Data consistency method, distributed coordinator and central coordinator

Abstract

The invention provides a data consistency method, a distributed coordinator and a central coordinator, and the data consistency method comprises the steps: recognizing the start of a transaction through a transaction start annotation in a transaction program, creating a global ID when the transaction starts, transmitting the global ID to the central coordinator, and building a global ID queue through the central coordinator; identifying the start of the transaction through the transaction start annotation, when the transaction is started, creating transaction information and sending the information to the central coordinator, wherein the transaction information comprises a transaction ID, a global ID and a transaction calling state, the central coordinator records the transaction ID into acorresponding global ID queue when the transaction calling state is finished, and if the transaction calling state is failed, a fallback instruction is generated according to all completed transaction ID queues in the corresponding global ID queue; and receiving a rollback instruction, and initiating rollback operation. According to the invention, the consistency of distributed server data can berealized.

Description

technical field [0001] The invention relates to the field of distributed servers, in particular to a distributed server data consistency method, a distributed coordinator and a central coordinator. Background technique [0002] The unification of distributed server data in the prior art is mainly realized in the following two ways: 1) two-phase commit (2PC), the principle used is the XA protocol principle (it is a standard protocol formulated by the X / Open organization for distributed transaction processing) ). 2) Compensation transaction (TCC), TCC adopts the compensation mechanism. For each operation, a corresponding confirmation and compensation (revocation) operation must be registered. It is divided into three stages: the Try stage is mainly to test the business system and resource reservation; the Confirm stage is mainly to confirm and submit the business system. When the Try stage is successfully executed and the Confirm stage is started, the Confirm stage will not m...

AI Technical Summary

Problems solved by technology

[0005] In order to solve the defects of information blockage, single point of failure, need to modify business logic and many times of communication in the distributed server data consistency method in the prior art, the first aspect of the present invention provides a distributed server data consistency method The method is applicable to the distributed coordinator, and the transaction start annotation and the transaction start annotation are established in the transaction program in advance, wherein the transaction program includes multiple transactions, and the method includes:

Images