Fault-tolerant distributed database system and method for the management of correctable subtransaction faults by the global transaction source node

• US 5,247,664 A
• Filed: 03/28/1991
• Issued: 09/21/1993
• Est. Priority Date: 03/28/1991
• Status: Expired due to Fees

• Current Assignee: Amoco Corporation (BP PLC)
• Original Assignee: Amoco Corporation (BP PLC)
• Inventors: Reyes, Tom C., Olson, Peter L., Thompson, Glenn R., Vassaur, Bing, Thomas, George F. III, Lee, Wen F.
• Primary Examiner(s): Lee, Thomas C.
• Assistant Examiner(s): AMSBURY, WAYNE P
• Application Number:
  US07/676,605
• Time in Patent Office: 908 Days
• Field of Search: 395/650, 395/700, 395/425, 395/575, 395/600, 364/200, 370/58.1, 371/12
• US Class Current: 1/1
• CPC Class Codes: G06F 11/1402 Saving, restoring, recovering or retrying
• CPC Class Codes: G06F 2201/87 Monitoring of transactions

Abstract

A fault-tolerant method and system for processing global transactions in a distributed database system. If a fault occurs in the distributed database system, a transaction management system will suspend the processing of a transaction and renew it when the fault is remedied. The transaction management system identifies one or more sites in the distributed database system that contain data needed for a global transaction and divides each transaction into subtransactions for processing at one or more identified sites. If a fault occurs in the distributed database system, the transaction management system maintains the status of each subtransaction. When a fault in the distributed database system is corrected, the transaction management system updates the status of each subtransaction and restarts processing of the transactions.

Claims

1. A fault-tolerant method of processing a global transaction in a distributed database system having a plurality of sites and affected by faults that prevent the processing of subtransactions and that are correctable, comprising:

(a) upon receipt of a global transaction at one of the sites in the distributed database system, identifying a number of sites among the plurality of sites in the distributed database system that contain data needed for the global transaction;

(b) dividing the global transaction by the said one site into subtransactions for processing at the identified sites and communicating the subtransactions to each appropriate identified site for processing;

(c)(1) monitoring at said one site the processing of all the subtransactions,

(c)(2) detecting at said one site a fault in the distributed database system that prevents the processing of a subset of the subtransactions and continuing the processing of the unaffected subtransactions,

(c)(3) interrupting by said one site the processing of the affected subset of subtransactions,

(c)(4) maintaining by said one site a status of all the subtransactions, and

(c)(5) correcting said fault;

(d) upon correction of the fault, updating the status by said one site of each subtransaction affected by the fault; and

(e)(1) restarting by said one site the processing of the affected subtransactions at the identified sites, and

(e)2) continuing the processing of the unaffected subtransactions.

2. The method of claim 1, wherein the distributed database system is a heterogeneous distributed database system.

3. The method of claim 1, wherein maintaining the status of each subtransaction further comprises sending the last executed portion of each subtransaction to stable storage at the local site.

4. The method of claim 3, further including presenting to a user an indication of the status of a subtransaction selected from the group including:

failed, suspended, completed and rescheduled.

5. The method of claim 3, wherein updating the status of each subtransaction comprises retrieving from stable storage previously executed portions of each subtransaction at the local site or other sites in the distributed database system.

6. The method of claim 5, further including presenting to a user an indication that the global transaction has recommenced processing after suspension.

7. The method of claim 1, wherein processing the subtransactions at the identified one or more sites further comprises:

(a) allocating a server at each site for each identified subtransactions;

(b) sending the subtransaction to each identified site;

(c) executing the subtransaction to perform an operation by each of the servers;

(d) rescheduling selected operations to be performed at selected sites;

(e) merging data from local servers and servers at other sites; and

(f) representing the merged data to a user at said one site.

8. A fault-tolerant transaction management system in a distributed database system that has a plurality of processing sites and that is affected by correctable faults, comprising:

(a) means for submitting a global transaction at a first one of the sites in the distributed database system;

(b) means for identifying a number of sites among the plurality of sites in the distributed database system that contain data needed for the global transaction;

(c) means for dividing the global transaction by the said one site into subtransactions for processing at the identified sites and for distributing each of said subtransactions to each of said identified sites;

(d)(1) monitoring means at said first site for monitoring the processing of all the subtransactions,

(d)(2) means for detecting at said one site a fault in the distributed database system that prevents the processing of a subset of the subtransactions and continuing the processing of the unaffected subtransactions,

(d)(3) means for interrupting by said one site the processing of the affected subset of subtransactions,

(d)(4) means for maintaining by said one site a status of all the subtransactions, and

(d)(5) means for correcting said fault;

(e) means for updating, upon correction of the fault, the status by said one site of each subtransaction affected by the fault; and

(f)(1) means for restarting by said one site the processing of the affected subtransactions at the identified sites, and

(f)(2) means for continuing the processing of the unaffected subtransactions.

9. The system of claim 8, wherein the distributed database system is a heterogeneous distributed database system.

10. The system of claim 8, wherein each identified site has stable storage, and means for sending a last executed portion of each subtransaction to stable storage at that site.

11. The system of claim 10, further including means for presenting to a user at the first site an indication that a subtransaction has not been processed due to said fault.

12. The system of claim 8 further including means for retrieving from stable storage previously executed portions of each subtransaction at each site in the distributed database system where the subtransaction was executed.

13. The system of claim 12, further including means for presenting to a user an indication that the transaction has recommenced processing after suspension.

14. The system of claim 8, wherein the means for continuing the processing further comprises:

(a) means for allocating a server at each site for each subtransaction;

(b) means for sending a subtransaction to each identified site;

(c) means for executing a subtransaction to perform an operation by each of the servers;

(d) means for rescheduling selected operations to be performed at selected sites;

(e) means for merging data from each server executing the subtransaction in the distributed database system; and

(f) means for presenting merged data to a user at the first site.