Repository files navigation

English |简体中文

DTM is a distributed transaction framework which provides cross-service eventual data consistency. It provides saga, tcc, xa, 2-phase message, outbox, workflow patterns for a variety of application scenarios. It also supports multiple languages and multiple store engine to form up a transaction as following:

Tencent

Bytedance

Ivydad

More

• Support for multiple transaction modes: SAGA, TCC, XA, Workflow, Outbox
• Multiple languages support: SDK for Go, Java, PHP, C#, Python, Nodejs
• Better Outbox: 2-phase messages, a more elegant solution than Outbox, support multi-databases
• Multiple database transaction support: MySQL/MariaDB, Redis, MongoDB, Postgres, TDSQL, etc.
• Support for multiple storage engines: MySQL/MariaDB (common), Redis (high performance), BoltDB (dev&test), MongoDB (under planning)
• Support for multiple microservices architectures:go-zero, go-kratos/kratos, polarismesh/polaris
• Support for high availability and easy horizontal scaling

DTM can be applied to data consistency issues in a large number of scenarios, here are a few common ones

• cache management: thoroughly guarantee the cache final consistency and strong consistency
• flash-sales to deduct inventory: in extreme cases, it is also possible to ensure that the precise inventory in Redis is exactly the same as the final order created, without the need for manual adjustment
• Non-monolithic order system: Dramatically simplifies the architecture
• Event publishing/subscription: better outbox pattern

git clone https://github.com/dtm-labs/dtm&&cd dtmgo run main.go

Suppose we want to perform an inter-bank transfer. The operations of transfer out (TransOut) and transfer in (TransIn) are coded in separate micro-services.

Here is an example to illustrate a solution of dtm to this problem:

git clone https://github.com/dtm-labs/quick-start-sample.git&&cd quick-start-sample/workflow-grpcgo run main.go

wfName:="workflow-grpc"err=workflow.Register(wfName,func(wf*workflow.Workflow,data []byte)error {// ...// Define a transaction branch for TransOutwf.NewBranch().OnRollback(func(bb*dtmcli.BranchBarrier)error {// compensation for TransOut_,err:=busiCli.TransOutRevert(wf.Context,&req)returnerr  })_,err=busiCli.TransOut(wf.Context,&req)// check error// Define another transaction branch for TransInwf.NewBranch().OnRollback(func(bb*dtmcli.BranchBarrier)error {_,err:=busiCli.TransInRevert(wf.Context,&req)returnerr  })_,err=busiCli.TransIn(wf.Context,&req)returnerr}// ...req:=busi.BusiReq{Amount:30,TransInResult:""}data,err:= proto.Marshal(&req)// Execute workflow_,err=workflow.ExecuteCtx(wfName,shortuuid.New(),data)logger.Infof("result of workflow.Execute is: %v",err)

When the above code runs, we can see in the console that servicesTransOut,TransIn has been called.

If any forward operation fails, DTM invokes the corresponding compensating operation of each sub-transaction to roll back, after which the transaction is successfully rolled back.

Let's purposely trigger the failure of the second sub-transaction and watch what happens

// req := busi.BusiReq{Amount: 30, TransInResult: ""}req:= busi.BusiReq{Amount:30,TransInResult:"FAILURE"}})

we can see in the console that servicesTransOut,TransIn,TransOutRevert has been called

If you want more quick start examples, please refer todtm-labs/quick-start-sample

The above example mainly demonstrates the flow of a distributed transaction. More on this, including practical examples of how to interact with an actual database, how to do compensation, how to do rollback, etc. please refer todtm-examples for more examples.

If you think this project is interesting, or helpful to you, please give a star!