of claim 10 wherein said system further includes a plurality of remotely located host systems, each connected to said communications link for transferring requests to said expert system specifying memory dumps for analysis, said batch control means including file means for listing said requests designating a number of dumps to be analyzed according to specific criteria.

12. A system for automatically analyzing dump files, said system comprising at least one host computer system and a local expert computer system interconnected together by a communications link, said host system including:

a memory for storing a retrieval module and external storage containing a number of memory dumps, each memory dump generated as a result of a problem occurring on a given computer system being operated under the control of an operating system, said each dump containing information pertaining to a number of operating system control structures containing information used by the operating system and user programs at the time the dump was generated; and

,

said local computer system including an expert system comprising;

a control module including;

a memory for storing knowledge in the form objects, said objects including information representative of a plurality of operating system control structures organized into frames and rulesets linked to said frames so as to create decisions using a premise-conclusion if-the form of reasoning; and

,

an inference engine coupled to said memory for interpreting said rulesets using said if-then reasoning; and

,

a knowledge representation module coupled to said control module and to said communications link, said knowledge representation module for providing dump control structures in response to requests received from said control module and said control module being operative to diagnose said problem causing said memory dump by analyzing the contents of said control structures contained in said memory dump in a predetermined sequence, said knowledge representation module in response to each request causing said retrieval module to transfer only that portion of said memory dump necessary for generating an operating system control structure requested by said control module thereby conserving time.

13. The system of claim 12 wherein said control module including a validation module, said validation module performing checks on said operating system control structures used by said control modules in performing diagnosis.

14. The system of claim 12 wherein said control module includes an information module, said information module displaying said operating system structure information necessary to explain the analysis being performed by said control module.

15. The system of claim 12 wherein said control module includes a diagnosis module, said diagnosis module determining the symptom of the problem under consideration and retrieving each specific instance of said operating system structures in which the cause of said problem is apparent.

16. The system of claim 12 wherein said control module includes a search module, said search module sequencing through all of said operating system control structures contained in said memory dump when said diagnosis module did not generate a conclusion about said problem.

17. The system of claim 12 wherein said control module includes a known problem file module, said known problem file module storing descriptions of problems previously diagnosed, identified, and resolved, said known problem file module further including a problem matching module for comparing the problem currently under consideration with the problems stored in said known problem file.

18. The system of claim 12 wherein said knowledge representation module is organized in a hierarchical data structure of objects in which said objects possess a number of related slots which give specific properties to said objects, a number of descriptors which represent specific constraints of said slots that can be used to modify their behavior, and a number of handlers and daemons associated with said objects to enable communications between said objects.

19. The system of claim 12 wherein said retrieval module includes a module for transmitting data strings back and forth over said communications link between said expert computer system and said remote host system thus enabling said expert system to access information contained in said memory dump.

20. The system of claim 12 wherein said control module includes a batch control module for enabling the analysis of said number of dumps sequentially without interruption according to specific criteria.

21. The system of claim 20 wherein said batch control module includes a module for indicating how each of said number of dumps are to be loaded, saved and cleared.

22. The system of claim 12 wherein said system further includes a plurality of remotely located host systems, each connected to said communications link for transferring requests to said expert system specifying memory dumps for analysis, said batch control module including a file for listing said requests designating a number of dumps to be analyzed according to specific criteria.

23. A method of organizing a system for automating the analysis of memory dumps generated as a result of a problem occurring on a given computer system being operated under the control of an operating system, each memory dump containing information pertaining to a number of operating system control structures containing information used by said operating system and user programs at the time the memory dump occurred, said method comprising the steps of:

storing the memory dumps to be analyzed in external storage of said system;

storing a retrieval module on said system for receiving requests, each request specifying a transfer of only a small portion of said memory dump at a time;

installing an expert system in another part of said system, said installing step including the steps of;

storing a control module in said expert system, said control module including a knowledge base containing knowledge in the form of objects, said objects including information representative of a plurality of operating system control structures organized into frames and rulesets linked to said frames so as to create decisions using a premise-conclusion if-then form of reasoning and an inference engine for interpreting said rulesets using said if-then reasoning;

storing a knowledge representation module in said expert system and coupling said representation module to said control module for providing dump control structures in response to requests received from said control module; and

,

connecting said knowledge representation module to communicate with said retrieval module to transfer requests for only specific portions of said memory dump required for creating said operating system control structures requested by said control module.

24. A method of organizing a system for automating the analysis of memory dumps generated as a result of a problem occurring on a given computer system being operated under the control of an operating system, each memory dump containing information pertaining to a number of operating system control structures containing information used by said operating system and user programs at the time the memory dump occurred, said method comprising the steps of:

storing the memory dumps to be analyzed on a remotely located host computer system;

storing a retrieval module on said host computer system for receiving requests, each request specifying a transfer of only a small portion of said memory dump at a time;

interconnecting a local computer system with said host computer system for communicating said requests through a communications link;

installing an expert system in said local computer system, said installing step including the steps of;

storing a control module in said expert system, said control module including a knowledge base containing knowledge in the form of objects, said objects including information representative of a plurality of operating system control structures organized into frames and rulesets linked to said frames so as to create decisions using a premise-conclusion if-then form of reasoning and an inference engine for interpreting said rulesets using said if-then reasoning;

storing a knowledge representation module in said expert system and coupling said representation module to said control module for providing dump control structures in response to requests received from said control module; and

,

connecting said knowledge representation module to communicate with said retrieval module to transfer requests for only specific portions of said memory dump required for creating said operating system control structures requested by said control module.