Methods and apparatus for monitoring events and implementing corrective action in a computer system

• US 5,893,083 A
• Filed: 03/19/1996
• Issued: 04/06/1999
• Est. Priority Date: 03/24/1995
• Status: Expired due to Term

• Current Assignee: QNAP Systems Incorporated
• Original Assignee: Hewlett-Packard Company (HP Inc.)
• Inventors: Eshghi, Kave, Moreau, Jean-Jacques
• Primary Examiner(s): Hafiz, Tariq R.
• Assistant Examiner(s): RHODES, JASON W
• Application Number:
  US08/618,159
• Time in Patent Office: 1,113 Days
• Field of Search: 395/50, 395/200.11, 395/60, 395/200.53, 395/200.54, 395/670-678, 395/180-181, 395/182.01-182.03, 395/183.01, 395/183.02 , 395/184.01, 395/185.01, 364/281.3, 364/550, 706/45, 706/46
• US Class Current: 706/45
• CPC Class Codes: G06F 11/2257 using expert systems
• CPC Class Codes: G06N 5/043 Distributed expert systems; Blackboards

Abstract

Apparatus for assisting management of services provided by a computer system includes an inferencing engine (30) for carrying out inferencing operations on a declarative model (24) of a service, using facts about the system stored in a fact base (32). A resident goal store (102) contains declarative definitions of goals which concern availability of services and which it is desirable for the system to continue to satisfy; these definitions are linked to associated facts in the fact base. The service model (24) includes definitions of events which can occur in the system and may affect availability of services, and definitions of actions which can be taken to modify the configuration of the system. When occurrence of an event defined in the service model is reported to the apparatus, the event definition is used to guide analysis of the event report and appropriate updating of the fact base. Goals which are linked to the updated facts are then examined to assess whether the goals are still satisfied. If a goal is no longer satisfied the service model is searched for actions which can re-configure the system to enable the goal to be re-satisfied. If a goal involves information about an entity in a part of the system managed by a second, different management apparatus (10D), the second apparatus can be requested to establish a sub-goal concerning the status of that entity. Thereafter the second apparatus takes appropriate action, autonomously, to keep the sub-goal satisfied, and reports back only if it is unable to satisfy the sub-goal.

Claims

1. A system management method of monitoring occurence of and attempting to remedy effects of events affecting a service provided by a computer system made up of cooperating physical and logical entities, said method comprising the steps of:

providing for said service a declarative model specifying requirements needing to be met for said service to be available, said requirements being set out in terms of the entities required and their inter-relationships;

specifying in respect of at least one aspect of said service a goal to be satisfied by said system;

providing a fact base for holding facts relating to the system;

identifying at least one fact which relates to the system and upon which said goal depends, and including that fact in said fact base;

determining whether said goal is satisfied and thereby establishing at least one link indicating a dependency relationship between said goal and said at least one fact, and including said link in said fact base;

defining at least one event which can occur in the system and whose occurrence in the system can affect validity of said fact; and

detecting occurrence of said event, and thereupon;

determining whether said fact is valid or invalid;

if said fact has become invalid, determining whether said goal is still satisfied by performing inferencing operations on the declarative model by referring to said fact base so as to ascertain whether a requirement relevant to said coal is met by the system;

if said goal is no longer satisfied, seeking an operation which will enable said goal to be re-satisfied; and

performing said operation.

2. The method of claim 1, wherein the step of determining whether said goal is satisfied is performed when said goal is first provided, wherein said links are established during that step.

3. The method of claim 1, including the step of inspecting said fact base for any link relating to a fact when that fact is updated, and, if such a link is found, determining whether a goal indicated by that link is still satisfied.

4. The method of claim 1, including the step of analyzing occurrence of said event to identify and establish validity of a fact related to the event.

5. The method of claim 1, including the step of testing whether an operation will enable said goal to be re-satisfied by temporarily updating said fact base to indicate consequences of effecting that operation, and determining whether said goal is satisfied by reference to said fact base in its temporarily updated state.

6. The method of claim 1, including the step of providing a description of reasons for a goal not being satisfied, if no operation can be found which will enable said goal to be re-satisfied.

7. The system management method of claim 1 wherein said event affects availability of the service.

8. System management apparatus for monitoring occurrence of and attempting to remedy effects of events affecting a service intended to be provided by a computer system made up of cooperating physical and logical entities, said apparatus comprising:

a declarative model specifying requirements needing to be met for said service to be available, said requirements being set out in terms of the entities required and their inter-relationships;

an inference engine for carrying out inferencing operations in relation to said declarative model;

a specification of a goal to be satisfied by said system in respect of at least one aspect of said service;

a fact base for holding facts relating to the system;

an identification of at least one fact which relates to the system and upon which said goal depends, said fact being included in said fact base;

at least one link being established by determining whether said goal is satisfied, said link indicating a dependency relationship between said goal and said at least one fact, and said link being included in said fact base;

a definition of at least one event which can occur in the system and whose occurrence in the system can affect validity of said fact; and

means for detecting occurrence of said event, and thereupon;

determining whether said fact is valid or invalid;

if said fact has become invalid, causing said inferencing engine to perform inferencing operations on the declarative model and causing reference to be made to said fact base for ascertaining whether a requirement relevant to said goal is met by the system, and determining whether said goal is still satisfied;

it said goal is no longer satisfied, seeking an operation for enabling said goal to be re-satisfied; and

performing said operation.

9. The system management apparatus of claim 8 wherein said event affects availability of the service.

10. The apparatus of claim 8, wherein the determination of whether the goal is satisfied is performed and the links are established when the goal is first provided.

11. A system management method of monitoring occurrence of and attempting to remedy effects of events in a computer system made up of cooperating physical and logical entities, said entities being logically arranged into groups and each group including a management entity, and said events affecting a service provided by entities in a first group, said method comprising the steps of:

providing for said service a declarative model specifying requirements needing to be met for said service to be available, said requirements being set out in terms of the entities required and their inter-relationships;

specifying in respect of at least one aspect of said service a goal to be satisfied by said system;

identifying, in a first management entity in said first group, that satisfaction of said goal requires a sub-goal to be satisfied, and that satisfaction of said sub-goal involves system entities in a second group different from said first group;

communicating to a second management entity in said second group a requirement to determine whether said sub-goal is satisfied;

providing a fact base for holding facts which relate to the second group;

identifying, in said second management entity, at least one fact which relates to said second group and upon which said sub-goal depends, and including that fact in said fact base;

determining, in said second management entity, whether said sub-goal is satisfied and thereby establishing at least one link indicating a dependency relationship between said sub-goal and said at least one fact, and including said link in said fact base;

defining, in said second management entity, at least one event which can occur in the second group whose occurrence in said second group can affect validity of said fact;

maintaining a watch, in said second management entity and autonomously of said first management entity, for occurrence of said event; and

upon detecting occurrence of said event, in said second management entity;

determining whether said fact is valid or invalid;

if said fact has become invalid, determining whether said sub-goal is still satisfied by performing inferencing operations on the declarative model and ascertaining whether a requirement relevant to said sub-goal is met by referring to said fact base;

if said sub-goal is no longer satisfied, seeking an operation which will enable said sub-goal to be re-satisfied;

performing said operation if one can be found; and

if no such operation can be found, communicating non-satisfaction of said sub-goal to said first management entity.

12. The system management method of claim 11 wherein the at least one event affects availability of the service.

13. The method of claim 11, wherein the determination of whether the goal is satisfied is performed and the links are established when the goal is first provided.

14. System management apparatus for monitoring occurrence of and attempting to remedy effects of events in a computer system made up of cooperating physical and logical entities, said entities being logically arranged into groups and each group including a management entity, and said events affecting a service intended to be provided by entities in a first group, said apparatus comprising:

a first management entity for a respective first one of said groups;

a second management entity for a respective second one of said groups different from said first group;

a declarative model specifying requirements needing to be met for said service to be available, said requirements being set out in terms of the entities required and their inter-relationships;

an inference engine in said second management entity for carrying out inferencing operations in relation to said declarative model;

a specification of a goal to be satisfied by said system in respect of at least one aspect of said service;

an identification, in said first management entity, that satisfaction of said goal requires a sub-goal to be satisfied, and that satisfaction of said sub-goal involves system entities in said second group;

means for communication to said second management entity from said first management entity a requirement to determine whether said sub-goal is satisfied;

a fact base storing facts relating to the second group;

an identification, in said second management entity, of at least one fact which relates to said second group and upon which said sub-goal depends, said fact being included in said fact base;

at least one link, said one link being established by determining, in said second management entity, whether said sub-goal is satisfied, said link indicating a dependency relationship between said sub-goal and said at least one fact, and being included in said fact base;

a definition, in said second management entity, of at least one event which can occur in the second group and whose occurrence in said second group can affect validity of said fact;

means in said second management entity for maintaining a watch, autonomously of said first management entity, for occurrence of said event and upon detection thereof;

determining whether said face is valid or invalid;

if said fact has become invalid, causing said inferencing engine to perform inferencing operations on the declarative model, ascertaining whether a requirement relevant to said sub-goal is met by making reference to said fact base to determine whether said sub-goal is still satisfied;

if said sub-goal is no longer satisfied, seeking an operation which will enable said sub-goal to be re-satisfied;

performing said operation if one can be found; and

if no such operation can be found, communicating non-satisfaction of said sub-goal to said first management entity.

15. The system management apparatus of claim 14 wherein the at least one event affects availability of the service.

16. The apparatus of claim 14, wherein the determination of whether the goal is satisfied is performed and the links are established when the goal is first provided.