Distributed transaction

The transaction ACID principle

A(Atomicity)： Atomicity ： All operations of the transaction , All or nothing , All or nothing

C(Consistency)： Uniformity ： To ensure the internal integrity of the database , Declarative constraints

I(Isolation)： Isolation, ： Transactions operating on the same resource cannot occur at the same time

D(Durability)： persistence ： On the database All modifications will be permanently saved , Whether or not there is a fault

  Micro service case

Micro service ordering business , When placing an order, the order service will be called , Create an order and write it to the database .

Then the order service invokes the account service and the inventory service ：
• The account service is responsible for deducting the user's balance
• Inventory service is responsible for deducting inventory of goods  ​​​​

 

But if the order microservices and account microservices operate the database, there is no problem , But when it comes to the inventory micro service, the inventory is insufficient , Then the inventory microservice will definitely operate, and the microservice will fail  . What do you do then ？ Ordinary transactions cannot listen to remotely invoked account services and inventory services ,@Transactional Is the use of the AOP To achieve , Enhanced our approach , Post enhancement is used. If there is an error, rollback the transaction , But we can't operate remote account service and inventory service , Because every micro service is independent , You can only access the open interface in the microservice and cannot operate it

CAP theory

C（Consistency）: Uniformity       A（Availability）： Usability      P（Partition tolerance）： Partition tolerance

These three indicators cannot be met in distributed systems