[microservice Architecture] distributed transactions

Phase II message

Local message table

In the scheme based on local message table , Operate this transaction outside , Record in the message table

Other business , Provide operation interface

The scheduled task polls the local message table , Send the unexecuted message to the operation interface .

Operation interface processing succeeded , Returns the success identifier , Processing failed , Return failure ID .

The scheduled task receives the identification , Update the status of the message

Regular tasks are executed repeatedly according to a certain period

For repeated failures , You can set the maximum number of failures

Messages that exceed the maximum number of failures , No interface calls

Waiting for manual processing

For example, the payment scenario using Alipay , The system generates orders , After Alipay system pays successfully , Call the callback interface provided by our system , The callback interface updates the order status to paid . Callback notification execution failed , Alipay will call again after a period of time .

Disadvantages of the local message table scheme ：

Message data and business data coupling , The message table needs to be formulated according to specific business scenarios , Not for public use . Even if you can share the message table , For the sub database business, each database needs a message table .

Only applicable to the final consistent business scenario .

RocketMQ Transaction message

Message queue RocketMQ Version provides similar XA or Open XA Distributed transaction function of , Through message queuing RocketMQ Version transaction message , It can achieve the final consistency of distributed transactions .

Rocketmq Disadvantages of message transaction

In extreme cases , Automatic processing of backchecks in transaction messages , It's solved RocketMQ Check the data inconsistency in extreme cases （ Has applied for a patent ）

Multiple distributed transaction protocols support

SAGA:

SAGA It first appeared in 1987 year Hector Garcaa-Molrna & Kenneth Salem Published papers SAGAS in . Its core idea is to split long transactions into multiple short transactions , from Saga Transaction coordinator coordination , If each short transaction is successfully committed , Then the global transaction will complete normally , If a step fails , The compensation operation is called once in reverse order .

SAGA The transaction mode is DTM The most common pattern in , Mainly because SAGA The model is simple and easy to use , Less work , And it can solve the needs of most businesses .

For example, we want to carry out a business similar to bank inter-bank transfer , take A Medium 30 Yuan to B, according to Saga How things work , We will deal with the overall situation , It is divided into the following services ：

• Transfer out （TransOut） service , Transfer out here will be operated A-30
• Transfer out compensation （TransOutCompensate） service , Rollback the above transfer out operation , namely A+30
• into （TransIn） service , The transfer will take place B+30
• Transfer in compensation （TransInCompensate） service , Rollback the above transfer in operation , namely B-30

Whole SAGA The logic of the transaction is ：

• Transfer out succeeded => Execution transfer succeeded => Global transaction completed

• If an error occurs in the middle , For example, transfer in B An error occurred , Then the compensation operation of the executed branch will be called , namely ：

  Transfer out succeeded => Failed to execute transfer in => Successful implementation of transfer in compensation => Successfully execute transfer out compensation => Global transaction rollback completed

In this picture , Our global business sponsor , Arrange the information of the whole global transaction , Including the definition of forward operation and reverse compensation operation of each step , Submit to server , The server will follow the steps above SAGA The logic of .

If any forward operation fails , For example, the account balance is insufficient or the account is frozen , that dtm It will call the compensation operation of each branch , Roll back , Finally, the transaction rolled back successfully .

TCC: Try-Confirm-Cancel

TCC Transaction mode

What is? TCC,TCC yes Try、Confirm、Cancel Abbreviations of three words , The first is the Pat Helland On 2007 A piece published in was called 《Life beyond Distributed Transactions:an Apostate’s Opinion》 The paper put forward .

TCC It is divided into 3 Stages

• Try Stage ： Trying to perform , Complete all business checks （ Uniformity ）, Reserve necessary business resources （ Quasi isolation ）
• Confirm Stage ： If all branches Try It's all done , Go to Confirm Stage .Confirm Really do business , No business checks , Use only Try Business resources reserved in the stage
• Cancel Stage ： If all branches Try One failed , Go to Cancel Stage .Cancel Release Try Business resources reserved in the stage .

In design ,TCC It is mainly used to deal with high consistency requirements 、 Short transactions that require more flexibility ：

TCC How to achieve better consistency

For our A Inter bank transfer to B Scene , If the SAGA, Adjust the balance in the forward operation middle note , In compensation operation , Reverse adjustment balance , Then this will happen ： If A Deduction successful , Amount transferred to B Failure , Finally roll back , hold A The balance of is adjusted to the initial value . In the whole process, if A I found my balance deducted , But the payee B Delayed receipt of funds , Then it will be right A Cause great trouble .

The above requirements are in SAGA Can't solve , But it can go through TCC To solve , The design skills are as follows ：

• In the account balance Outside the field , Introduce another trading_balance Field
• Try Stage check whether the account is frozen , Check whether the account balance is sufficient , No problem , adjustment trading_balance （ That is, frozen funds in business ）
• Confirm Stage , adjustment balance , adjustment trading_balance （ That is, unfrozen funds in business ）
• Cancel Stage , adjustment trading_balance （ That is, unfrozen funds in business ）

In this case , End user A You won't see your balance deducted , however B And the delay in receiving funds

XA Transaction mode

With the distributed transaction scenario , There will be ways to solve the problem ,XA Specification is the specification for distributed transactions . There are many ways to implement distributed transactions , The most representative is by Oracle Tuxedo The system proposes XA Distributed transaction protocol .XA The agreement consists of two phases （2PC） And three phase submission （3PC） Two implementations .

XA What is it?

XA By X/Open The specification of distributed transaction proposed by the organization ,XA The specification mainly defines ( overall situation ) Transaction manager and ( Local ) Explorer (RM) Interface between . Local databases such as mysql stay XA It plays RM role

AT Pattern is a non intrusive distributed transaction solution .

XA It is divided into two stages ：

• The first stage （prepare）： That is, all participants RM Prepare to execute the transaction and lock the required resources . participants ready when , towards TM The report is ready .
• The second stage (commit/rollback)： When the transaction manager confirms that all participants (RM) all ready after , Send... To all participants commit command .

At present, the mainstream databases basically support XA Business , Include mysql、oracle、sqlserver、postgre

How does the local database support XA Of ：

XA start '4fPqCNTYeSG' --  To start a  xa  Business 
UPDATE `user_account` SET `balance`=balance + 30,`update_time`='2021-06-09 11:50:42.438' WHERE user_id = '1'
XA end '4fPqCNTYeSG'
XA prepare '4fPqCNTYeSG' --  Before this call , Connection is broken , Then the transaction will be rolled back automatically 
--  When all participants have finished prepare, Then enter the second stage   Submit 
xa commit '4fPqCNTYeSG'

XA The characteristics of transactions are ：

Easy to understand
Easy to develop , Operations like rollback , Automatically completed by the underlying database
Locked the resource for a long time , Low concurrency , It is not suitable for high concurrency businesses

AT(Auto Transaction) Pattern

Distributed transaction framework

SEATA

Seata Is an open source distributed transaction solution , Committed to providing high-performance and easy-to-use distributed transaction services .Seata Will provide users with AT、TCC、SAGA and XA Transaction mode , Create a one-stop distributed solution for users .