Distributed transaction execution method and device

Abstract

The embodiment of the invention provides a distributed transaction execution method and device, and the method comprises the steps: updating a field value of a target field in a service table from first service data to second service data in a first stage of first distributed transaction execution; correspondingly recording the identifier of the target field and the second service data in a redo log; then, in the second stage of the execution of the first distributed transaction, when it is determined that rollback is needed, judging whether the field value of the target field in the businesstable is consistent with that in the redo log or not; and when the judgment result is that the field values of the target fields in the service table and the redo log are consistent, executing a rollback operation, thereby checking whether dirty writing occurs or not, and avoiding service data loss and inconsistency caused by dirty writing.

Description

technical field [0001] One or more embodiments of this specification relate to the field of computers, and in particular, to methods and devices for executing distributed transactions. Background technique [0002] If there is only one resource for a transaction operation, then this transaction is a local transaction; the transaction on a single database usually seen is a local transaction; if there are multiple resources for a transaction operation, and these resources are distributed on different nodes on the network , then this transaction is a distributed transaction. [0003] A transaction modifies the data from state A to another state B. Before, during, and after the execution of the transaction, the outside world will not see other states except A and B. This consistency is called strong consistency; A transaction modifies the data from state A to another state B. During the execution of the transaction, the outside world will see the state C other than A and B, but...

AI Technical Summary

Problems solved by technology

That is to say, when multiple different distributed transactions modify the same row of data at the same time, dirty writes will occur, resulting in inconsistencies in business data

[0004] At present, "row lock" is usually used to prevent multiple distributed transactions from modifying the same business data concurrently, which can achieve the purpose of avoiding dirty writing; but "row lock" will prevent parallel execution of distributed transactions, so that distributed transactions must be serialized row execution, which reduces the concurrency of distributed transactions and affects the performance of distributed transactions

Images