Seata之分布式事务问题及解决方案_Java

seata–分布式事务解决方案

简介

seata 是阿里开源的分布式事务解决方案，提供 at、xa、tcc、saga 四种事务模式，支持微服务架构下的数据一致性。

官网：https://seata.apache.org/zh-cn/

方案	核心特点	适用场景
seata	多模式支持，代码侵入性低，社区活跃	复杂业务场景，需灵活选择模式
阿里云 gts	商业版方案，功能全面，性能强	企业级付费场景
rocketmq 事务消息	基于消息队列实现最终一致性	异步高吞吐场景
lcn	基于代理模式，实现简单	轻量级快速接入

环境搭建

1.微服务

创建 spring cloud 项目，推荐使用以下组件：

<!-- spring cloud alibaba 依赖 -->
<dependency>
    <groupid>com.alibaba.cloud</groupid>
    <artifactid>spring-cloud-starter-alibaba-seata</artifactid>
</dependency>

2.sql

每个微服务数据库需创建 undo_log 表（at模式必需）：

-- 每个业务数据库均需执行
create table if not exists `undo_log`
(
    `id`            bigint(20)   not null auto_increment,
    `branch_id`     bigint(20)   not null,
    `xid`           varchar(100) not null,
    `context`       varchar(128) not null,
    `rollback_info` longblob     not null,
    `log_status`    int(11)      not null,
    `log_created`   datetime     not null,
    `log_modified`  datetime     not null,
    primary key (`id`),
    unique key `ux_undo_log` (`xid`, `branch_id`)
) engine = innodb
  auto_increment = 1
  default charset = utf8;

3.seata-server

1.下载 :https://seata.apache.org/zh-cn/download/seata-server
2.解压并启动：seata-server.bat
3.控制台:http://127.0.0.1:7091/#/transaction/list

4.微服务配置

依赖

除基础依赖外需添加：

<dependency>
  <groupid>com.alibaba.cloud</groupid>
  <artifactid>spring-cloud-starter-alibaba-seata</artifactid>
</dependency>

配置

每个微服务创建 file.conf文件，完整内容如下；

【微服务只需要复制 service 块配置即可】

#
# licensed to the apache software foundation (asf) under one or more
# contributor license agreements.  see the notice file distributed with
# this work for additional information regarding copyright ownership.
# the asf licenses this file to you under the apache license, version 2.0
# (the "license"); you may not use this file except in compliance with
# the license.  you may obtain a copy of the license at
#
#     http://www.apache.org/licenses/license-2.0
#
# unless required by applicable law or agreed to in writing, software
# distributed under the license is distributed on an "as is" basis,
# without warranties or conditions of any kind, either express or implied.
# see the license for the specific language governing permissions and
# limitations under the license.
#

transport {
  # tcp, unix-domain-socket
  type = "tcp"
  #nio, native
  server = "nio"
  #enable heartbeat
  heartbeat = true
  # the tm client batch send request enable
  enabletmclientbatchsendrequest = false
  # the rm client batch send request enable
  enablermclientbatchsendrequest = true
   # the rm client rpc request timeout
  rpcrmrequesttimeout = 2000
  # the tm client rpc request timeout
  rpctmrequesttimeout = 30000
  # the rm client rpc request timeout
  rpcrmrequesttimeout = 15000
  #thread factory for netty
  threadfactory {
    bossthreadprefix = "nettyboss"
    workerthreadprefix = "nettyservernioworker"
    serverexecutorthread-prefix = "nettyserverbizhandler"
    sharebossworker = false
    clientselectorthreadprefix = "nettyclientselector"
    clientselectorthreadsize = 1
    clientworkerthreadprefix = "nettyclientworkerthread"
    # netty boss thread size
    bossthreadsize = 1
    #auto default pin or 8
    workerthreadsize = "default"
  }
  shutdown {
    # when destroy server, wait seconds
    wait = 3
  }
  serialization = "seata"
  compressor = "none"
}
service {
  #transaction service group mapping
  vgroupmapping.default_tx_group = "default"
  #only support when registry.type=file, please don't set multiple addresses
  default.grouplist = "127.0.0.1:8091"
  #degrade, current not support
  enabledegrade = false
  #disable seata
  disableglobaltransaction = false
}

client {
  rm {
    asynccommitbufferlimit = 10000
    lock {
      retryinterval = 10
      retrytimes = 30
      retrypolicybranchrollbackonconflict = true
    }
    reportretrycount = 5
    tablemetacheckenable = false
    tablemetacheckerinterval = 60000
    reportsuccessenable = false
    sagabranchregisterenable = false
    sagajsonparser = "fastjson"
    sagaretrypersistmodeupdate = false
    sagacompensatepersistmodeupdate = false
    tccactioninterceptororder = -2147482648 #ordered.highest_precedence + 1000
    sqlparsertype = "druid"
    branchexecutiontimeoutxa = 60000
    connectiontwophaseholdtimeoutxa = 10000
  }
  tm {
    commitretrycount = 5
    rollbackretrycount = 5
    defaultglobaltransactiontimeout = 60000
    degradecheck = false
    degradecheckperiod = 2000
    degradecheckallowtimes = 10
    interceptororder = -2147482648 #ordered.highest_precedence + 1000
  }
  undo {
    datavalidation = true
    onlycareupdatecolumns = true
    logserialization = "jackson"
    logtable = "undo_log"
    compress {
      enable = true
      # allow zip, gzip, deflater, lz4, bzip2, zstd default is zip
      type = zip
      # if rollback info size > threshold, then will be compress
      # allow k m g t
      threshold = 64k
    }
  }
  loadbalance {
      type = "xid"
      virtualnodes = 10
  }
}
log {
  exceptionrate = 100
}
tcc {
  fence {
    # tcc fence log table name
    logtablename = tcc_fence_log
    # tcc fence log clean period
    cleanperiod = 1h
  }
}

事务模式

1.at模式（推荐:自动补偿）

二阶提交协议原理

原理：基于反向sql补偿，自动生成回滚日志

使用：

@service
public class orderservice {
    @globaltransactional(name = "createorder", timeoutmills = 60000)
    public void createorder(orderdto order) {
        // 1. 扣减库存（调用库存服务）
        storagefeignclient.deduct(order.getproductid());
        // 2. 创建订单（本地事务）
        ordermapper.insert(order);
        // 3. 模拟异常触发回滚
        int i = 1 / 0; 
    }
}

关键机制：

一阶段：提交本地事务，生成回滚日志（undo_log）
二阶段：成功则异步删除日志，失败则通过日志反向补偿

特点：

对代码无侵入
需创建undo_log表
适用于大多数crud场景

2.xa模式（强一致）

原理：基于数据库xa协议的两阶段提交

配置：

seata:
  data-source-proxy-mode: xa  # 默认at

特点：

基于数据库xa协议
两阶段提交（2pc）
事务持有锁时间较长，适合短事务

适用场景：强一致性需求，支持xa协议的数据库（如mysql 5.7+）

3.tcc模式（手动补偿）

原理：try-confirm-cancel 三阶段控制

实现：

// 1. 定义tcc接口
public interface storagetccservice {
    @twophasebusinessaction(name = "deduct", 
                          commitmethod = "confirm", 
                          rollbackmethod = "cancel")
    boolean deduct(@businessactioncontextparameter(paramname = "productid") string productid,
                   @businessactioncontextparameter(paramname = "count") integer count);
    
    boolean confirm(businessactioncontext context);
    boolean cancel(businessactioncontext context);
}

// 2. 实现try逻辑
@service
public class storagetccserviceimpl implements storagetccservice {
    @override
    public boolean deduct(string productid, integer count) {
        // try阶段：资源预留（例如冻结库存）
        return storagemapper.freezestock(productid, count) > 0;
    }
    
    @override
    public boolean confirm(businessactioncontext context) {
        // confirm阶段：真实扣减（例如删除冻结记录）
        string productid = context.getactioncontext("productid");
        integer count = context.getactioncontext("count");
        return storagemapper.reducestock(productid, count) > 0;
    }
    
    @override
    public boolean cancel(businessactioncontext context) {
        // cancel阶段：释放资源（例如恢复冻结库存）
        string productid = context.getactioncontext("productid");
        integer count = context.getactioncontext("count");
        return storagemapper.unfreezestock(productid, count) > 0;
    }
}

使用限制：

需自行实现try/confirm/cancel方法
confirm和cancel需保证幂等性

特点：高性能，但需手动编写补偿逻辑

4.saga模式（长事务）

实现方式：

通过状态机配置补偿策略：

@sagastart
public void createordersaga(order order) {
    // 1. 创建订单
    orderservice.create(order);
    // 2. 调用支付服务（若失败则触发逆向操作）
    paymentservice.pay(order.getid());
}
// 定义补偿方法
@compensate
public void compensateorder(order order) {
    orderservice.delete(order.getid());
}

原理：长事务拆分+逆向补偿

适用场景：跨系统长时间操作（如订单+物流+支付）

总结

模式选型对照表：

模式	一致性	性能	侵入性	适用场景
at	弱一致	高	低	常规业务（库存扣减、订单创建）
tcc	强一致	中	高	资金交易、需精准控制
xa	强一致	低	低	银行转账、短事务
saga	最终一致	高	中	跨系统长流程（订单+物流+支付）

核心要点：

分类	要点说明
选型	at模式适用于大多数场景，tcc适合高性能要求，xa适合强一致性
配置	确保seata-server与微服务的registry配置一致
事务id	通过rootcontext.getxid()可获取当前事务id
排错	检查undo_log表记录，查看seata-server控制台日志

最佳实践：