Error detection/correction and fault detection/recovery – Data processing system error or fault handling – Reliability and availability
Reexamination Certificate
1998-07-30
2001-09-04
Le, Dieu-Minh (Department: 2184)
Error detection/correction and fault detection/recovery
Data processing system error or fault handling
Reliability and availability
C714S010000
Reexamination Certificate
active
06286110
ABSTRACT:
BACKGROUND OF THE INVENTION
The invention relates generally to fault-tolerant transaction processing systems formed from multiple processor units to maintain information collections (e.g., a database), and to from time-to-time modify that collection. More particularly, the invention relates to a method for detecting the loss of a processor unit participating in a transaction that is in the process of changing the state of the information collection maintained by the system.
Concern about protecting and maintaining the integrity of information collections in the face of updates and changes to that information has resulted in the development of a programmatic construct called a transaction. A useful definition of a transaction is that it is an explicitly delimited operation, or set of related operations, that change or otherwise modify the content of the information collection or database from one consistent state to another. Changes are treated as a single unit in that all changes of a transaction are formed and made permanent (the transaction is “committed”) or none of the changes are made permanent (i.e., the transaction is “aborted”). If a failure occurs during the execution of a transaction, the transaction can be aborted and whatever partial changes were made to the collection can be undone to leave it in a consistent state.
Typically, transactions are performed under the supervision of a transaction manager facility (TMF). In geographically distributed systems, such as multiple processor unit systems or “clusters” (i.e., a group of independent processor units managed as a single system), each processor unit will have its own TMF component to coordinate transaction operations conducted on that processor unit. The processor unit at which (or on which) a transaction begins is sometimes called the “beginner” processor unit, and the TMF component of that processor unit will operate to coordinate those transactional resources remote from its resident processor unit (i.e., resources managed by other processor units). Those TMF components running on processor units managing resources enlisted in a transaction are “participants” in the transaction. And, it is the TMF component of the beginner processor unit that initiates the steps taken.
Fault tolerance is another important feature of transaction processing. Being able to detect and tolerate faults allows the integrity of the collection being managed by the system to be protected. Although a number of different methods and facilities exist, one particularly effective fault tolerant technique is the “process-pair” technique as it is sometimes called. According to this technique, each process running on each processor unit of a multiple processor system will have a backup process on another processor unit of the system. If a process, or the processor unit upon which the process is running, fails, that failure will bring into operation the backup process to take over the operation of the lost (primary) process. If that failure occurs during a transaction in which the lost process was a participant, the backup will decide whether or not to notify the beginner processor unit to abort the transaction and begin over again. In this way the state of the collection managed by the system remains consistent.
The process-pair paradigm uses what is sometimes called a “Heartbeat” or “I'm Alive” approach to detecting failure of a processor unit. Briefly, according to this approach, each processor unit is required to periodically broadcast an “I'm Alive” message to all other processor units of the system. If the heartbeat message of a particular processor unit has not received its siblings within a predetermined period of time, the silent processor unit is assumed to have failed and all primary processes resident on or associated with the now assumed failed processor unit will be taken over by their backup processes on the other processor units of the system. Each backup process, when taking over, will investigate whether or not it was involved in a transaction, and if so, decide whether or not to abort the transaction. An example of the process-pair concept using “I'm Alive” detection of processor failures can be found in U.S. Pat. No. 4,817,091.
But there are times when a process may not have a back-up process—even though resident in a multiple processor system employing process-pair fault tolerance. If that process is a participant in a transaction, and the processor unit upon which that process runs fails, the TMF component on the beginner processor unit may be aware of the failure and the loss of the processor unit, but not of the participant process. If a modification to be made by the participant process was never made, yet the other participants were able to complete their modifications, the result can severely damage the integrity of the managed collection, i.e., the collection is now inconsistent.
Accordingly, it can be seen that there exists a need for a fault-tolerant method of notifying a transaction manager of the loss of a participant process as a result of the associated processor unit failing, separate and apart from employment of a process-pair fault detection technique.
SUMMARY OF THE INVENTION
Whereas prior implementations of transaction processing systems had available the process-pair paradigm to notify a transaction manager of the loss of a resource participating in a transaction, the present invention provides an alternate approach. The invention provides a simple yet effective facility for allowing a transaction manager to know what processor units of a multiple processor system are participating in a transaction. When the transaction manager is notified of a processor unit failure, a check is made to determine if the failed processor unit has resources that were called upon to participate in any transaction. If so, a decision is made as to whether or not to abort the transaction.
According to the present invention, when a transaction is started on a “beginner” processor unit of a multiple processor system, and there are resources managed by other processor units of the system enlisted to perform work on behalf of the transaction, information identifying each of these other processor units is maintained by the transaction manager. If, before the transaction is completed, a processor unit as having failed, the TMF component on the beginner processor unit will be notified so that it can examine the information. If that examination reveals a transaction that involved resources managed by the failed processor unit, a decision is made whether to abort the transaction.
In an alternate embodiment, the invention is used in association with the process-pair technique (described above); that is, the invention may be employed in a system in which the process-pair technique is used to notify a transaction manager that a participant resource has been lost, yet detect the loss of a participant processor that does not have backup process. According to this embodiment of the invention, if a process that lacks a backup process (and, therefore is not able to use the process pair technique) is enlisted as a participant in a transaction a processor unit of the system other than the beginner processor unit, the identity of that process will be associated with that processor unit on which it is running in the information maintained by the TMF component on that processor unit. And, at the same time, the identity of that processor unit is made known to the TMF component of the beginner processor unit.
A number of advantages are achieved by the present invention. First, the present invention provides a fault-tolerant environment for multi-processor architectures without the addition of a process-pair implementation.
Further, even if process-pair fault-tolerant techniques are employed, those processes that may be too extravagant for a backup process, the present invention provides a technique for permitting such processes to notify a beginner TMF component of its loss.
These and other advantages and aspects of the invention will be
Carley William J.
Gondi Albert C.
Klein Johannes
Lanka Sitaram V.
Compaq Computer Corporation
Le Dieu-Minh
Oppenheimer Wolff & Donnelly LLP
Sherry Leah
LandOfFree
Fault-tolerant transaction processing in a distributed... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Fault-tolerant transaction processing in a distributed..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Fault-tolerant transaction processing in a distributed... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2504164