System for localizing field replaceable unit failures employing automated isolation procedures and weighted fault probability encoding

• US 5,561,760 A
• Filed: 09/22/1994
• Issued: 10/01/1996
• Est. Priority Date: 09/22/1994
• Status: Expired due to Fees

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Gross, Andrew L., Torrens, Douglas E., Ferris, Thomas M., Hill, Mark E.
• Primary Examiner(s): Beausoliel, Jr., Robert W.
• Assistant Examiner(s): Wright, Norman M.
• Application Number:
  US08/310,540
• Time in Patent Office: 740 Days
• Field of Search: 395/575, 395/182.06, 395/183.01, 395/185.07, 395/264.4, 395/267.7, 395/181, 395/183.02, 395/184.01, 364/264.4, 364/267.7
• US Class Current: 714/25
• CPC Class Codes: G06F 11/2273 Test methods

Abstract

A system for the automated isolation and report of a Field Replaceable Unit (FRU) and/or a Replaceable Component (RC) responsive to a stream of Fault Symptom Code (FSCs) produced in response to a system fault. The process is entirely automated and can be stored in microcode to allow immediate updating of FSC to FRU/RC correspondences responsive to system design revisions. Selected FRU/RC isolation procedures are included in the FRUs/RCs for automatic execution as part of the isolation process. Failure probability data, including probability distribution data, are included in each FRU/RC to support an alternative weighted failure probability (WFP) isolation scheme to isolate a faulty FRU/RC even when the automated isolation procedures do not provide an unambiguous result.

Claims

1. In a system having a plurality of elements including Field Replaceable Units (FRUs) each having one or more Replaceable Components FRCs), said system having means for producing, responsive to a failure, a plurality of error data, a machine implemented process for isolating and reporting an element failure in said system, said process comprising the steps of:

initially, detecting an element failure, followed by;

(a) in response to said element failure producing a plurality of Fault Symptom Code error data structures (FSCs) each describing a fault symptom with associated probability data and one or more Failure Candidate (FC) identifiers, said FC identifiers including FRU identifiers (FIDs) each listing one or more FRUs and RC identifiers (CIDs) each listing one or more RCs, selected ones of said FC identifiers each further including isolation procedure data for isolating one of a plurality of said FRUs or RCs;

(b) combining, for each said FID in said FSC plurality, said associated probability data with the number of times that said each FID is specified in said FSC plurality to produce a Weighted Failure Probability (WFP) for said each FID; and

(c) selecting and reporting the FRU having the maximum said WFP value as a faulty FRU.

2. The machine-implemented process of claim 1 wherein said selecting step (c) comprises the steps of:

(c.1) selecting from said FSC plurality each said FID having a WFP value within a predetermined deviation from said maximum WFP value to form a first FRU group;

(c.2) selecting a first said FID that is associated with said first FRU group;

(c.3) executing one or more of any said isolation procedures associated with said first FID to identify an isolated FRU if any; and

(c.4) selecting and reporting said isolated FRU if any as said faulty FRU, otherwise biasing said WFP values to reduce the contributions from said first FSC before selecting and reporting said FRU having the maximum said WFP value as said faulty FRU.

3. The machine-implemented process of claim 2 further comprising the steps of:

(d) identifying in said FSC plurality all said CIDs associated with said faulty FRU;

(e) combining, for each said associated CID in said FSC plurality, said associated probability data with the number of times that said each CID is specified in said FSC plurality to produce a second WFP value for said each associated CID; and

(f) selecting and reporting the RC having the maximum said second WFP value as a faulty RC.

4. The machine-implemented process of claim 3 wherein said second selecting step (f) comprises the steps of:

(f.1) selecting from said FSC plurality each said associated CID having a second WFP value within a predetermined deviation from said maximum second WFP value to form a first RC group;

(f.2) selecting a first said FRU that is associated with said first RC group;

(f.3) executing one or more of any said isolation procedures associated with said first FRU to identify an isolated RC if any; and

(f.4) selecting and reporting said isolated RC if any as said faulty RC, otherwise biasing said second WFP values to reduce the contributions from said first FRU before selecting and reporting said RC having the maximum said WFP value as said faulty RC.

5. The machine-implemented process of claim 4 wherein said producing step (a) comprises the steps of:

(a.1) if said FSC plurality is small, performing the steps of

(a.1.1) executing one or more of said isolation procedures if any associated with said FSC plurality to identify an isolated FRU if any, and

(a.1.2) selecting and reporting said isolated FRU if any as said faulty FRU;

otherwise

(a.2) doing nothing.

6. The machine-implemented process of claim 1 wherein said producing step (a) comprises the steps of:

(a.1) if said FSC plurality is small, performing the steps of

(a.1.1) executing one or more of said isolation procedures if any associated with said FSC plurality to identify an isolated FRU if any, and

(a.1.2) selecting and reporting said isolated FRU if any as said faulty FRU;

otherwise

(a.2) doing nothing.

7. A fault reporting system for isolating and reporting a faulty element in a system having a plurality of elements including Field Replaceable Units (FRUs) each having one or more Replaceable Components (RCs), said fault reporting system comprising:

error output means coupled to said faulty element for producing a plurality of error data structures including one or more Fault Symptom Codes (FSCs) each describing a fault symptom with associated probability data and one or more Failure Candidate (FC) identifiers, said FC identifiers including FRU identifiers (FIDs) each listing one or more FRUs and RC identifiers (CIDs) each listing one or more RCs, selected ones of said FC identifiers each further including isolation procedure data for isolating one of a plurality of said FRUs or RCs;

first combining means coupled to said error output means for combining said associated probability data with the number of times that said each FRU is specified in said FSC plurality to produce a Weighted Failure Probability (WFP) for said each FID; and

FRU output means coupled to said combining means and said error output means for selecting and reporting said FRU having the maximum said WFP value as a faulty FRU.

8. The fault reporting system of claim 7 further comprising:

FRU selecting means in said FRU output means for selecting from said FSC plurality each said FID having a WFP value within a predetermined deviation from said maximum WFP value to form a first FRU group;

FSC mapping means in said FRU output means coupled to said FRU selecting means for selecting a first said FID that is associated with said first FRU group;

FRU isolating means coupled to said FRU selecting means for executing one or more of any said isolation procedures associated with said first FID to identify an isolated FRU if any; and

FRU biasing means coupled to said FRU isolating means for biasing said WFP values to reduce the contributions from said first FID responsive to failure of said isolation procedures associated with said first FID.

9. The fault reporting system of claim 8 further comprising:

RC identifying means coupled to said FRU output means for identifying in said FSC plurality all said CIDs associated with said faulty FRU;

second combining means coupled to said RC identifying means for combining said associated probability data with the number of times that said each CID is specified in said FSC plurality to produce a second WFP value for said each associated CID; and

RC output means coupled to said second combining means for selecting and reporting said associated RC having the maximum said second WFP value as a faulty RC.

10. The fault reporting system of claim 9 further comprising:

RC selecting means in said RC output means for selecting from said FSC plurality each said associated CID having a second WFP value within one standard deviation of said maximum second WFP value to form a first RC group;

FRU mapping means in said RC output means coupled to said second selecting means for selecting a first said FRU that is associated with said first RC group;

RC isolating means coupled to said FRU mapping means for executing one or more of any said isolation procedures associated with said first FRU to identify an isolated RC if any; and

RC biasing means coupled to said RC isolating means for biasing said second WFP values to reduce the contributions from said first FRU responsive to failure of said any isolation procedures.

11. The fault reporting system of claim 7 further comprising:

RC identifying means coupled to said FRU output means for identifying in said FSC plurality all said CIDs associated with said faulty FRU;

second combining means coupled to said RC identifying means for combining said associated probability data with the number of times that said each CID is specified in said FSC plurality to produce a second WFP value for said each associated CID; and

RC output means coupled to said second combining means for selecting and reporting said associated RC having the maximum said second WFP value as a faulty RC.