WO2021103036A1 - Transaction commit system and method, and related device - Google Patents

Abstract

The present application provides a transaction commit system and method, and a related device. The system comprises a coordination server and a plurality of storage servers. The method comprises: the coordination server receives a transaction sent by a service server, generates at least one write request according to the transaction, and sends write requests to the storage servers which store data required to be operated by the write requests; the storage servers which receive the write requests execute the received write requests, and respectively send the execution results to the coordination server; the coordination server determines the transaction processing result of the transaction according to the execution results of the write requests fed back by the storage servers, and sends the transaction processing result to the service server. The coordination server can determine the transaction processing result of the transaction according to the execution results of the write requests sent by the storage servers, and returns the transaction processing result of the transaction to the service server, thereby achieving one-phase transaction commit, reduces the time delay of write request processing, and improves the processing efficiency and the working efficiency of the service server.

Description

Transaction submission system, method and related equipment Technical field

This application relates to the field of computer technology, in particular to a transaction submission system, method and related equipment.

Background technique

A transaction is a logical unit in the execution of a database management system, including a limited sequence of database operations. In a distributed system, in order to improve the reliability of data, it is necessary to establish a redundant relationship for the data. For example, multiple copies of the data are stored to establish a redundant relationship, and the multiple copies are distributed in different nodes. When a transaction is executed across multiple participants, each participant can determine the execution result of the transaction executed by each participant, but cannot know the execution result of other participants. In order to maintain the atomicity, consistency, isolation, and durability of the transaction, it is necessary to introduce a coordinator to manage the participants in the execution of the transaction. Two-phase commit is a method designed for transaction commit based on the system architecture of a distributed system. However, in this method, multiple interactions are required between the coordinator and the participants and the master copy in the distributed system, which will cause the transaction to be committed. The time delay is longer.

Summary of the invention

The embodiment of the present invention provides a transaction submission system, method and related equipment, which can reduce the time delay of transaction submission.

In the first aspect, this application provides a transaction processing system, including a coordination server and multiple storage servers, where:

The coordination server is configured to receive transactions, the transaction includes at least one transaction data information, and each transaction data information includes the storage location information of a copy of the original data on the storage server and the write operator; according to the pre-stored data distribution Information to determine transaction data distribution information, wherein the data distribution information indicates a storage server in the transaction processing system that stores each copy of each original data, and the transaction data distribution information indicates that each transaction included in the transaction is stored The storage server for each copy of the original data corresponding to the transaction data information; generate at least one write request, each write request including the transaction data distribution information and one transaction data information; send each write request to the each The storage server where the copy of the original data corresponding to the transaction data information included in each write request is located;

The storage server that receives any write request is configured to store the transaction data distribution information; operate on a copy of the original data corresponding to the transaction data information included in the write request according to the transaction data information included in the write request; Sending the execution result of the write request to the coordination server;

The coordination server is further configured to determine the transaction processing result of the transaction according to the execution result of the write request sent by each storage server that receives the write request, and send the transaction processing result to the service server.

The coordination server generates one or more write requests according to the transaction data information in the received transaction, and then sends each write request directly to the transaction data included in each write request according to the data distribution information of the distributed system stored in advance The storage server where each copy of the original data corresponding to the information is located, so as to avoid the coordination server from first sending the write request to the storage server where the master copy of the original data corresponding to the transaction data information is located, and then the master copy sends the write request to the corresponding The backup copy process improves the efficiency of sending write requests to each storage server. In addition, after the storage server that received the write request has executed the received write request, each storage server feeds back the execution result of the write request to the coordination server, and the coordination server can determine the transaction based on the execution result fed back by each storage server. Then, the transaction processing result of the above transaction is returned to the business server, so as to realize the one-stage transaction submission, reduce the time delay of write request processing, and improve the processing efficiency of write request and the work efficiency of the business server.

In a possible implementation manner, the coordination server is configured to receive a read request, where the read request includes a storage location and a read operator of a copy of the target data on the storage server; and the storage is determined according to the data distribution information. A storage server for the copy of the target data; sending the read request to the storage server for the copy of the target data;

The storage server that has received the read request is configured to query each storage server that has received the write request for the execution result of the write request according to the transaction data distribution information; according to each received write request The execution result of the execution of the write request sent by the storage server determines the execution of the commit operation or the rollback operation; after the execution of the commit operation or the rollback operation, according to the storage location of the copy of the target data on the storage server Read the target data; send the read target data to the coordination server.

In a possible implementation manner, the storage server that receives the read request is further configured to generate operation information according to the execution result of the write request, and the operation information indicates a commit operation or a rollback operation, and the operation The information is sent to each storage server that receives the write request.

The coordination server does not send commit instructions or rollback instructions to the storage server. If the storage server receives a read request, the storage server can send the transaction data to the storage server according to the transaction data identifier and transaction data distribution information in the received write request. The storage server of each copy of the original data corresponding to each transaction data information included queries the execution result of the write request, and determines the commit operation or rollback operation according to the execution result of the queried write request, and then executes the received read request . The storage server is used to determine the transaction corresponding to the write request to perform the commit operation or rollback operation, which reduces the amount of data processed by the business server and improves the efficiency of the business server.

In a possible implementation, the storage location of the copy of the target data on the storage server and the storage location of the copy of the original data on the storage server overlap, or the copy of the target data and the original Copies of data are stored in data blocks that need to be consistent.

In a possible implementation manner, the coordination server is configured to determine multiple storage servers that store copies of the target data according to the data distribution information; from the determined multiple storage servers that store the target data Select a target storage server from the storage servers of the copy; send the read request to the target storage server.

The target storage server may be the closest to the coordination server among the multiple storage servers that store the copies of the target data, or the coordination server has the fastest access speed among the multiple storage servers that store the copies of the target data, so as to improve The speed at which the business server reads the target data improves work efficiency.

In a possible implementation manner, the coordination server is further configured to determine that each write request is received according to the execution result of the write request sent by each storage server that receives the write request The operation information that the storage server needs to perform, the operation information indicates the execution of a commit operation or a rollback operation, and the operation information is sent to each storage server that receives the write request.

In a possible implementation manner, the storage server that receives the write request is further used to determine that no data sent by the coordination server is received within a preset time after sending the execution result of the operation to the coordination server. The operation information, according to the transaction data distribution information, queries the storage server storing each copy of the original data corresponding to each transaction data information included in the transaction for the execution result of the write request; and determines according to the execution result Perform the commit operation or the rollback operation.

After the storage server feeds back the execution result to the coordination server, it does not receive the operation information sent by the coordination server within a preset time period, and the storage server determines that the coordination server is faulty. According to the transaction data distribution information, the storage server queries other storage servers included in the transaction data distribution information for the execution results of other storage servers executing write requests, and then determines to perform the commit operation or rollback operation based on the execution results of other storage servers to prevent the storage server from being When the commit command or rollback command from the coordination server is not received, the storage server cannot release the system resources applied for to execute the above write request, which leads to system congestion and inability to execute new transactions, and can improve the reliability of the system.

In a second aspect, the present application provides a transaction processing method, which is applied to a transaction processing system including a coordination server and multiple storage servers, and the method includes:

The coordination server receives a transaction sent by a business server, the transaction includes at least one transaction data information, and each transaction data information includes storage location information of a copy of the original data on the storage server and a write operator;

The coordination server determines transaction data distribution information according to pre-stored data distribution information, where the data distribution information indicates a storage server that stores each copy of each original data in the transaction processing system, and the transaction data distribution The information indicates a storage server that stores each copy of original data corresponding to each transaction data information included in the transaction;

The coordination server generates at least one write request, and each write request includes the transaction data distribution information and one transaction data information;

The coordination server sends each write request to a storage server where a copy of the original data corresponding to the transaction data information included in each write request is located;

The storage server that receives any write request stores the transaction data distribution information, and according to the transaction data information included in the write request, operates on a copy of the original data corresponding to the transaction data information included in the write request, and reports to the Coordinating the server to send the execution result of the write request;

The coordination server determines the transaction processing result of the transaction according to the execution result of the write request sent by each storage server that receives the write request, and sends the transaction processing result to the service server.

In a possible implementation manner, the coordination server receives a read request sent by the business server, where the read request includes a storage location and a read operator of the copy of the target data on the storage server;

The coordination server determines a storage server that stores a copy of the target data according to the data distribution information, and sends the read request to the storage server that stores the copy of the target data;

According to the transaction data distribution information, the storage server that has received the read request queries each storage server that has received the write request for the execution result of the write request; according to each storage server that has received the write request The sent execution result of the execution of the write request, confirming the execution of the commit operation or the rollback operation;

The storage server that receives the read request reads the target data according to the storage location of the copy of the target data on the storage server, and sends the read target data to the coordination server.

In a possible implementation manner, the storage server that receives the read request generates operation information according to the execution result of the write request, and sends the operation information to each storage server that receives the write request, The operation information indicates a commit operation or a rollback operation.

In a possible implementation, the storage location of the copy of the target data on the storage server and the storage location of the copy of the original data on the storage server overlap, or the copy of the target data and the original Copies of data are stored in data blocks that need to be consistent.

In a possible implementation manner, the coordination server determines a storage server that stores a copy of the target data according to the data distribution information, and sends the read request to the storage server that stores the copy of the target data, include:

The coordination server determines a plurality of storage servers that store copies of the target data according to the data distribution information;

The coordination server selects a target storage server from the determined storage servers that store copies of the target data, and sends the read request to the target storage server.

In a possible implementation manner, the method further includes: the coordination server determines that each write request is received according to the execution result of the write request sent by each storage server that receives the write request Operation information that the storage server needs to perform, where the operation information indicates the execution of a commit operation or a rollback operation;

The coordination server sends the operation information to each storage server that receives the write request.

In a possible implementation manner, the method further includes:

After receiving the write request, the storage server determines that it has not received the operation information sent by the coordination server within a preset time after sending the execution result of the operation to the coordination server, and according to the transaction data distribution information, The storage server storing each copy of the original data corresponding to each transaction data information included in the transaction queries the execution result of the write request;

The storage server receiving the write request determines to execute the commit operation or the rollback operation according to the execution result.

In a third aspect, the present application provides a transaction processing method, which is applied to a coordination server in a transaction processing system including a coordination server and multiple storage servers, and the method includes:

The coordination server receives the transaction sent by the business server, the transaction includes at least one transaction data information, and each transaction data information includes the storage location information of a copy of the original data on the storage server and the write operator;

The coordination server determines transaction data distribution information according to pre-stored data distribution information, where the data distribution information indicates a storage server that stores each copy of each original data in the transaction processing system, and the transaction data distribution information indicates A storage server that stores each copy of original data corresponding to each transaction data information included in the transaction;

The coordination server generates at least one write request, and each write request includes the transaction data distribution information and one transaction data information;

The coordination server sends each write request to a storage server where a copy of the original data corresponding to the transaction data information included in each write request is located;

Receiving, by the coordination server, an execution result of the write request sent by each storage server that receives the write request;

The coordination server determines the transaction processing result of the transaction according to the execution result of the write request sent by each storage server that receives the write request, and sends the transaction processing result to the service server.

In a possible implementation manner, the method further includes: the coordination server determines that each write request is received according to the execution result of the write request sent by each storage server that receives the write request Operation information that the storage server needs to perform, where the operation information indicates the execution of a commit operation or a rollback operation;

The coordination server sends the operation information to each storage server that receives the write request.

In a fourth aspect, the present application provides a transaction processing method, which is applied to a storage server in a transaction processing system including a coordination server and multiple storage servers, and the method includes:

The storage server receives a write request sent by the coordination server. The write request includes transaction data distribution information and one transaction data information. The write request is one of at least one write request generated by the coordination server according to the received transaction. The storage server is one of multiple storage servers where a copy of the original data corresponding to the transaction data information included in the at least one write request is located, the transaction includes at least one transaction data information, and each transaction data information includes a copy of the original data The storage location information and the write operator on the storage server, and the transaction data distribution information indicates the storage server that stores each copy of the original data corresponding to each transaction data information included in the transaction;

The storage server operates on a copy of the original data corresponding to the transaction data information included in the write request according to the transaction data information, and sends the execution result of the write request to the coordination server;

The storage server receives a read request sent by the coordination server, where the read request includes a storage location and a read operator of the copy of the target data on the storage server;

According to the transaction data distribution information, the storage server queries a storage server that stores each copy of original data corresponding to each transaction data information included in the transaction for the execution result of the at least one write request;

The storage server executes a commit operation or a rollback operation according to the execution result of the at least one write request obtained by the query.

In a possible implementation manner, the method further includes: the storage server generates operation information according to the execution result of the at least one write request obtained by the query, the operation information indicating a commit operation or a rollback operation;

The storage server sends the operation information to a storage server that stores each copy of original data corresponding to each transaction data information included in the transaction.

In a possible implementation, the storage location of the copy of the target data on the storage server and the storage location of the copy of the original data on the storage server overlap, or the copy of the target data and the original Copies of data are stored in data blocks that need to be consistent.

In a fifth aspect, the present application provides a transaction processing method, which is applied to a storage server in a transaction processing system including a coordination server and multiple storage servers, and the method includes:

The storage server receives a write request sent by the coordination server. The write request includes transaction data distribution information and one transaction data information. The write request is one of at least one write request generated by the coordination server according to the received transaction. The storage server is one of multiple storage servers where a copy of the original data corresponding to the transaction data information included in the at least one write request is located, the transaction includes at least one transaction data information, and each transaction data information includes a copy of the original data The storage location information and the write operator on the storage server, and the transaction data distribution information indicates the storage server that stores each copy of the original data corresponding to each transaction data information included in the transaction;

The server operates on a copy of the original data corresponding to the transaction data information included in the write request according to the transaction data information, and sends the execution result of the write request to the coordination server;

The storage server determines that the operation information sent by the coordination server is not received within a preset time after sending the execution result of the write request to the coordination server, and according to the transaction data distribution information, the storage of the transaction includes The storage server of each copy of the original data corresponding to each transaction data information inquires the execution result of the at least one write request;

The storage server executes a commit operation or a rollback operation according to the execution result of the at least one write request obtained by the query.

In a sixth aspect, the present application provides a coordination device, and the coordination device includes:

The communication unit is configured to receive a transaction sent by the service server, the transaction includes at least one transaction data information, and each transaction data information includes storage location information of a copy of the original data on the storage server and a write operator;

The processing unit is configured to determine transaction data distribution information according to pre-stored data distribution information, where the data distribution information indicates a storage server that stores each copy of each original data in the transaction processing system, and the transaction data distribution information Indicating a storage server that stores each copy of original data corresponding to each transaction data information included in the transaction;

Generating at least one write request, each write request including the transaction data distribution information and one transaction data information;

The communication unit is further configured to send each write request to a storage server where a copy of the original data corresponding to the transaction data information included in each write request is located;

Receiving an execution result of the write request sent by each storage server that receives the write request;

The processing unit is further configured to determine the transaction processing result of the transaction according to the execution result of the write request sent by each storage server that receives the write request;

The communication unit is further configured to send the transaction processing result to the service server.

In a possible implementation manner, the processing unit is specifically configured to determine each storage server that receives the write request according to the execution result of the write request sent by each storage server that receives the write request The operation information that needs to be executed, the operation information indicating the execution of a commit operation or a rollback operation;

The communication unit is further configured to send the operation information to each storage server that receives the write request.

In a seventh aspect, the present application provides a storage device, the device including:

The communication unit is configured to receive a write request sent by the coordination server, the write request includes transaction data distribution information and one transaction data information, and the write request is one of at least one write request generated by the coordination server according to the received transaction The storage server is one of a plurality of storage servers where a copy of the original data corresponding to the transaction data information included in the at least one write request is located, the transaction includes at least one transaction data information, and each transaction data information includes an original The storage location information and the write operator of the copy of the data on the storage server, and the transaction data distribution information indicates the storage server that stores each copy of the original data corresponding to each transaction data information included in the transaction;

A processing unit, based on the transaction data information, operating on a copy of the original data corresponding to the transaction data information included in the write request;

The communication unit is further configured to send the execution result of the write request to the coordination server; receive a read request sent by the coordination server, where the read request includes the storage location and the reading of the copy of the target data on the storage server. Operator

The processing unit is further configured to query the execution result of the at least one write request from the storage server storing each copy of the original data corresponding to each transaction data information included in the transaction according to the transaction data distribution information;

According to the execution result of the at least one write request obtained by the query, a commit operation or a rollback operation is performed.

In a possible implementation manner, the processing unit is further configured to generate operation information according to the execution result of the at least one write request obtained by the query, where the operation information indicates a commit operation or a rollback operation;

The communication unit is further configured to send the operation information to a storage server that stores each copy of the original data corresponding to each transaction data information included in the transaction.

In a possible implementation, the storage location of the copy of the target data on the storage server and the storage location of the copy of the original data on the storage server overlap, or the copy of the target data and the original Copies of data are stored in data blocks that need to be consistent.

In an eighth aspect, the present application provides a storage device, the device including:

The communication unit is configured to receive a write request sent by the coordination server, the write request includes transaction data distribution information and one transaction data information, and the write request is one of at least one write request generated by the coordination server according to the received transaction The storage server is one of a plurality of storage servers where a copy of the original data corresponding to the transaction data information included in the at least one write request is located, the transaction includes at least one transaction data information, and each transaction data information includes an original The storage location information and the write operator of the copy of the data on the storage server, and the transaction data distribution information indicates the storage server that stores each copy of the original data corresponding to each transaction data information included in the transaction;

A processing unit, configured to operate on a copy of the original data corresponding to the transaction data information included in the write request according to the transaction data information;

The communication unit is further configured to send the execution result of the write request to the coordination server;

The processing unit is further configured to determine that the operation information sent by the coordination server has not been received within a preset time after the execution result of the write request is sent to the coordination server, and send data to the storage system according to the transaction data distribution information. The storage server of each copy of the original data corresponding to each transaction data information included in the transaction queries the execution result of the at least one write request;

According to the execution result of the at least one write request obtained by the query, a commit operation or a rollback operation is performed.

In a ninth aspect, the present application provides a coordination server, including a processor and a memory; the memory is used to store instructions, and the processor is used to execute the instructions. When the processor executes the instructions, the processing The device executes the method provided in the third aspect or any one of the possible implementation manners of the third aspect.

In a tenth aspect, the present application provides a storage server, including a processor and a memory; the memory is used to store instructions, the processor is used to execute the instructions, and when the processor executes the instructions, the processing The device executes the method provided by the fourth aspect or any one of the possible implementation manners of the fourth aspect.

In an eleventh aspect, the present application provides a storage server, including a processor and a memory; the memory is used to store instructions, the processor is used to execute the instructions, and when the processor executes the instructions, the The processor executes the fifth aspect or the method provided in any possible implementation manner of the fifth aspect.

In the twelfth aspect, the present application provides a non-transitory computer-readable storage medium, the computer-readable storage medium stores a computer program, and the computer program implements the third aspect or the third aspect when the computer program is executed by the coordination server. The method provided by any one of the possible implementations.

In the thirteenth aspect, the present application provides a non-transitory computer-readable storage medium, the computer-readable storage medium stores a computer program, and the computer program implements the fourth aspect or the fourth aspect when the computer program is executed by a storage server. The method provided by any one of the possible implementations.

In a fourteenth aspect, the present application provides a non-transitory computer-readable storage medium, the computer-readable storage medium stores a computer program, and the computer program implements the fifth aspect or the fifth aspect when the computer program is executed by a storage server The method provided by any one of the possible implementations.

Description of the drawings

In order to explain the embodiments of the present invention or the technical solutions in the prior art more clearly, the following will briefly introduce the drawings that need to be used in the description of the embodiments or the prior art.

FIG. 1 is a schematic diagram of a system architecture of a distributed system provided by an embodiment of the present application;

FIG. 2 is a schematic diagram of a two-stage submission interaction provided by an embodiment of the present application;

FIG. 3 is a schematic diagram of performing a commit operation or a rollback operation provided by an embodiment of the present application;

FIG. 4 is another schematic diagram of the interaction of two-stage submission provided by an embodiment of the present application;

FIG. 5 is a schematic diagram of interaction of transaction commit provided by an embodiment of the present application;

FIG. 6 is a schematic diagram of interaction for executing a read request according to an embodiment of the present application;

FIG. 7 is a schematic diagram of another transaction submission interaction provided by an embodiment of the present application;

FIG. 8 is a schematic structural diagram of a coordination device provided by an embodiment of the present application;

FIG. 9 is a schematic structural diagram of a storage device provided by an embodiment of the present application;

FIG. 10 is a schematic structural diagram of another storage device provided by an embodiment of the present application;

FIG. 11 is a schematic structural diagram of a coordination server provided by an embodiment of the present application;

FIG. 12 is a schematic structural diagram of a storage server provided by an embodiment of the present application;

FIG. 13 is a schematic structural diagram of another storage server provided by an embodiment of the present application.

Detailed ways

The terms "first" and "second" in the embodiments of the present application are only used for descriptive purposes, and cannot be understood as indicating or implying relative importance or implicitly indicating the number of indicated technical features. Thus, the features defined with "first" and "second" may explicitly or implicitly include one or more of these features.

First, some concepts involved in the embodiments of the present application are introduced.

A transaction is a logical unit in the execution of a database management system, which includes a limited sequence of database operations. In order to ensure the correctness and reliability of data, transactions must have four characteristics: atomicity, consistency, isolation, and durability. Among them, atomicity refers to all operations in a transaction , Either all is completed, or all is not completed, it will not end in a certain link in the middle, the transaction will be rolled back to the state before the start of the transaction if an error occurs during the execution of the transaction. Consistency means that the integrity of the database has not been destroyed before the start of the transaction and after the end of the transaction. Isolation means that the database allows multiple concurrent transactions to read, write and modify data at the same time. Isolation can prevent data inconsistencies caused by cross execution when multiple transactions are executed concurrently. Persistence means that after the transaction is executed, the modification of the data is permanent, and it will not be lost even if the system fails.

Participants assume the role of executor in the transaction execution process, used to receive the transaction request sent by the coordinator, and perform specified operations on the data specified in the transaction request. Among them, a transaction includes a transaction request for one or more data, and the transaction request is a read request or a write request. In a distributed system, data is stored in different storage servers. Since each storage server may fail, in order to improve the reliability of the data, it is necessary to establish a redundant relationship for the data, such as by creating a copy or erasure of the data Code verification improves the fault tolerance of the system. In the embodiments of the present application, the redundancy relationship is taken as an example for description, and other redundancy relationships are also applicable. During the execution of a transaction, multiple copies of the same target data are called a participant, where the target data is one or more data requested in the transaction request, and one of the multiple copies is the master Copies, the rest are backup copies. It can be understood that when server resources are sufficient, and multiple copies of the same data are distributed on different storage servers, one participant corresponds to one or more storage servers. In the embodiment of the present application, multiple copies of the same data are distributed on different storage servers as an example.

The coordinator assumes the coordinating role in the execution of the transaction. In a distributed system, when a transaction is executed across multiple participants (that is, the transaction request included in a transaction is executed by multiple participants), each participant can determine the result of executing the transaction request, but cannot know the other participants The result of executing the transaction request. In order to maintain the above transaction characteristics, a coordinator is therefore introduced to manage the results of executing the transaction request and instruct the storage server whether to submit the result of executing the transaction request, for example, to replace the original data with the updated data.

Exemplarily, the distributed system includes one or more business servers, one or more coordinator servers (coordinators), and multiple storage servers. In Figure 1, a distributed system including a business server, a coordinator server and multiple storage servers is taken as an example. The coordinator is used to receive the transaction request sent by the business server and send the transaction request to one or more participants for execution. Among them, each participant includes multiple copies, each storage server includes multiple logical storage nodes, and multiple copies of the same data are stored in the logical storage nodes of different storage servers. Wherein, each storage server includes multiple logical storage nodes, and each logical storage node may be a chunk in the storage server. For example, if the storage server P1 in FIG. 1 includes logical storage nodes P11, P12, and P13, the storage server P2 includes logical storage nodes P21, P22, and P23, and the storage server P3 includes logical storage nodes P31, P32, and P33. If the logical storage nodes P11, P21, and P31 include three copies of the same data, the logical storage nodes P12, P22, and P32 include three copies of the same data, and the logical storage nodes P13, P23, and P33 include three copies of the same data. The logical storage nodes P11, P21, and P31 are logically one participant; P12, P22, and P32 are logically one participant, and P13, P23, and P33 are logically one participant. One of the multiple copies is the primary copy, and the rest are backup copies. For example, P11 is the primary copy of the participant, and P12 and P13 are the backup copies of the participant.

Two-phase commit is a method designed based on transaction commit under a distributed system architecture. Two-phase commit includes a prepare phase and a commit phase. Specifically, as shown in Figure 2, taking the transaction request included in the transaction sent by the business server as a write request as an example, the prepare phase includes:

a. After receiving the transaction sent by the service server, the coordinator forwards one or more write requests included in the transaction to the participant where the original data requested for the operation in each write request is located. In FIG. 2, the participant 1 receives the first write request, the participant 2 receives the second write request, and the participant n receives the nth write request as an example.

b. After each participant receives the write request, it executes the write request they received. Participants perform write requests on the original data to obtain the target data, but do not replace the original data with the target data, and only store the target data temporarily.

c. The participant sends the operation result to the coordinator. If the participant can successfully obtain the target data based on the original data requested in the write request in step b, then the operation result of the successful operation is sent to the coordinator. If the participant is in step b If the target data cannot be successfully obtained based on the original data of the requested operation in the write request, the operation result of the operation failure is sent to the coordinator. In Figure 2, participant 1 generates the first operation result according to the execution of the first write request, participant 2 generates the second operation result according to the execution of the second write request, and participant n generates the second operation result according to the execution of the nth write request. Take the nth operation result as an example.

The commit phase includes:

d. The coordinator determines the execution result of the write request according to the operation result sent by each participant, and sends a commit instruction or a rollback instruction to each participant. Among them, in the case that the operation result of each participant is successful, the coordinator determines that the execution result of the write request is successful, and the coordinator sends a commit instruction to each participant. In the case where at least one of the operation results of each participant is an operation failure, the coordinator determines that the execution result of the write request is an execution failure, and the coordinator sends a rollback instruction to each participant. In Figure 2, the execution result of the write request is the commit instruction as an example.

e, each participant executes the received commit instruction or rollback instruction. When the coordinator sends a commit command, each participant performs the commit operation and replaces the original data with the above target data; when the coordinator sends a rollback command, each participant performs a rollback operation and saves the above temporarily stored target data delete. As shown in Figure 3, the target data obtained after the original data A is written is B. Before the commit command or rollback command is received, the target data B is temporarily stored. If the participant performs the commit operation according to the commit command, use B replaces A. If the participant performs a rollback operation according to the rollback instruction, the temporarily stored target data B will be deleted.

f. Each participant sends a confirmation message to the coordinator, and the confirmation message indicates the completion of the submission operation or the completion of the rollback operation. In FIG. 2, the participant 1 generates the first confirmation information, the participant 2 generates the second confirmation information, and the participant n generates the nth confirmation information as an example.

The coordinator returns the execution result of the transaction to the service server after receiving the confirmation message fed back by each participant.

The above describes the interaction process between the coordinator and the participant in the two-phase submission. In a distributed system, a participant includes one or more copies of the same data, and the logical storage node where each copy of the same data is located. Distributed on different storage servers. Therefore, in the preparation phase, after a participant receives a write request sent by the coordinator, there is an interactive relationship among multiple copies of the participant. Specifically, the coordination server sends the write request corresponding to each participant to the logical storage node where the master copy of each participant is located, and the storage server where the master copy is located forwards the write request to the storage server storing the backup copy, and then receives it. The multiple logical storage nodes to the write request execute the operation in the write request, and send the operation result in the storage server of each storage backup copy to the coordination server through the storage server storing the primary copy. In the commit phase, the coordination server determines whether the write request is executed successfully according to the received operation results of each participant, and sends a commit command or rollback command to the master copy, and then the master copy forwards the received instructions to the corresponding backup copy. It should be noted that the interaction between the logical storage nodes and the interaction between the logical storage nodes and the coordinator are all performed through the storage server where each logical storage node is located.

Exemplarily, taking three copies of data in a distributed system as an example, two original data need to be modified in the above write request, where the first original data is distributed in logical storage nodes P11, P21, and P31, and the above three The logical storage nodes are located in storage servers P1, P2, and P3, and the second original data is distributed in logical storage nodes P33, P43, and P53. These three logical storage nodes are located in storage servers P3, P4, and P5, namely, P11, P21 and P31 are a participant, P33, P43, and P53 are a participant, and P11 and P43 are stored as master copies. As shown in Figure 4, during the two-phase commit process, the interaction process between logical storage nodes in the distributed system is as follows:

The prepare phase includes:

a. The coordination server first sends the first write request to the logical storage node P11 where the master copy of the first original data is stored, and sends the second write request to the logical storage node P43 where the master copy of the second original data is stored, where The first write request and the second write request are included in the transaction received by the coordination server, the first write request indicates to modify the first original data, and the second write request indicates to modify the second original data.

b. P11 sends the received first write request to P21 and P31 of the storage backup copy, and P43 sends the received second write request to P33 and P53 of the storage backup copy.

c, P11, P21, and P31 perform the corresponding operations in the first write request respectively, P33, P43, and P53 perform the corresponding operations in the second write request respectively, and P21 and P31 send the operation result of the first write request to the storage server P11 , P33 and P53 send the operation result of the second write request to P43, where the operation result indicates the success or failure of the write request operation.

P11 determines the execution result of the first write request according to the operation results of P11, P21, and P31, and P43 determines the execution result of the second write request according to the operation results of P33, P43, and P53, and then P11 and P43 respectively determine the execution result of the first write request. The execution result and the execution result of the second write request are sent to the coordination server. Among them, when the operation results of P11, P21, and P31 are all successful, P11 determines that the execution result of the first write request is successful. When the operation result of any one or more of P11, P21, and P31 is operation failure, P11 determines that the execution result of the first write request is execution failure. The method for P43 to determine whether the execution result of the second write request is successful or failed is similar.

The commit phase includes:

d. The coordination server sends a commit command or a rollback command to P11 and P43 according to the execution result of the first write request and the execution result of the second write request fed back by P11 and P43. In Figure 4, the execution result of the write request is the commit instruction as an example.

e, P11 sends the received commit command or rollback command to P21 and P31, and P43 sends the received commit command or rollback command to P33 and P53. Take the commit command as an example in Figure 4.

f, P11, P21, P31 and P33, P43, P53 execute the received commit command or rollback command. Among them, when a commit instruction is received, each logical storage node performs a commit operation, replacing the original data with the target data obtained after the update; when a rollback instruction is received, each logical storage node performs a rollback operation. Then P21 and P31 feed back the first confirmation information (commit operation completion or rollback operation complete) to P11, and P33 and P53 feed back the second confirmation information (commit operation completion or operation rollback completion) to P43.

During the execution of the above transaction, the coordinator and the participant need to interact twice, and the primary and backup copies in the same participant need to interact twice, which will lead to a longer delay in transaction submission. In addition, during the execution of the transaction, if the coordination server fails, the resources requested by the storage server during the execution of the transaction cannot be released in time, which will cause the system to be in a blocked state, and other blocked transactions will wait until the original Resources related to the transaction request are released.

The embodiment of the present application provides a transaction commit method, which is applied to the distributed system as shown in FIG. 1, in which the business server sends the transaction to the coordination server, and the coordination server is based on the data distribution information of the distributed system and the transactions in the transaction. The data information sends one or more write requests included in the transaction to the corresponding storage server for execution, and each storage server executes the corresponding operation according to the received write request and sends the respective execution result to the coordination server. The coordination server determines the transaction processing result according to the execution result of each storage server, and returns the transaction processing result to the business server. In the following, the transaction submission method provided by the present application will be described in detail from two aspects of the transaction request being a write request and a read request in conjunction with the accompanying drawings. As shown in Figure 4, when the transaction request is a write request, the transaction commit method includes:

S301. The business server sends a transaction to the coordination server.

In the embodiment of the present application, the above-mentioned transaction includes one or more transaction data information and one or more transaction data identifiers, wherein each transaction data information corresponds to a transaction data identifier, and each transaction data information indicates a reference to the distributed database. One or more raw data is operated, and each transaction data information includes storage location information of the original data and an operator, the storage location information includes an address or a data identifier, and the operator instructs to perform a write operation on the original data.

The coordination server stores data distribution information of the distributed system, and the data distribution information records the logical storage node where each data is located in the distributed system and the storage server to which each logical storage node belongs. The coordination server may determine the logical storage node where the original data is located in the distributed database and the storage server to which the logical storage node belongs based on the storage location information and data distribution information of the original data.

S302. The coordination server generates one or more write requests according to the transaction sent by the business server, and sends each write request to a corresponding storage server.

After receiving the transaction, the coordination server queries the data distribution information according to the storage location information of the original data in each transaction data information, and can obtain the query result corresponding to the transaction data information. The query result includes the transaction data information instruction to write The logical storage node where each copy of the original data of the operation is located and the storage server where each logical storage node is located, where each copy of the original data corresponds to a logical storage node. The coordination server can determine the query results corresponding to all transaction data information in the transaction according to each transaction data information and data distribution information, and merge the query results corresponding to all transaction data information included in the transaction to obtain transaction data distribution information, and the transaction data distribution information Indicate the storage server where each copy of the original data corresponding to the storage location information in each transaction data information in the transaction is located.

After determining the transaction data distribution information, the coordination server generates multiple write requests according to the transaction data information, the transaction data identifier, and the transaction data distribution information in the transaction. Wherein, each write request includes a transaction data information, a transaction data identifier, and the foregoing transaction data distribution information. Then each write request is sent to the storage server where each copy of the original data corresponding to the transaction data information included in each write request is located.

Exemplarily, a transaction includes 3 transaction data information, and each transaction data information indicates a write operation to an original data as an example. The first transaction data information indicates an increase by 1 operation to the data of the second field recorded in page1. ; The second transaction data information indicates that the data in the fifth field recorded in page2 should be added by 1; the third transaction data information indicates that the data in the third field recorded in page3 should be subtracted by 1, that is, the transaction instruction is three The original data is written.

The coordination server determines that page1 is located in the logical storage node 11 of the storage server 1, the logical storage node 21 of the storage server 2, and the logical storage node 31 of the storage server 3 according to the above-mentioned data distribution information. The aforementioned page2 is located in the logical storage of the storage server 1. In node 12, logical storage node 22 of storage server 2 and logical storage node 42 of storage server 4, page 3 is located in logical storage node 13 of storage server 1, logical storage node 33 of storage server 3, and logical storage of storage server 4 Node 43. The three transaction data identifiers corresponding to the above three transaction data information are block1, block2, and block3. The coordination server determines the transaction data distribution information corresponding to the transaction according to the logical storage node and the storage server corresponding to each of the determined three transaction data information. The transaction data distribution information corresponding to this transaction is shown in Table 1 below.

Table 1

After the coordination server determines the logical storage node and storage server corresponding to each transaction data information, the coordination server generates three write requests according to the three transaction data information, the three transaction data identifiers, and the transaction data distribution information. A write request includes the above-mentioned first transaction data information, three transaction data identifiers, and transaction data distribution information, that is, the first write request instructs to add 1 to the data in the second field recorded in page1; the second write request includes the above The second transaction data information, the three transaction data identifiers, and the transaction data distribution information, that is, the second write request indicates to add 1 to the data in the fifth field recorded in page2; the third write request includes the aforementioned third transaction data information , Three transaction data identifiers and transaction data distribution information, that is, the third write request instructs to subtract 1 from the data in the third field recorded in page3. Then the coordination server sends the first write request to storage server 1, storage server 2 and storage server 3, sends the second write request to storage server 1, storage server 2 and storage server 4, and sends the third write request to storage server 1. Storage server 3 and storage server 4.

The coordination server generates one or more write requests according to the transaction data information in the received transaction, and then sends each write request directly to the transaction data included in each write request according to the data distribution information of the distributed system stored in advance The storage server where each copy of the original data corresponding to the information is located, so as to avoid the coordination server from first sending the write request to the storage server where the master copy of the original data corresponding to the transaction data information is located, and then the master copy sends the write request to the corresponding The backup copy process improves the efficiency of sending write requests to each storage server.

S303. The storage server executes a corresponding write operation according to the received write request.

After each storage server receives the write request sent by the coordination server, each storage server performs a corresponding write operation on the original data indicated by the transaction data information in the write request according to the received write request. Exemplarily, after the storage server 1, the storage server 2, the storage server 3, and the storage server 4 receive the above-mentioned write request, according to the specific operation indicated by the respective received write request, perform processing on the original data of the specified field in the transaction data information. The corresponding write operation. For example, the storage server 1 receives the first write request, the second write request, and the third write request. The logical storage node 11 in the storage server 1 adds 1 to the data in the second field of page 1, and the logical storage node 12 The data in the fifth field in page2 is added by 1, and the logical storage node 13 is subtracted by 1 in the data in the third field in page3. The storage server 2 receives the first write request and the second write request. The logical storage node 21 of the storage server 2 adds 1 to the data in the second field of page1, and the logical storage node 22 adds 1 to the data in the fifth field of page2. Add 1 to the data. Each storage server temporarily stores the target data obtained after performing a write operation on the original data, but does not replace the original data with the target data.

S304. The storage server sends the execution result of executing the write request to the coordination server.

After receiving the write request, the storage server performs the corresponding write operation on the data in the specified field in the received write request, and then sends the execution result to the coordination server, where each write request corresponds to an execution result, and the execution result instructs the storage server The execution of the write request is successful or failed. Successful execution means that the storage server performs a write operation on the original data specified in the write request to obtain the target data, and execution failure means that the storage server cannot perform a write operation on the original data specified in the write request to obtain the target data. .

If the storage server executes the write operation of the specified data in the received write request successfully, it will send the successful execution result of the write request to the coordination server, if the storage server executes the write operation of the specified data in the received write request If it fails, the execution result of the failed execution of the write request is sent to the coordination server. Exemplarily, if the coordination server sends the first write request, the second write request, and the third write request to the storage server 1, the storage server 1 needs to feed back the three execution results corresponding to the three write requests to the coordination server. For example, the logical storage node 11 in the storage server 1 can successfully execute the write operation to the data in page1, that is, the storage server 1 can successfully execute the first write request, and the storage server 1 will send the execution of the first write request to the coordination server. Successful execution result; if the logical storage node 12 in the storage server 1 cannot successfully execute the write operation to the data in page2, that is, the storage server 1 cannot successfully execute the second write request, the storage server 1 sends the execution second to the coordination server The execution result of the failed execution of the write request.

S305. The coordination server determines the transaction processing result according to the execution result sent by the storage server, and sends the transaction processing result to the business server.

After receiving the execution result corresponding to each write request sent by each storage server, the coordination server determines the transaction processing result according to the execution result sent by each storage server, and sends the transaction processing result to the service server. Specifically, when the coordination server receives that the execution results of the received write requests from each storage server are all executed successfully, the coordination server determines that the transaction processing result of the transaction is successful, and sends the transaction successful to the business server News. When the coordination server receives any one or more execution results that are execution failures, the coordination server determines that the transaction processing result of the transaction is a processing failure, and sends a transaction processing failure message to the business server. For example, the execution results sent by the storage server 1 and the storage server 2 and the storage server 4 are all successful, and the logical storage node 33 in the storage server 3 fails to perform the subtraction operation of the data in page 3, that is, the storage server 3 executes the third write If the request is an execution failure, the coordination server will receive a failed execution result sent by the storage server 3. The coordination server determines that the transaction processing result of the write request is a processing failure, and sends a transaction failure message to the business server.

After the storage server that has received the write request has executed the received write request, each storage server will feed back the execution result of the write request to the coordination server, and the coordination server can determine the transaction according to the execution result fed back by each storage server. After processing the result, the transaction processing result of the above transaction is returned to the service server, so as to realize the one-stage transaction submission, reduce the time delay of the write request processing, and improve the processing efficiency of the write request and the work efficiency of the business server.

The aforementioned coordination server determines the transaction processing result according to the execution result sent by the storage server, and after sending the transaction processing result to the service server, the coordination server does not send a commit instruction or a rollback instruction to each storage server. If the storage server receives a read request after executing a write request, the storage server that receives the read request needs to determine whether the target data that needs to be read by the received read request is related to the original data that the previously executed write request needs to operate, if relevant, The storage server that receives the read request needs to first determine whether the transaction to which the previously executed write request belongs needs to perform a commit operation or a rollback operation, and then execute the received read request after the commit operation or rollback operation is performed. Among them, the target data that the read request needs to read is related to the original data that the write request needs to operate on, which means that the target data that the read request needs to read is the same as the original data that the write request needs to operate, and the target data that the read request needs to read is the same as the write request. The original data that needs to be operated overlaps, the storage location of the target data that needs to be read by the read request overlaps the storage location of the original data that needs to be operated on the write request, or the target data that needs to be read by the read request and the original data that needs to be operated on the write request belong to The same storage area, the storage area can be a data block (block), can be a data area (chunk). The following describes the processing procedure when the storage server receives the read request after executing the write request with reference to FIG. 6.

S601. The coordination server receives the read request sent by the service server, and sends the read request to the target storage server.

The above-mentioned read request includes the storage location information and the operator of the target data to be read. After receiving the read request, the coordination server first queries the data distribution information of the distributed system according to the storage location information of the target data in the read request, and obtains the query result corresponding to the read request. The query result includes the information that the read request needs to read. Multiple logical storage nodes where multiple copies of the target data are located and multiple storage servers where the multiple logical storage nodes are located, where each copy corresponds to a logical storage node, and each logical storage node corresponds to a storage server. Then the coordination server determines the target storage server from the multiple storage servers included in the query result; and sends the above-mentioned read request to the target storage server. Wherein, the target storage server is any one of the multiple storage servers where the above-mentioned target data is located. For example, the target storage server may be the closest to the coordination server among the multiple storage servers where the target data to be read in the read request is located. , Or the fastest coordinated server access among multiple storage servers where the target data to be read in the read request is located.

S602. The target storage server sends a query request to one or more storage servers according to the read request.

After receiving the above-mentioned read request, the target storage server determines that the target data that the read request needs to read is related to the original data that the target write request needs to write based on the read request and one or more previously received write requests, then the target The storage server cannot directly read the above target data. The target storage server must first determine the transaction processing result of the transaction to which the target write request belongs, and determine the commit operation or rollback operation based on the transaction processing result to ensure that the read request needs to read Consistency of multiple copies of target data. The foregoing target write request includes part or all of one or more write requests received by the target storage server before the read request is received. In the embodiment of the present application, the target write request includes one write request as an example.

The target storage server will determine the multiple storage servers included in the transaction data distribution information in the target write request according to the transaction data identification and transaction data distribution information in the target write request, and then send a query request to these multiple storage servers, the query request includes The target write request includes multiple transaction data identifiers, and the query request instructs the storage server to query the execution result of the write request corresponding to the multiple transaction data identifiers in the query request, and send the queried execution result to the target storage server.

S603: The target storage server receives the query result, performs a submit operation or a rollback operation according to the query result, executes a read request, and obtains target data and sends it to the coordination server.

The target storage server determines the transaction processing result of the transaction to which the target write request belongs according to the execution results returned by multiple storage servers. If the transaction processing result is successful, the target storage server sends the transaction data distribution information in the target write request to the target write request. A storage server sends a commit instruction and performs a commit operation on the transaction to which the target write request belongs. If the transaction processing result is a processing failure, the target storage server sends a rollback to multiple storage servers included in the transaction data distribution information in the target write request Instruction, and perform a rollback operation on the transaction to which the target write request belongs. Finally, the received read request is executed, the target data to be read by the read request is obtained and the target data is sent to the coordination server.

For example, the target storage server is the storage server 1. The storage server 1 receives a read request after executing the first write request, the second write request, and the third write request. The target data that the read request needs to read is located in the logic Storage node 11, the storage server 1 first determines the first write request, the second write request, and the third write request included in the transaction to which the first write request belongs according to the transaction data identifier and transaction data distribution information sent by the coordination server. The write request is sent to the storage server 1, the storage server 2, the storage server 3, and the storage server 4. Then the storage server 1 determines the execution results of the first write request, the second write request, and the third write request, and sends a query request to the storage server 2, the storage server 3, and the storage server 4. The query request includes the first write request. The transaction data identifier block1 corresponding to one write request, the transaction data identifier block2 corresponding to the second write request, and the transaction data identifier block3 corresponding to the third write request. The query request instructs the storage server 2, the storage server 3 and the storage server 4 to query the execution result of the write request corresponding to the transaction data identifier in the query request, and send the execution result of each write request to the storage server 1. For example, if the storage server 2 receives the first write request and the second write request, the storage server 2 queries the execution result of the first write request and the execution result of the second write request, and executes the first write request and the second write request. The two execution results of are sent to the storage server 1 respectively.

After the storage server 1 receives the execution results returned by the storage server 2, the storage server 3, and the storage server 4, if the storage server 1, the storage server 2, the storage server 3, and the storage server 4 all successfully execute the write request received respectively, then The storage server 1 performs the commit operation. If any write request in the storage server 1, the storage server 2, the storage server 3, or the storage server 4 fails, the storage server 1 performs a rollback operation.

After the storage server 1 performs the commit operation or rollback operation, it reads the target data specified by the above read request, and sends the read target data to the coordination server. After the coordination server receives the target data sent by the storage server 1, the The target data is sent to the business server.

If the coordination server does not send a commit command or rollback command to the storage server, if the storage server receives a read request, and the target data that the read request needs to read is related to the original data that the storage server needs to operate on the write request previously received The storage server can determine the transaction processing result of the transaction to which the write request belongs based on the transaction data identifier and transaction data distribution information in the write request, and perform the commit operation or rollback operation on the transaction to which the write request belongs based on the transaction processing result, and then Execute the received read request. The storage server determines the transaction processing result of the transaction corresponding to the write request, reduces the amount of data processed by the business server, and improves the efficiency of the business server.

In a possible implementation manner, after the target storage server receives the read request and determines the target write request, since the coordination server has determined the transaction result of the transaction to which the target write request belongs, the target storage server may send the request to the coordination server Send a query request to query the transaction processing result of the transaction to which the target write request belongs. If the target storage server receives the transaction processing result sent by the coordination server as processing success, the target storage server sends a commit instruction to multiple storage servers included in the transaction data distribution information in the target write request, and executes a commit to the transaction to which the target write request belongs operating. If the target storage server receives the transaction processing result sent by the coordination server as a processing failure, the target storage server sends a rollback instruction to multiple storage servers included in the transaction data distribution information in the target write request, and performs a rollback operation on the transaction to which the target write request belongs . Finally, the received read request is executed, the target data to be read by the read request is obtained and the target data is sent to the coordination server.

Referring to Figure 7, in the above S305, after the coordination server sends the transaction processing result to the business server, the coordination server may generate operation information according to the execution result of the write request fed back by the storage server that has received the write request, and the operation information instructs the storage server to perform submission Operation or rollback operation. If the execution result fed back by the storage server that receives the write request is successful, the coordination server generates the operation information as a commit command and sends a commit command to each storage server. After each storage server receives the commit command sent by the coordination server, Each logical storage node in each storage server executes the commit operation. If any one of the execution results fed back by the storage server that receives the write request is execution failure, the coordination server generates the operation information as a rollback instruction, and sends the rollback instruction to each storage server, and each storage server receives the coordination server's sending After the rollback instruction, perform the rollback operation.

In the foregoing S302, the coordination server generates multiple write requests according to the transaction, sends the multiple write requests to some or all of the corresponding multiple storage servers, and sends the commit instruction or rollback instruction to the multiple storage servers. Before all the servers in the server, the coordination server may fail or the network card may lose packets, causing the storage server that received the write request to not receive the commit command or rollback command sent by the coordination server. At this time, the storage server that received the write request needs to determine Whether the transaction to which the received write request belongs needs to perform a commit operation or a rollback operation. The following describes the execution process of the transaction when the coordination server fails.

In the embodiment of the present application, the storage server has the ability to sense and coordinate server failures. For example, the storage server that receives the write request executes the operations in S303 and S304 above, and if the above S304 is executed, the storage server does not receive it within a preset period of time. If the commit instruction or rollback instruction sent by the coordination server is sent, the storage server that has received the write request determines that the coordination server is faulty.

The storage server that receives the write request determines the multiple storage servers included in the transaction data distribution information according to the transaction data distribution information in the received write request, and then sends a query to the storage servers other than itself among the multiple storage servers Request, the query request includes multiple transaction data identifiers included in the write request, the query request instructs the storage server that has received the query request to query the execution result of the write request corresponding to the multiple transaction data identifiers in the query request, and the query result The execution result is sent to the storage server that sent the query request. The storage server receiving the write request generates operation information according to the received execution result.

The storage servers that have received the write request may be all storage servers that have received the write request, and each storage server that has received the write request needs to send a query request to other storage servers, and then generate operation information according to the received execution result. The storage server that receives the write request may also be any storage server, for example, the first storage server that determines that the coordination server has failed. In the case that the storage server that received the write request is a storage server, among the storage servers that received the write request, the first storage server that determines that the coordination server is faulty generates operation information based on the query execution results, and determines it based on the operation information After the commit operation or the rollback operation is performed, the generated operation information is sent to multiple storage servers included in the transaction data distribution information.

Exemplarily, after the storage server 1 determines that the coordination server is faulty, the storage server 1 determines according to the transaction data distribution information that the coordination server has sent the first write request, the second write request, and the third write request to the storage server 1, the storage server 1. Server 2, storage server 3, and storage server 4. Therefore, the storage server 1 sends query requests to the storage server 2, the storage server 3, and the storage server 4, instructing the storage server 2, the storage server 3, and the storage server 4 to query the first write request, the second write request, and the first write request received. Three write the execution result of the request, and send the execution result of the execution write request to the storage server 1.

After the storage server 1 receives the execution results returned by the storage server 2, the storage server 3, and the storage server 4, if the storage server 1, the storage server 2, the storage server 3, and the storage server 4 all successfully execute the write request received respectively, then The storage server 1 determines that the transaction processing result is successful, and the operation information generated by the storage server 1 is a commit instruction, and sends the commit instruction to the storage server 2, the storage server 3, and the storage server 4, and then executes the commit operation. If any one or more of the storage server 1, the storage server 2, the storage server 3, or the storage server 4 fails to execute, the storage server 1 determines that the transaction processing result is a processing failure, and the operation information generated by the storage server 1 is a rollback instruction. And send the rollback instruction to the storage server 2, the storage server 3, and the storage server 4, and then perform the rollback operation.

Optionally, when the coordination server fails, some storage servers may receive the operation information sent by the coordination server. After receiving the operation information, the storage server receives the query request sent by other storage servers according to the transaction in the query request. The data identifier determines the write request of the query request, and then determines the operation information corresponding to the transaction to which the write request belongs, and then sends the operation information to the storage server that sent the query request. For example, the storage server 1 receives the operation information sent by the coordination server as a commit instruction. After the storage server 1 receives a query request sent by another storage server, the storage server 1 determines that the transaction to which the write request of the query request belongs corresponds to commit. Command, and then send a commit command to the storage server that sent the query request.

By sending transaction data distribution information to the storage server, when the storage server senses the failure of the coordination server, it queries other storage servers included in the transaction data distribution information for the execution results of other storage servers executing write requests based on the transaction data distribution information, and then According to the execution results of other storage servers, the commit operation or rollback operation is determined to prevent the storage server from not releasing the system resources requested for executing the above write request when the storage server does not receive the operation information of the coordinating server, which causes the system to be blocked and cannot be executed. Problems with new transactions and can improve the reliability of the system.

In a specific implementation manner, during the execution of the transaction, the storage server may also fail, causing the storage server to fail to execute the write request or fail to return the execution result to the coordination server after executing the write request. When the coordination server does not receive the execution result of the storage server within the preset time period, it is determined that the storage server is faulty. At this time, in order to keep the number of data backups in the system consistent, the distributed system will select a replacement storage server from other storage servers to replace the failed storage server, and restore the failure in the replacement server based on the corresponding copy of the data in the failed storage server Store the data in the server. After the coordination server determines the information of the above-mentioned alternative storage server, the coordination server resends the corresponding write request to the above-mentioned alternative storage server, and then the alternative storage server that receives the write request executes the write operation specified in the write request and sends it to the coordination server The execution result of the execution of the write request.

Continue to take the aforementioned coordination server sending write requests to storage server 1, storage server 2, storage server 3, and storage server 4 as an example. When storage server 2 fails, the distributed system selects storage server m to replace storage server 2, and according to the storage The corresponding copy of the data in the server 2 restores the data in the storage server 2 in the storage server m. After the coordination server determines that the storage server m replaces the storage server 2, it updates the transaction data distribution information, for example, replaces the storage server 2 in the transaction data distribution information with the storage server m, and then sends the write request sent to the storage server 2 to the storage server m, and send the updated transaction data distribution information to the storage server 1, the storage server 3, and the storage server 4. After the storage server m receives the write request, it executes the operation specified in the write request, and then sends the execution result after executing the write request to the coordination server, and the coordination server according to the storage server 1, storage server 3, storage server 4, and storage server m The result of the execution determines the result of the transaction.

The transaction submission system provided by the present application is described in detail above with reference to Figs. 1 to 7, and the related devices and equipment for implementing transaction submission provided by the present application will be described below with reference to Figs. 8-13. Referring to FIG. 8, FIG. 8 is a schematic structural diagram of a coordination device provided by an embodiment of the present application. The coordination device 800 includes a communication unit 810 and a processing unit 820. Among them, the communication unit 810 is used to communicate with a service server and a storage server (storage device).

The communication unit 810 is configured to receive a transaction sent by the service server, the transaction includes at least one transaction data information, and each transaction data information includes storage location information of a copy of the original data on the storage server and a write operator. For the information included in the foregoing transaction, refer to the related description in the foregoing S302, which is not repeated here.

The processing unit 820 is configured to determine transaction data distribution information according to pre-stored data distribution information, where the data distribution information indicates a storage server that stores each copy of each original data in the transaction processing system, and the transaction data distribution information indicates storage of the foregoing A storage server for each copy of the original data corresponding to each transaction data information included in the transaction.

The processing unit 820 is further configured to generate at least one write request according to the received transaction, and each write request includes the foregoing transaction data distribution information and one transaction data information. For the specific operation of the processing unit 820 to generate the write request, reference may be made to the related description in the foregoing S302.

The communication unit 810 is also configured to send each write request to the storage server where the copy of the original data corresponding to the transaction data information included in each write request is located. The execution result sent by each storage server that receives the write request is received, and the execution result is the execution result of the storage server that received the write request operating the copy of the original data corresponding to the transaction data information included in the received write request. The operation of the communication unit 810 sending each write request to the corresponding storage server may be the relevant description in S302 above.

The processing unit 820 is further configured to determine the transaction processing result of the transaction according to the execution result of executing the write request sent by each storage server that receives the write request. For the operation of the processing unit 820 to determine the transaction processing result, reference may be made to the relevant description in the foregoing S305, which will not be repeated here.

The communication unit 810 is also configured to send the transaction processing result to the service server.

The coordination device 800 may further include a storage unit 830, and the storage unit 830 is configured to store the foregoing data distribution information.

Specifically, the specific operation of the coordination device 800 to implement distributed fact submission can refer to the operations performed by the coordination server in the foregoing method embodiment. For example, the processing unit 820 is configured to perform the operations performed by the coordination server in S301 to S305 or S601 to S603. The operation is not repeated here.

Refer to FIG. 9, which is a schematic structural diagram of a storage device provided by the present application. The storage device 900 includes a communication unit 910 and a processing unit 920. Among them, the communication unit 910 is used to communicate with other storage servers (storage devices) and coordination servers (coordination devices).

The communication unit 910 is configured to receive a write request sent by the coordination server. The write request includes transaction data distribution information and one transaction data information. The write request is any one or more of at least one write request generated by the coordination server according to the received transaction. A write request, the target storage server is any one of the multiple storage servers where the copy of the original data corresponding to the transaction data information included in the at least one write request is located, the transaction includes at least one transaction data information, and the transaction data distribution information indicates the storage transaction The storage server for each copy of the original data corresponding to each transaction data information included. For the information included in the above transaction, transaction data information, etc., please refer to the description in the above S301, which will not be repeated here.

The processing unit 920 operates on a copy of the original data corresponding to the transaction data information included in the write request according to the transaction data information, and the communication unit 910 sends the execution result of the above write request to the coordination server. For operations performed by the processing unit 920 after receiving the write request, reference may be made to operations performed by the storage server in S303 and S304, which will not be repeated here.

The communication unit 910 is also configured to receive a read request sent by the coordination server, where the read request includes the storage location of the copy of the target data on the storage server and the read operator. For the specific information included in the read request, refer to the description in S601 above.

The processing unit 920, according to the transaction data distribution information in the received write request, queries the storage server storing each copy of the original data corresponding to each transaction data information included in the transaction for the execution result of the at least one write request. Execute the commit operation or rollback operation according to the execution result of the at least one write request obtained by the query; read the target data according to the storage location of the copy of the target data on the storage device, and send the read target data to The coordination server. For the specific operation of the processing unit 920 to query the execution result of the at least one write request, refer to the operation of the target storage server sending a query request to one or more storage servers in S602; the processing unit 920 feedbacks according to each storage server that receives the write request. For the method for determining the execution of the commit operation or the rollback operation according to the execution result, refer to the description of the target storage server performing the commit operation or the rollback operation according to the query result in S603, which will not be repeated here.

The processing unit 920 is further configured to generate operation information according to the execution result of the at least one write request obtained by the query, the operation information indicating a commit operation or a rollback operation; the communication unit 910 is also configured to transfer the operation The information is sent to a storage server that stores each copy of the original data corresponding to each transaction data information included in the transaction. The method for the processing unit 920 to generate the operation information may refer to the method for the target storage server to generate the operation information in S603.

The storage device 900 further includes a storage unit 930, which is used to store the original data that needs to be manipulated by the write request and the target data that needs to be read by the read request.

Specifically, for the specific operations of the storage device 900 to implement distributed fact submission, refer to the operations performed by the storage server in the foregoing method embodiment. For example, the processing unit 920 is configured to perform the operations performed by the storage server in S301 to S305 or S601 to S603. The operation is not repeated here.

Refer to FIG. 10, which is a schematic structural diagram of a storage device provided by the present application. The storage device 1000 includes a communication unit 1010 and a processing unit 1020. Among them, the communication unit 910 is used to communicate with other storage servers (storage devices) and coordination servers (coordination devices).

The communication unit 1010 is configured to receive a write request sent by the coordination server. The write request includes transaction data distribution information and one transaction data information. The write request is one of at least one write request generated by the coordination server according to the received transaction, and the storage server Is one of the multiple storage servers where a copy of the original data corresponding to the transaction data information included in the at least one write request is located, the transaction includes at least one transaction data information, and each transaction data information includes a copy of the original data on the storage server The storage location information and the write operator, and the transaction data distribution information indicate a storage server that stores each copy of the original data corresponding to each transaction data information included in the transaction. For the information included in the above transaction, transaction data information, etc., please refer to the description in the above S301, which will not be repeated here.

The processing unit 1020 is configured to operate on a copy of the original data corresponding to the transaction data information included in the write request according to the transaction data information; the communication unit 1010 is also configured to send the execution result of the write request to the coordination server. For operations performed by the processing unit 1020 after receiving the write request, reference may be made to operations performed by the storage server in S303 and S304, which will not be repeated here.

The processing unit 1020 is further configured to determine that the operation information sent by the coordination server is not received within a preset time after sending the execution result of the write request to the coordination server, and send the operation information to the storage according to the transaction data distribution information. The storage server of each copy of the original data corresponding to each transaction data information included in the transaction queries the execution result of the at least one write request; according to the execution result of the at least one write request obtained by the query, the commit operation or the response is executed Roll operation. The processing unit 1020 executes, according to transaction data distribution information, querying the storage server storing each copy of the original data corresponding to each transaction data information included in the transaction for the execution result of the at least one write request, and according to the at least one write request. For the operations performed as a result of the execution result of, refer to the operations performed after the storage server perceives the failure of the coordination server in the foregoing method embodiment, which will not be repeated here.

The storage device 1000 further includes a storage unit 1030, which is configured to store the original data that needs to be operated on by the write request and the target data that needs to be read by the read request. The processing unit 1020 executes the operations performed by the storage server in S301 to S305 and the operations performed by the storage server after sensing the failure of the coordination server, which will not be repeated here.

Refer to FIG. 11, which is a schematic structural diagram of a coordination server provided by an embodiment of the present application. The coordination server 1100 includes: one or more processors 1110, a communication interface 1120, and a memory 1130. Optionally, the processor 1110, the communication interface 1120, and the memory 1130 are connected to each other through a bus 1140, where:

The processor 1110 is configured to implement the operations performed by the processing unit 820, and the specific implementation of the various operations performed by the processor 1110 may refer to the specific operations performed by the coordination server in the foregoing method embodiment. For example, the processor 1110 is configured to perform operations of the coordination server in S303 and S304 in FIG. 5, and details are not described herein again.

The processor 1110 may have a variety of specific implementation forms. For example, the processor 1110 may be a central processing unit (CPU) or an graphics processing unit (GPU), and the processor 1110 may also be a single-core processor or Multi-core processor. The processor 1110 may be a combination of a CPU and a hardware chip. The above-mentioned hardware chip may be an application-specific integrated circuit (ASIC), a programmable logic device (PLD) or a combination thereof. The aforementioned PLD may be a complex programmable logic device (CPLD), a field-programmable gate array (FPGA), a generic array logic (GAL), or any combination thereof. The processor 1110 may also be implemented solely by a logic device with built-in processing logic, such as an FPGA or a digital signal processor (digital signal processor, DSP).

The communication interface 1120 can be a wired interface or a wireless interface for communicating with other modules or devices. The wired interface can be an Ethernet interface, a local interconnect network (LIN), etc., and the wireless interface can be a cellular network interface or use Wireless LAN interface, etc.

In the embodiment of the present application, the communication interface 1120 executes the operations implemented by the communication unit 810, for example, it can be used to execute the transaction sent by the service server in S301, or the execution result of the storage server. Specifically, the actions performed by the communication interface 1120 may refer to the actions received or sent by the coordination server in the foregoing method embodiments, and details are not described herein again.

The memory 1130 may be a non-volatile memory, for example, read-only memory (ROM), programmable read-only memory (programmable ROM, PROM), erasable programmable read-only memory (erasable PROM, EPROM), Electrically erasable programmable read-only memory (EPROM, EEPROM) or flash memory. The memory 1130 may also be a volatile memory, and the volatile memory may be a random access memory (RAM), which is used as an external cache.

The memory 1130 may also be used to store instructions and data, so that the processor 1110 can call the instructions stored in the memory 1130 to implement the operations performed by the processing unit 820, such as coordinating operations performed by the server in the foregoing method embodiment. In addition, the coordination server 1100 may include more or fewer components than those shown in FIG. 11, or may have different component configurations.

The bus 1140 may be a peripheral component interconnect standard (PCI) bus or an extended industry standard architecture (EISA) bus, etc. The bus 1140 can be divided into an address bus, a data bus, a control bus, and so on. For ease of representation, only one thick line is used to represent in FIG. 11, but it does not mean that there is only one bus or one type of bus.

Optionally, the coordination server 1100 may further include an input/output interface 1150, and the input/output interface 1150 is connected to an input/output device for receiving input information and outputting operation results.

Specifically, for the specific implementation of various operations performed by the aforementioned coordination server 1100, reference may be made to the specific operations performed by the coordination server in the foregoing method embodiments, and details are not described herein again.

Referring to FIG. 12, FIG. 12 is a schematic structural diagram of a storage server according to an embodiment of the present application. The storage server 1200 includes: one or more processors 1210, a communication interface 1220, and a memory 1230. Optionally, the processor 1210, the communication interface 1220, and the memory 1230 are connected to each other through a bus 1240, where:

The processor 1210 is configured to implement the operations performed by the processing unit 920. For the specific implementation of the various operations performed by the processor 1210, refer to the specific operations performed by the storage server in the foregoing method embodiment. For example, the processor 1210 is used to perform the operations performed by the storage server in S303 and S304 in FIG. 5, or the operations performed by the storage servers in S301 to S305, and the operations performed by the storage server after sensing that the coordination server has failed. Go into details again.

The processor 1210 may have a variety of specific implementation forms. For example, the processor 1210 may be a central processing unit or an image processor, and the processor 1210 may also be a single-core processor or a multi-core processor. The processor 1210 may be a combination of a CPU and a hardware chip.

The communication interface 1220 may be a wired interface or a wireless interface. For details, refer to the communication interface 1120.

In the embodiment of the application, the communication interface 1220 executes the operations implemented by the communication unit 910, which can be specifically used to perform operations such as receiving transaction data identifiers, transaction data information, and transaction data distribution information in S303, or perform operations of storing server execution results in S304. Send to the coordination server and other operations. Specifically, the actions performed by the communication interface 1220 may refer to the actions received or sent by the storage server in the foregoing method embodiment, and details are not described herein again.

The memory 1230 may be a non-volatile memory or a volatile memory. For details, refer to the memory 1130.

The memory 1230 may also be used to store instructions and data, so that the processor 1210 can call the instructions stored in the memory 1230 to implement the operations performed by the processing unit 920, such as the operations performed by the storage server in the foregoing method embodiment. In addition, the storage server 1200 may include more or fewer components than those shown in FIG. 12, or may have different component configurations.

The bus 1240 may be a PCI bus, an EISA bus, or the like. For details, refer to the bus 1140.

Optionally, the storage server 1200 may further include an input/output interface 1250, and the input/output interface 1250 is connected to an input/output device for receiving input information and outputting operation results.

Specifically, for the specific implementation of the various operations performed by the foregoing server 1200, reference may be made to the specific operations performed by the storage server in the foregoing method embodiment, and details are not described herein again.

Referring to FIG. 13, FIG. 13 is a schematic structural diagram of a storage server according to an embodiment of the present application. The storage server 1300 includes: one or more processors 1310, a communication interface 1320, and a memory 1330. Optionally, the processor 1310, the communication interface 1320, and the memory 1330 are connected to each other through a bus 1340, where:

The processor 1310 is configured to implement the operations performed by the processing unit 920, and the specific implementation of various operations performed by the processor 1310 may refer to the specific operations performed by the storage server in the foregoing method embodiment. For example, the processor 1310 is configured to perform the operations performed by the storage server in S301 to S305 in FIG. 5 and the operations performed by the storage server in the foregoing method-side embodiment after it perceives a failure of the coordination server, which will not be repeated here.

The processor 1310 can have a variety of specific implementation forms. For example, the processor 1310 can be a central processing unit or an image processor, the processor 1310 can also be a single-core processor or a multi-core processor, and the processor 1310 can also be composed of a CPU and a hardware chip. The combination.

The communication interface 1320 may be a wired interface or a wireless interface. For details, refer to the communication interface 1120.

In this embodiment of the application, the communication interface 1320 executes the operations implemented by the communication unit 910, which can be specifically used to perform operations such as receiving transaction data identifiers, transaction data information, and transaction data distribution information in S303, or execute S304 to store the execution results of the server Send to the coordination server and other operations. Specifically, the actions performed by the communication interface 1320 may refer to the actions received or sent by the storage server in the foregoing method embodiment, and details are not described herein again.

The memory 1330 may be a non-volatile memory or a volatile memory. For details, refer to the memory 1130.

The memory 1330 may also be used to store instructions and data, so that the processor 1310 can call the instructions stored in the memory 1330 to implement the operations performed by the processing unit 820, for example, the operations performed by the storage server in the foregoing method embodiment. In addition, the storage server 1300 may include more or fewer components than those shown in FIG. 13, or may have different component configurations.

The bus 1340 may be a PCI bus, an EISA bus, or the like. For details, refer to the bus 1140.

Optionally, the storage server 1300 may further include an input/output interface 1350. The input/output interface 1350 is connected to an input/output device for receiving input information and outputting operation results.

The embodiment of the present application also provides a non-transitory computer-readable storage medium in which a computer program is stored. When the computer program runs on a processor, the coordination server in the above-mentioned method embodiment can be implemented. For the method steps to be executed, the processor of the computer storage medium can refer to the specific operations of the coordination server in the method embodiments for the specific implementation of the above method steps, which will not be repeated here.

The embodiments of the present application also provide a non-transitory computer-readable storage medium in which a computer program is stored. When the computer program runs on a processor, the storage server in the above method embodiment can be executed. For the method steps of the computer storage medium, the processor of the computer storage medium can refer to the specific operation of the storage server in the above method embodiment for the specific implementation of the above method steps, which will not be repeated here.

In the above-mentioned embodiments, the description of each embodiment has its own focus. For a part that is not described in detail in an embodiment, reference may be made to related descriptions of other embodiments.

The above-mentioned embodiments may be implemented in whole or in part by software, hardware, firmware or any other combination. When implemented using software, the above-mentioned embodiments may be implemented in the form of a computer program product in whole or in part. The computer program product includes one or more computer instructions. When the computer program instructions are loaded or executed on a computer, the processes or functions described in the embodiments of the present invention are generated in whole or in part. The computer may be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices. The computer instructions may be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions may be transmitted from a website, computer, server, or data center. Transmission to another website, computer, server or data center via wired (such as coaxial cable, optical fiber, digital subscriber line (DSL)) or wireless (such as infrared, wireless, microwave, etc.). The computer-readable storage medium may be any available medium that can be accessed by a computer or a data storage device such as a server or a data center that includes one or more sets of available media. The usable medium may be a magnetic medium (for example, a floppy disk, a hard disk, a magnetic tape), an optical medium (for example, a DVD), or a semiconductor medium. The semiconductor medium may be a solid state drive (SSD).

The steps in the method in the embodiment of the present application can be adjusted, merged, or deleted in order according to actual needs; the modules in the device in the embodiment of the present application can be divided, merged, or deleted according to actual needs.

The embodiments of the application are described in detail above, and specific examples are used in this article to illustrate the principles and implementation of the application. The descriptions of the above embodiments are only used to help understand the methods and core ideas of the application; at the same time, for Those of ordinary skill in the art, based on the idea of the application, will have changes in the specific implementation and the scope of application. In summary, the content of this specification should not be construed as a limitation to the application.

Claims (28)

A transaction processing system, characterized in that it includes a coordination server and a plurality of storage servers; The coordination server is configured to receive transactions, the transaction includes at least one transaction data information, and each transaction data information includes the storage location information of a copy of the original data on the storage server and the write operator; according to the pre-stored data distribution Information to determine transaction data distribution information, wherein the data distribution information indicates a storage server in the transaction processing system that stores each copy of each original data, and the transaction data distribution information indicates that each transaction included in the transaction is stored The storage server for each copy of the original data corresponding to the transaction data information; generate at least one write request, each write request including the transaction data distribution information and one transaction data information; send each write request to the each The storage server where the copy of the original data corresponding to the transaction data information included in each write request is located; The storage server that receives any write request is configured to store the transaction data distribution information; operate on a copy of the original data corresponding to the transaction data information included in the write request according to the transaction data information included in the write request; Sending the execution result of the write request to the coordination server; The coordination server is further configured to determine the transaction processing result of the transaction according to the execution result of the write request sent by each storage server that receives the write request, and send the transaction processing result to the service server. The transaction processing system of claim 1, wherein: The coordination server is configured to receive a read request, where the read request includes a storage location and a read operator of the copy of the target data on the storage server; according to the data distribution information, the storage server that stores the copy of the target data is determined ; Send the read request to a storage server that stores a copy of the target data; The storage server that has received the read request is configured to query each storage server that has received the write request for the execution result of the write request according to the transaction data distribution information; according to each received write request The execution result of the execution of the write request sent by the storage server determines the execution of the commit operation or the rollback operation; after the execution of the commit operation or the rollback operation, according to the storage location of the copy of the target data on the storage server Read the target data; send the read target data to the coordination server. The transaction processing system of claim 2, wherein: The storage server that receives the read request is further configured to generate operation information according to the execution result of the write request, the operation information indicating a submit operation or a rollback operation, and send the operation information to each receiving The storage server for the write request. The transaction processing system according to claim 2 or 3, wherein the storage location of the copy of the target data on the storage server overlaps with the storage location of the copy of the original data on the storage server, or the The copy of the target data and the copy of the original data are stored in a data block that needs to be consistent. The transaction processing system according to any one of claims 2 to 4, wherein: The coordination server is configured to determine a plurality of storage servers that store copies of the target data according to the data distribution information; select the target storage server from the determined multiple storage servers that store copies of the target data ; Send the read request to the target storage server. The transaction processing system of claim 1, wherein: The coordination server is further configured to determine the operation information that each storage server that receives the write request needs to perform based on the execution result of the write request sent by each storage server that receives the write request, The operation information indicates the execution of a commit operation or a rollback operation, and the operation information is sent to each storage server that receives the write request. The transaction processing system of claim 6, wherein: The storage server that receives the write request is further configured to determine that the operation information sent by the coordination server is not received within a preset time after sending the execution result of the operation to the coordination server, and according to the transaction The data distribution information queries the storage server storing each copy of the original data corresponding to each transaction data information included in the transaction for the execution result of the write request; according to the execution result, it is determined to execute the commit operation or the return Roll operation. A transaction processing method, characterized in that it is applied to a transaction processing system including a coordination server and a plurality of storage servers, and the method includes: The coordination server receives a transaction sent by a business server, the transaction includes at least one transaction data information, and each transaction data information includes storage location information of a copy of the original data on the storage server and a write operator; The coordination server determines transaction data distribution information according to pre-stored data distribution information, where the data distribution information indicates a storage server that stores each copy of each original data in the transaction processing system, and the transaction data distribution The information indicates a storage server that stores each copy of original data corresponding to each transaction data information included in the transaction; The coordination server generates at least one write request, and each write request includes the transaction data distribution information and one transaction data information; The coordination server sends each write request to a storage server where a copy of the original data corresponding to the transaction data information included in each write request is located; The storage server that receives any write request stores the transaction data distribution information, and according to the transaction data information included in the write request, operates on a copy of the original data corresponding to the transaction data information included in the write request, and reports to the Coordinating the server to send the execution result of the write request; The coordination server determines the transaction processing result of the transaction according to the execution result of the write request sent by each storage server that receives the write request, and sends the transaction processing result to the service server. The method according to claim 8, wherein the method further comprises: Receiving, by the coordination server, a read request sent by the business server, the read request including a storage location and a read operator of the copy of the target data on the storage server; The coordination server determines a storage server that stores a copy of the target data according to the data distribution information, and sends the read request to the storage server that stores the copy of the target data; According to the transaction data distribution information, the storage server that has received the read request queries each storage server that has received the write request for the execution result of the write request; according to each storage server that has received the write request The sent execution result of the execution of the write request, confirming the execution of the commit operation or the rollback operation; The storage server that receives the read request reads the target data according to the storage location of the copy of the target data on the storage server, and sends the read target data to the coordination server. The method according to claim 9, wherein the method further comprises: The storage server that receives the read request generates operation information according to the execution result of the write request, and sends the operation information to each storage server that receives the write request, and the operation information indicates submitting an operation or returning Roll operation. The method according to claim 9 or 10, wherein the storage location of the copy of the target data on the storage server overlaps with the storage location of the copy of the original data on the storage server, or the target data The copy of and the copy of the original data are stored in a data block that needs to be consistent. The method according to any one of claims 9 to 11, wherein the coordination server determines a storage server that stores a copy of the target data according to the data distribution information, and sends the read request to the storage The storage server for the copy of the target data includes: The coordination server determines a plurality of storage servers that store copies of the target data according to the data distribution information; The coordination server selects a target storage server from the determined storage servers that store copies of the target data, and sends the read request to the target storage server. The method according to claim 8, wherein the method further comprises: The coordination server determines the operation information that each storage server that receives the write request needs to perform according to the execution result of the write request sent by each storage server that receives the write request, and the operation information indicates the execution Submit operation or rollback operation; The coordination server sends the operation information to each storage server that receives the write request. The method of claim 13, wherein the method further comprises: After receiving the write request, the storage server determines that it has not received the operation information sent by the coordination server within a preset time after sending the execution result of the operation to the coordination server, and according to the transaction data distribution information, The storage server storing each copy of the original data corresponding to each transaction data information included in the transaction queries the execution result of the write request; The storage server receiving the write request determines to execute the commit operation or the rollback operation according to the execution result. A transaction processing method, characterized in that the method includes: The coordination server receives the transaction sent by the business server, the transaction includes at least one transaction data information, and each transaction data information includes the storage location information of a copy of the original data on the storage server and the write operator; The coordination server determines transaction data distribution information according to pre-stored data distribution information, where the data distribution information indicates a storage server that stores each copy of each original data in the transaction processing system, and the transaction data distribution information indicates A storage server that stores each copy of original data corresponding to each transaction data information included in the transaction; The coordination server generates at least one write request, and each write request includes the transaction data distribution information and one transaction data information; The coordination server sends each write request to a storage server where a copy of the original data corresponding to the transaction data information included in each write request is located; Receiving, by the coordination server, an execution result of the write request sent by each storage server that receives the write request; The coordination server determines the transaction processing result of the transaction according to the execution result of the write request sent by each storage server that receives the write request, and sends the transaction processing result to the service server. The method of claim 15, wherein the method further comprises: The coordination server determines the operation information that each storage server that receives the write request needs to perform according to the execution result of the write request sent by each storage server that receives the write request, and the operation information indicates the execution Submit operation or rollback operation; The coordination server sends the operation information to each storage server that receives the write request. A transaction processing method, characterized in that the method includes: The storage server receives a write request sent by the coordination server. The write request includes transaction data distribution information and one transaction data information. The write request is one of at least one write request generated by the coordination server according to the received transaction. The storage server is one of multiple storage servers where a copy of the original data corresponding to the transaction data information included in the at least one write request is located, the transaction includes at least one transaction data information, and each transaction data information includes a copy of the original data The storage location information and the write operator on the storage server, and the transaction data distribution information indicates the storage server that stores each copy of the original data corresponding to each transaction data information included in the transaction; The storage server operates on a copy of the original data corresponding to the transaction data information included in the write request according to the transaction data information, and sends the execution result of the write request to the coordination server; The storage server receives a read request sent by the coordination server, where the read request includes a storage location and a read operator of the copy of the target data on the storage server; According to the transaction data distribution information, the storage server queries a storage server that stores each copy of original data corresponding to each transaction data information included in the transaction for the execution result of the at least one write request; The storage server executes a commit operation or a rollback operation according to the execution result of the at least one write request obtained by the query. The method of claim 17, wherein the method further comprises: The storage server generates operation information according to the execution result of the at least one write request obtained by the query, and the operation information indicates a commit operation or a rollback operation; The storage server sends the operation information to a storage server that stores each copy of original data corresponding to each transaction data information included in the transaction. The method of claim 17 or 18, wherein: The storage location of the copy of the target data on the storage server and the storage location of the copy of the original data on the storage server overlap, or the copy of the target data and the copy of the original data are stored in a storage that needs to be consistent In the data block. A coordination device, characterized in that the coordination device includes: The communication unit is configured to receive a transaction sent by the service server, the transaction includes at least one transaction data information, and each transaction data information includes storage location information of a copy of the original data on the storage server and a write operator; The processing unit is configured to determine transaction data distribution information according to pre-stored data distribution information, where the data distribution information indicates a storage server that stores each copy of each original data in the transaction processing system, and the transaction data distribution information Indicating a storage server that stores each copy of original data corresponding to each transaction data information included in the transaction; Generating at least one write request, each write request including the transaction data distribution information and one transaction data information; The communication unit is further configured to send each write request to a storage server where a copy of the original data corresponding to the transaction data information included in each write request is located; Receiving an execution result of the write request sent by each storage server that receives the write request; The processing unit is further configured to determine the transaction processing result of the transaction according to the execution result of the write request sent by each storage server that receives the write request; The communication unit is further configured to send the transaction processing result to the service server. The device of claim 20, wherein: The processing unit is specifically configured to determine the operation information that each storage server that receives the write request needs to perform according to the execution result of the write request sent by each storage server that receives the write request, the operation The information indicates the execution of a commit operation or a rollback operation; The communication unit is further configured to send the operation information to each storage server that receives the write request. A storage device, characterized in that the device comprises: The communication unit is configured to receive a write request sent by the coordination server, the write request includes transaction data distribution information and one transaction data information, and the write request is one of at least one write request generated by the coordination server according to the received transaction The storage server is one of a plurality of storage servers where a copy of the original data corresponding to the transaction data information included in the at least one write request is located, the transaction includes at least one transaction data information, and each transaction data information includes an original The storage location information and the write operator of the copy of the data on the storage server, and the transaction data distribution information indicates the storage server that stores each copy of the original data corresponding to each transaction data information included in the transaction; A processing unit, based on the transaction data information, operating on a copy of the original data corresponding to the transaction data information included in the write request; The communication unit is further configured to send the execution result of the write request to the coordination server; receive a read request sent by the coordination server, where the read request includes the storage location and the reading of the copy of the target data on the storage server. Operator The processing unit is further configured to query the execution result of the at least one write request from the storage server storing each copy of the original data corresponding to each transaction data information included in the transaction according to the transaction data distribution information; According to the execution result of the at least one write request obtained by the query, a commit operation or a rollback operation is performed. The device of claim 22, wherein: The processing unit is further configured to generate operation information according to the execution result of the at least one write request obtained by the query, where the operation information indicates a commit operation or a rollback operation; The communication unit is further configured to send the operation information to a storage server that stores each copy of the original data corresponding to each transaction data information included in the transaction. The device according to claim 22 or 23, wherein the storage location of the copy of the target data on the storage server overlaps with the storage location of the copy of the original data on the storage server, or the target data The copy of and the copy of the original data are stored in a data block that needs to be consistent. A coordination server, characterized by comprising a processor and a memory; the memory is used to store instructions, the processor is used to execute the instructions, and when the processor executes the instructions, the processor executes The method of any one of claims 15 or 16. A storage server, characterized by comprising a processor and a memory; the memory is used to store instructions, the processor is used to execute the instructions, and when the processor executes the instructions, the processor executes The method of any one of claims 17-19. A non-transitory computer-readable storage medium, the computer-readable storage medium stores a computer program, and when the computer program is executed by a coordination server, the method according to any one of claims 15 or 16 is implemented. A non-transitory computer-readable storage medium, the computer-readable storage medium stores a computer program, and when the computer program is executed by a storage server, the method according to any one of claims 17 to 19 is implemented.

Images