Replicated data processing method in distributed processing system

• US 5,333,265 A
• Filed: 10/22/1991
• Issued: 07/26/1994
• Est. Priority Date: 10/22/1990
• Status: Expired due to Term

• Current Assignee: Hitachi, Ltd.
• Original Assignee: Hitachi, Ltd.
• Inventors: Suzuki, Hitoshi, Fujise, Hiroshi, Mori, Kinji, Hirasawa, Shigeki, Orimo, Masayuki, Takeuchi, Masuyuki
• Primary Examiner(s): Shaw, Dale M.
• Assistant Examiner(s): DINH, DUNG C
• Application Number:
  US07/780,337
• Time in Patent Office: 1,008 Days
• Field of Search: 395/200, 395/600, 395/650, 395/575, 371/7, 371/47.1
• US Class Current: 709/201
• CPC Class Codes: G06F 15/161 Computing infrastructure, e.g. computer clusters, blade chassis or hardware partitioning casings, cabinets, racks or drawers for data centers H05K5/00

Abstract

A replicated data processing method and apparatus for a distributed processing system are provided in which a plurality of processors are connected to one another through a common transmission line. The processors perform processes using replicated data produced by resources such as files replicated in respective processors. Any processor issues an access request and broadcasts the request to the common transmission line. User'"'"'s program of the processor is initiated by a first-received data in response to the request, and checks matching between the first-received data of the replicated data and the subsequently received data to repeat a retry processing for the request until a mismatch is removed.

Claims

1. A replicated data processing method for a distributed processing system having a plurality of processors connected to each other via a common transmission medium and information resources replicated in different processors, said method, being executed in each processor of said plurality of processors, comprising the steps of:

outputting data from said information resources in said processor to said common transmission medium;

receiving, at said processor, replicated data from said plurality of processors on said common transmission medium, said replicated data having first received replicated data and replicated data subsequently received;

performing processing using said first received replicated data beginning at a time before receipt of said replicated data subsequently received;

checking content correspondence between said first received replicated data and said replicated data subsequently received; and

performing an abnormality processing in response to a result of said checking step indicating that a content correspondence is determined to be abnormal.

2. A replicated data processing method for a distributed processing system according to claim 1, wherein said information resources correspond to files replicated in each of external storage media respectively connected to said plurality of processors, and an arbitrary processor among said plurality of processors issues an access request to access said replicated files, receives replicated response data from said replicated files corresponding to said access request, performs said processings by using first received replicated response data from among said replicated response data, and performs said checking between first received replicated response data and replicated response data and replicated response data received thereafter.

3. A replicated data processing method for a distributed processing system according to claim 2, wherein said arbitrary processor which has issued said access request does not generate a next file access request until checking of content correspondence between said replicated response data is completed.

4. A replicated data processing method for a distributed processing system according to claim 3, wherein upon detection of an abnormality by checking of content correspondence between said replicated response data, said arbitrary processor which has issued said access request performs a first retry processing to issue an instruction for cancelling data established by said access request, performs a second retry processing for generating said access request again after said data has been cancelled by said first retry processing, and repeats said first retry processing and said second retry processing until an abnormality between said replicated response data is no longer detected.

5. A replicated data processing method for a distributed processing system according to claim 4, wherein said arbitrary processor which has issued said access request executes said second retry processing at a predetermined time period after said first retry processing.

6. A replicated data processing method for a distributed processing system according to claim 4, wherein said arbitrary processor which has issued said access request collects all replicated response data fed from said replicated files and performs said processing after checking of content correspondence among all of said collected replicated response data and determination of content correspondence between said replicated response data conducted by said first retry processing and said second retry processing.

7. A replicated data processing method for a distributed processing system according to claim 4, wherein said arbitrary processor which has issued said access request performs said first retry processing and said second retry processing on a basis of a check on content correspondence among all replicated response data fed from said replicated files and collected within a preset time.

8. A replicated data processing method for a distributed processing system according to claim 7, wherein if all of replicated response data from said replicated files cannot be collected within said preset time, said arbitrary processor which has issued said access request performs a fault processing for limiting said checking on content correspondence to replicated response data collected within said preset time and performs said checking, said first retry processing and said second retry processing after performing said fault processing.

9. A replicated data processing method for a distributed processing system according to claim 4, wherein as said first retry processing, said arbitrary processor which has issued said access request issues, if said access request was a read command, a free command for canceling an occupation state of a record established by exclusive control of said read command, issues, if said access request was an addition command, an erase command for delating a record added by said addition command, and issues, if said access request was an exclusive open command, a close command for canceling an occupation state of a file established by said exclusive open command.

10. A replicated data processing apparatus in a distributed processing system having a plurality of processors connected to each other via a common transmission medium and information resources replicated in different processors, wherein said apparatus, being included in each processor of said plurality of processors, comprises:

an output circuit for outputting data from said information resources in said processor to said common transmission medium;

a receiving circuit for receiving, at said processor, replicated data from said plurality of processors on said common transmission medium, said replicated data having first received replicated data and replicated data subsequently received;

a processing circuit for performing processing using said first received replicated data beginning at a time before receipt of replicated data subsequently received;

a checking circuit for checking content correspondence between said first received replicated data and said replicated data subsequently received; and

an abnormality testing circuit for performing an abnormality processing in response to a output of said checking circuit indicating that a content correspondence is determined to be abnormal.

11. A replicated data processing apparatus according to claim 10, further comprising:

external storage media respectively connected to said plurality of processors for storing, as said information resources, files replicated with respect to said plurality of processors;

wherein an arbitrary processor among said plurality of processors issues an access request to access said replicated files, receives replicated response data from said replicated files corresponding to said access request, performs said processing by using first received replicated response data from among said replicated response data, and performs said checking between first received replicated response data and replicated response data received thereafter.

12. A replicated data processing apparatus according to claim 11, wherein said arbitrary processor which has issued said access request does not generate a next file access request until checking of content correspondence between said replicated response data is completed.

13. A replicated data processing according to claim 12, wherein upon detection of an abnormality by checking of content correspondence between said replicated response data, said arbitrary processor which has issued said access request performs a first retry processing to issue an instruction for cancelling data established by said access request, performs a second retry processing for generating said access request again after said data has been cancelled by said first retry processing, and repeats said first retry processing and said second retry processing until an abnormality between said replicated response data is no longer detected.

14. A replicated data processing according to claim 13, wherein said arbitrary processor which has issued said access request executes said second retry processing at a predetermined time period after said first retry processing.

15. A replicated data processing according to claim 13, wherein said arbitrary processor which has issued said access request collects all replicated response data fed from said replicated files and performs said processing after checking of content correspondence among all of said collected replicated response data and determination of content correspondence between said replicated response data conducted by said first retry processing and said second retry processing.

16. A replicated data processing according to claim 13, wherein said arbitrary processor which has issued said access request performs said first retry processing and said second retry processing on a basis of a check on content correspondence among all replicated response data fed from said replicated files and collected within a preset time.

17. A replicated data processing according to claim 16, wherein if all of replicated response data from said replicated files cannot be collected within said preset time, said arbitrary processor which has issued said access request performs a fault processing for limiting said checking on content correspondence to replicated response data collected within said preset time and performs said checking, said first retry processing and said second retry processing after performing said fault processing.

18. A replicated data processing according to claim 13, wherein as said first retry processing, said arbitrary processor which has issued said access request issues, if said access request was a read command, a free command for canceling an occupation state of a record established by exclusive control of said read command, issues, if said access request was an addition command, an erase command for delating a record added by said addition command, and issues, if said access request was an exclusive open command, a close command for canceling an occupation state of a file established by said exclusive open command.