Error detection/correction and fault detection/recovery – Data processing system error or fault handling – Reliability and availability
Reexamination Certificate
1998-01-05
2001-02-06
Wiley, David A (Department: 2781)
Error detection/correction and fault detection/recovery
Data processing system error or fault handling
Reliability and availability
C714S002000
Reexamination Certificate
active
06185699
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
The present invention relates generally to data recovery in data processing systems employing a database. More particularly, the present invention relates to a method and apparatus for providing processing system availability during database management system restart recovery without loss of data integrity.
2. Description of the Related Art
Data processing systems typically manage large amounts of customer data and data generated by users within the data processing system. For many businesses, any loss in the ability to access data severely impacts the success of the business. Indeed, it is deemed catastrophic if the data is unavailable for a prolonged period of time. Thus, when system failures occur, the system must restart rapidly to minimize data outage.
Much business data is stored in databases, under the management of a database management system (DBMS). When a DBMS is restarted after system failure, part of the DBMS's restart function is to attend to work that was interrupted by the failure. This function is typically called restart recovery. It involves ensuring that committed work is not lost as a result of the failure, and ensuring that uncommitted work that persists the failure is undone.
In DBMSs that employ write-ahead logging (WAL), committed transactions are typically recovered by performing REDO (forward) processing of the log during restart recovery: uncommitted transactions are backed out by performing an UNDO (backward) processing of the recovery log. A further discussion of write-ahead logging, UNDO, REDO and other commonly known recovery methods may be found in U.S. Pat. No. 5,333,303 issued Jul. 26, 1994, entitled, “METHOD FOR PROVIDING DATA AVAILABILITY IN A TRANSACTION-ORIENTED SYSTEM DURING RESTART AFTER A FAILURE,” assigned to the assignee of the current invention and incorporated herein, and in the following references:
1. C. J. Date, “AN INTRODUCTION TO DATABASE SYSTEM”, Vol. 1, Fourth Edition, Addison Wesley Publishing Company, Copyright 1986, Chapter 18;
2. H. F. Korth et al., “DATABASE SYSTEM CONCEPTS”, McGraw-Hill Book Company, Copyright 1986, Chapter 10;
3. C. Mohan et al., “ARIES: A TRANSACTION RECOVERY METHOD SUPPORTING FINE-GRANULARITY LOCKING AND PARTIAL ROLLBACKS USING WRITE-AHEAD LOGGING”, IBM Research Report R J 6649 (63960), Jan. 23, 1989, Revised Nov. 2, 1990;
4. C. Mohan, “COMMIT_LSN: A NOVEL AND SIMPLE METHOD FOR REDUCING LOCKING AND LATCHING IN TRANSACTION PROCESSING SYSTEM”, Proceedings of the Sixteenth VLDB Conference, August 1990; and
5. C. Mohan et al., “TRANSACTION MANAGEMENT IN THE R* DISTRIBUTED DATABASE MANAGEMENT SYSTEM”, ACM transactions on Database Systems, Vol 11, No. 4, December 1986, pgs. 378-396.
DBMSs that do not use WAL typically employ force-at-commit protocols. In these systems, the result of work completed is held in cache and the resultant data is not updated until the work has been fully committed. With force-at-commit systems, it is not necessary to perform REDO processing; UNDO processing is sufficient.
In both WAL and non-WAL systems, a DBMS's restart recovery processing can take a considerable amount of time. This is especially true if one or more long-running transactions were active at the time of failure. This is because the operations performed by the transactions must be backed out and prior art DBMSs do not traditionally start to process new transactions until all data recovery processing has been completed. Hence the data outage can be large.
What is needed is a method and apparatus for performing a DBMS restart recovery with partially delayed UNDO processing that provides processing system availability. The method would allow a portion of the restart recovery to be postponed until after the DBMS involved begins accepting new work (transactions). Further, the method and apparatus should ensure data consistency during recovery, where required, to preserve data integrity.
SUMMARY OF THE INVENTION
Broadly, the present invention concerns a method and apparatus that provides processing system availability during DBMS restart recovery while maintaining data consistency. The invention delays at least part of the recovery processing until after the DBMS has begun accepting new work.
In one embodiment, the invention may be implemented to provide a method of DBMS restart recovery that allows transactions to access data that does not have restart recovery work pending. Access to data having pending restart recovery work is restricted, and a transaction is denied access to this data. In another embodiment, the inventive method allows certain transactions to access restricted data that has restart recovery work pending. These transactions are transactions that do not require data consistency. Regardless of the embodiment, the invention allows full recovery to be completed concurrent to data transaction processing that requires access to the database data. An amount of restart recovery processing may be postponed until after the DBMS has begun accepting new work requests.
In another embodiment, the invention may be implemented to provide an apparatus that provides a method of DBMS restart recovery that allows transactions to access the data being recovered. The apparatus may include a processor, storage, a user interface with or without a graphic interface, and other digital data processing components commonly found in such digital processing systems.
In still another embodiment, the invention may be implemented to provide an article of manufacture comprising a data storage device tangibly embodying a program of machine-readable instructions executable by a digital data processing apparatus to perform method steps for DBMS restart recovery that allows transactions to access data that does not have restart recovery work pending as described below. As mentioned above, the method steps allow postponement of the restart recovery of the data until after the DBMS begins accepting new work.
The invention affords its users with a number of distinct advantages. One advantage the invention provides is a way to recover data in a processing system concurrent to allowing transactions to access data so that the system can be restarted and system “down time” reduced. Another advantage is that transactions requiring data consistency are denied access to restricted data, that is, data that has restart recovery work pending. Still a further example is that the invention allows transactions not requiring data consistency to access restricted data.
The invention also provides a number of other advantages and benefits, which should be apparent from the following description of the invention.
REFERENCES:
patent: 4674038 (1987-06-01), Brelsford et al.
patent: 5155678 (1992-10-01), Fukumoto et al.
patent: 5317739 (1994-05-01), Elko et al.
patent: 5333303 (1994-07-01), Mohan
patent: 5386559 (1995-01-01), Eisenberg et al.
patent: 5421006 (1995-05-01), Jablon et al.
patent: 5455944 (1995-10-01), Haderle et al.
patent: 5530855 (1996-06-01), Satoh et al.
patent: 5561798 (1996-10-01), Haderle et al.
patent: 5574902 (1996-11-01), Josten et al.
patent: 5581750 (1996-12-01), Haderle et al.
patent: 5596710 (1997-01-01), Voigt
patent: 5640561 (1997-06-01), Satoh et al.
“Asynchronous Recovery and Reliability for a Multiuser Database System”, S. Sadjadi, et al.IBM—Technical Disclosure Bulletin,34-10B, Mar. 1992.
“Scheme for Reduced Logging for Forward-Recoverable Databases”, D. Bezviner, et al.IBM—Technical Disclosure Bulletin,37-05, May 1994.
“Object Relocation File Recovery”IBM—Technical Disclosure Bulletin,35-4B, Sep. 1992.
C. Mohan, “A Cost-Effective Method for Providing Improved Data Availability During DBMS Restart Recovery After a Failure”, 19th International Conference on Very Large Data Bases, Proceedings, pp. 368-379, Published: Palo Alto, CA, USA, 1993.
C. Mohan, et al. “Recovery and Coherency-Control Protocols for Fast Intersystem Page Transfer and Fine-Granularity Locking in a Shared Disks Transaction Environment”, 17th International Conference on Very Large Data Bases, Pr
Haderle Donald J.
Watts Julie Ann
Foerster Ingrid M.
International Business Machines - Corporation
Sawyer Law Group LLP
Wiley David A
LandOfFree
Method and apparatus providing system availability during... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Method and apparatus providing system availability during..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method and apparatus providing system availability during... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2571645