AT模式执行流程分析
第一阶段
在事务分支中执行sql前解析将要执行的sql，生成相对应的undolog,同时入库。
第二阶段
分布式事务如果操作成功，TC通知RM异步删除undolog，如果分布式事务操作失败，TC向RM发送回滚请求，RM通过XID和BRANCHID找到相应的undolog，通过回滚记录反向更新sql，完成事务回滚。
源码分析
本源码分析从seata提供的seata-samples项目中开始分析。
初始化
初始时bean配置
1.Seata中的代理数据源对普通数据源做一层代理，代理了阿里的druid数据源。
@Bean
public DataSourceProxy dataSourceProxy(DruidDataSource druidDataSource){
return new DataSourceProxy(druidDataSource);
}
2.配置了一个GlobalTransactionScanner的bean，第一个构造参数为应用id，第二个参数为事务分组。
@Bean
public GlobalTransactionScanner globalTransactionScanner(){
return new GlobalTransactionScanner("account-gts-seata-example", "my_test_tx_group");
}
GlobalTransactionScanner是关键的入口。代码如下：
public class GlobalTransactionScanner extends AbstractAutoProxyCreator
implements InitializingBean, ApplicationContextAware,
DisposableBean {
private static final long serialVersionUID = 1L;
private static final Logger LOGGER = LoggerFactory.getLogger(GlobalTransactionScanner.class);
private static final int AT_MODE = 1;
private static final int MT_MODE = 2;
private static final int ORDER_NUM = 1024;
private static final int DEFAULT_MODE = AT_MODE + MT_MODE;
private static final Set<String> PROXYED_SET = new HashSet<>();
private MethodInterceptor interceptor;
private final String applicationId;
private final String txServiceGroup;
private final int mode;
private final boolean disableGlobalTransaction = ConfigurationFactory.getInstance().getBoolean(
ConfigurationKeys.DISABLE_GLOBAL_TRANSACTION, DEFAULT_DISABLE_GLOBAL_TRANSACTION);
private final FailureHandler failureHandlerHook;
private ApplicationContext applicationContext;
该类分别实现了Spring的3个接口InitializingBean，ApplicationContextAware，DisposableBean三个接口。
该类前面都是一些构造方法，关键点在于afterPropertiesSet方法；
@Override
public void afterPropertiesSet() {
if (disableGlobalTransaction) {
if (LOGGER.isInfoEnabled()) {
LOGGER.info("Global transaction is disabled.");
}
return;
}
initClient();
}
该方法中调用了initClient方法：
private void initClient() {
if (LOGGER.isInfoEnabled()) {
LOGGER.info("Initializing Global Transaction Clients ... ");
}
if (StringUtils.isNullOrEmpty(applicationId) || StringUtils.isNullOrEmpty(txServiceGroup)) {
throw new IllegalArgumentException(String.format("applicationId: %s, txServiceGroup: %s", applicationId, txServiceGroup));
}
//init TM
TMClient.init(applicationId, txServiceGroup);
if (LOGGER.isInfoEnabled()) {
LOGGER.info("Transaction Manager Client is initialized. applicationId[{}] txServiceGroup[{}]", applicationId, txServiceGroup);
}
//init RM
RMClient.init(applicationId, txServiceGroup);
if (LOGGER.isInfoEnabled()) {
LOGGER.info("Resource Manager is initialized. applicationId[{}] txServiceGroup[{}]", applicationId, txServiceGroup);
}
if (LOGGER.isInfoEnabled()) {
LOGGER.info("Global Transaction Clients are initialized. ");
}
registerSpringShutdownHook();
}
该方法中初始化了TMClient和RMClient，初始化参数为Configuration配置类中传递过来的applicationId和txServiceGroup，并且执行registerSpringShutdownHook方法注册了一个spring的ShutdownHook。
TMClient和RMClient初始化
TMClient.init():
public static void init(String applicationId, String transactionServiceGroup) {
TmRpcClient tmRpcClient = TmRpcClient.getInstance(applicationId, transactionServiceGroup);
tmRpcClient.init();
}
从init方法中可以看出调用了tmRpcClient.init()方法，在这个方法中又调用了它的父类初始化方法：
@Override
public void init() {
clientBootstrap.setChannelHandlers(new ClientHandler());
clientBootstrap.start();
timerExecutor.scheduleAtFixedRate(new Runnable() {
@Override
public void run() {
clientChannelManager.reconnect(getTransactionServiceGroup());
}
}, SCHEDULE_DELAY_MILLS, SCHEDULE_INTERVAL_MILLS, TimeUnit.MILLISECONDS);
if (NettyClientConfig.isEnableClientBatchSendRequest()) {
mergeSendExecutorService = new ThreadPoolExecutor(MAX_MERGE_SEND_THREAD,
MAX_MERGE_SEND_THREAD,
KEEP_ALIVE_TIME, TimeUnit.MILLISECONDS,
new LinkedBlockingQueue<>(),
new NamedThreadFactory(getThreadPrefix(), MAX_MERGE_SEND_THREAD));
mergeSendExecutorService.submit(new MergedSendRunnable());
}
super.init();
}
底层采用了netty作为客户端，并且在该初始化方法中还开启了一个定时任务线程，定时重连，clientChannelManager.reconnect(getTransactionServiceGroup())
void reconnect(String transactionServiceGroup) {
List<String> availList = null;
try {
availList = getAvailServerList(transactionServiceGroup);
} catch (Exception e) {
LOGGER.error("Failed to get available servers: {}", e.getMessage(), e);
return;
}
if (CollectionUtils.isEmpty(availList)) {
String serviceGroup = RegistryFactory.getInstance()
.getServiceGroup(transactionServiceGroup);
LOGGER.error("no available service '{}' found, please make sure registry config correct", serviceGroup);
return;
}
for (String serverAddress : availList) {
try {
acquireChannel(serverAddress);
} catch (Exception e) {
LOGGER.error("{} can not connect to {} cause:{}",FrameworkErrorCode.NetConnect.getErrCode(), serverAddress, e.getMessage(), e);
}
}
}
首先根据事务的分组名获取获取到对应的seata-server的ip地址列表，然后进行重连，getAvailServerList方法：
private List<String> getAvailServerList(String transactionServiceGroup) throws Exception {
List<InetSocketAddress> availInetSocketAddressList = RegistryFactory.getInstance()
.lookup(transactionServiceGroup);
if (CollectionUtils.isEmpty(availInetSocketAddressList)) {
return Collections.emptyList();
}
return availInetSocketAddressList.stream()
.map(NetUtil::toStringAddress)
.collect(Collectors.toList());
}
该方法中对不同的注册中心做了适配，默认情况下是以FILE作为注册中心的，还有consul、etcd3、eureka、nacos、redis、sofa、zk这几种注册中心。
综上所述，TMClient的初始化为以下两个个步骤：
启动netty客户端> 延迟60秒每10秒执行一次重新连接seata-server
RmClient.init()
public static void init(String applicationId, String transactionServiceGroup) {
RmRpcClient rmRpcClient = RmRpcClient.getInstance(applicationId, transactionServiceGroup);
rmRpcClient.setResourceManager(DefaultResourceManager.get());
rmRpcClient.setClientMessageListener(new RmMessageListener(DefaultRMHandler.get(), rmRpcClient));
rmRpcClient.init();
}
从init方法中可以看出调用了 rmRpcClient.init()方法，后面的代码基本和TMClient的初始化方法一致。
小结：在Spring应用启动时启动两个客户端TMClient和RMClient，TMClient与Server建立连接并发送消息，RMClient与Server建立连接并负责接收二阶段提交、回滚的消息并在RMHandler（回调执行器）中处理。
两阶段提交过程分析
第一阶段
拦截器中开启事务
在需要加全局事务的方法中，会加上GlobalTransactional注解，注解往往对应着拦截器，Seata中拦截全局事务的拦截器是`GlobalTransactionalInterceptor
@Override
@GlobalTransactional(timeoutMills = 300000, name = "dubbo-gts-seata-example")
public ObjectResponse handleBusiness(BusinessDTO businessDTO) {
System.out.println("开始全局事务，XID = " + RootContext.getXID());
ObjectResponse<Object> objectResponse = new ObjectResponse<>();
//1、扣减库存
CommodityDTO commodityDTO = new CommodityDTO();
commodityDTO.setCommodityCode(businessDTO.getCommodityCode());
commodityDTO.setCount(businessDTO.getCount());
ObjectResponse storageResponse = storageDubboService.decreaseStorage(commodityDTO);
//2、创建订单
OrderDTO orderDTO = new OrderDTO();
orderDTO.setUserId(businessDTO.getUserId());
orderDTO.setCommodityCode(businessDTO.getCommodityCode());
orderDTO.setOrderCount(businessDTO.getCount());
orderDTO.setOrderAmount(businessDTO.getAmount());
ObjectResponse<OrderDTO> response = orderDubboService.createOrder(orderDTO);
if (storageResponse.getStatus() != 200 || response.getStatus() != 200) {
throw new DefaultException(RspStatusEnum.FAIL);
}
objectResponse.setStatus(RspStatusEnum.SUCCESS.getCode());
objectResponse.setMessage(RspStatusEnum.SUCCESS.getMessage());
objectResponse.setData(response.getData());
return objectResponse;
}
拦截方法如下
@Override
public Object invoke(final MethodInvocation methodInvocation) throws Throwable {
Class<?> targetClass = methodInvocation.getThis() != null ? AopUtils.getTargetClass(methodInvocation.getThis())
: null;
Method specificMethod = ClassUtils.getMostSpecificMethod(methodInvocation.getMethod(), targetClass);
final Method method = BridgeMethodResolver.findBridgedMethod(specificMethod);
final GlobalTransactional globalTransactionalAnnotation =
getAnnotation(method, targetClass, GlobalTransactional.class);
final GlobalLock globalLockAnnotation = getAnnotation(method, targetClass, GlobalLock.class);
if (!disable && globalTransactionalAnnotation != null) {
return handleGlobalTransaction(methodInvocation, globalTransactionalAnnotation);
} else if (!disable && globalLockAnnotation != null) {
return handleGlobalLock(methodInvocation);
} else {
return methodInvocation.proceed();
}
}
判断方法上是否有全局事务注解，调用handleGlobalTransaction方法开启全局事务；否则按普通方法执行。handleGlobalTransaction方法如下：
private Object handleGlobalTransaction(final MethodInvocation methodInvocation,
final GlobalTransactional globalTrxAnno) throws Throwable {
try {
return transactionalTemplate.execute(new TransactionalExecutor() {
@Override
public Object execute() throws Throwable {
return methodInvocation.proceed();
}
public String name() {
String name = globalTrxAnno.name();
if (!StringUtils.isNullOrEmpty(name)) {
return name;
}
return formatMethod(methodInvocation.getMethod());
}
@Override
public TransactionInfo getTransactionInfo() {
TransactionInfo transactionInfo = new TransactionInfo();
transactionInfo.setTimeOut(globalTrxAnno.timeoutMills());
transactionInfo.setName(name());
transactionInfo.setPropagation(globalTrxAnno.propagation());
Set<RollbackRule> rollbackRules = new LinkedHashSet<>();
for (Class<?> rbRule : globalTrxAnno.rollbackFor()) {
rollbackRules.add(new RollbackRule(rbRule));
}
for (String rbRule : globalTrxAnno.rollbackForClassName()) {
rollbackRules.add(new RollbackRule(rbRule));
}
for (Class<?> rbRule : globalTrxAnno.noRollbackFor()) {
rollbackRules.add(new NoRollbackRule(rbRule));
}
for (String rbRule : globalTrxAnno.noRollbackForClassName()) {
rollbackRules.add(new NoRollbackRule(rbRule));
}
transactionInfo.setRollbackRules(rollbackRules);
return transactionInfo;
}
});
} catch (TransactionalExecutor.ExecutionException e) {
TransactionalExecutor.Code code = e.getCode();
switch (code) {
case RollbackDone:
throw e.getOriginalException();
case BeginFailure:
failureHandler.onBeginFailure(e.getTransaction(), e.getCause());
throw e.getCause();
case CommitFailure:
failureHandler.onCommitFailure(e.getTransaction(), e.getCause());
throw e.getCause();
case RollbackFailure:
failureHandler.onRollbackFailure(e.getTransaction(), e.getCause());
throw e.getCause();
case RollbackRetrying:
failureHandler.onRollbackRetrying(e.getTransaction(), e.getCause());
throw e.getCause();
default:
throw new ShouldNeverHappenException(String.format("Unknown TransactionalExecutor.Code: %s", code));
}
}
}
该方法中调用了TransactionalTemplate的execute方法：
public Object execute(TransactionalExecutor business) throws Throwable {
// 1 get transactionInfo
TransactionInfo txInfo = business.getTransactionInfo();
if (txInfo == null) {
throw new ShouldNeverHappenException("transactionInfo does not exist");
}
// 1.1 get or create a transaction
GlobalTransaction tx = GlobalTransactionContext.getCurrentOrCreate();
// 1.2 Handle the Transaction propatation and the branchType
Propagation propagation = txInfo.getPropagation();
SuspendedResourcesHolder suspendedResourcesHolder = null;
try {
switch (propagation) {
case NOT_SUPPORTED:
suspendedResourcesHolder = tx.suspend(true);
return business.execute();
case REQUIRES_NEW:
suspendedResourcesHolder = tx.suspend(true);
break;
case SUPPORTS:
if (!existingTransaction()) {
return business.execute();
}
break;
case REQUIRED:
break;
case NEVER:
if (existingTransaction()) {
throw new TransactionException(
String.format("Existing transaction found for transaction marked with propagation 'never',xid = %s"
,RootContext.getXID()));
} else {
return business.execute();
}
case MANDATORY:
if (!existingTransaction()) {
throw new TransactionException("No existing transaction found for transaction marked with propagation 'mandatory'");
}
break;
default:
throw new TransactionException("Not Supported Propagation:" + propagation);
}
try {
// 2. begin transaction
beginTransaction(txInfo, tx);
Object rs = null;
try {
// Do Your Business
rs = business.execute();
} catch (Throwable ex) {
// 3.the needed business exception to rollback.
completeTransactionAfterThrowing(txInfo, tx, ex);
throw ex;
}
// 4. everything is fine, commit.
commitTransaction(tx);
return rs;
} finally {
//5. clear
triggerAfterCompletion();
cleanUp();
}
} finally {
tx.resume(suspendedResourcesHolder);
}
}
可以看出该方法中主要有以下几个步骤：
1、获取事务信息，
2、开启事务
3、执行业务方法
4、提交事务（没有抛出异常）
5、回滚操作（抛出异常）
beginTransaction最终调用了DefaultGlobalTransaction的begin方法，代码如下：
@Override
public void begin(int timeout, String name) throws TransactionException {
if (role != GlobalTransactionRole.Launcher) {
assertXIDNotNull();
if (LOGGER.isDebugEnabled()) {
LOGGER.debug("Ignore Begin(): just involved in global transaction [{}]", xid);
}
return;
}
assertXIDNull();
if (RootContext.getXID() != null) {
throw new IllegalStateException();
}
xid = transactionManager.begin(null, null, name, timeout);
status = GlobalStatus.Begin;
RootContext.bind(xid);
if (LOGGER.isInfoEnabled()) {
LOGGER.info("Begin new global transaction [{}]", xid);
}
}
通过transactionManager.begin()方法通过TmRpcClient与server通信并生成一个xid，再将将xid绑定到Root上下文中。
全局事务拦截成功后还是会执行原业务方法，但是由于seata代理了数据源，sql解析undolog是在代理数据源中完成的。seata不止会代理数据源，还会对Connection，Statement做的代理封装。
对sql解析实际是发生在StatementProxy中：
@Override
public ResultSet executeQuery(String sql) throws SQLException {
this.targetSQL = sql;
return ExecuteTemplate.execute(this, (statement, args) -> statement.executeQuery((String) args[0]), sql);
}
最终执行了ExecuteTemplate类的execute方法：
public static <T, S extends Statement> T execute(List<SQLRecognizer> sqlRecognizers,
StatementProxy<S> statementProxy,
StatementCallback<T, S> statementCallback,
Object... args) throws SQLException {
if (!shouldExecuteInATMode()) {
// Just work as original statement
return statementCallback.execute(statementProxy.getTargetStatement(), args);
}
if (sqlRecognizers == null) {
sqlRecognizers = SQLVisitorFactory.get(
statementProxy.getTargetSQL(),
statementProxy.getConnectionProxy().getDbType());
}
Executor<T> executor;
if (CollectionUtils.isEmpty(sqlRecognizers)) {
executor = new PlainExecutor<>(statementProxy, statementCallback);
} else {
if (sqlRecognizers.size() == 1) {
SQLRecognizer sqlRecognizer = sqlRecognizers.get(0);
switch (sqlRecognizer.getSQLType()) {
case INSERT:
executor = new InsertExecutor<>(statementProxy, statementCallback, sqlRecognizer);
break;
case UPDATE:
executor = new UpdateExecutor<>(statementProxy, statementCallback, sqlRecognizer);
break;
case DELETE:
executor = new DeleteExecutor<>(statementProxy, statementCallback, sqlRecognizer);
break;
case SELECT_FOR_UPDATE:
executor = new SelectForUpdateExecutor<>(statementProxy, statementCallback, sqlRecognizer);
break;
default:
executor = new PlainExecutor<>(statementProxy, statementCallback);
break;
}
} else {
executor = new MultiExecutor<>(statementProxy, statementCallback, sqlRecognizers);
}
}
T rs;
try {
rs = executor.execute(args);
} catch (Throwable ex) {
if (!(ex instanceof SQLException)) {
// Turn other exception into SQLException
ex = new SQLException(ex);
}
throw (SQLException) ex;
}
return rs;
}
首先判断是否存在全局事务，不在全局事务中按普通方法执行，如果在全局事务中则开始解析sql，对不同的DML语句做响应的处理，在调用执行方法。
以上流程为：
1、先判断是否开启了全局事务，如果没有，不走代理，不解析sql。
2、调用SQLVisitorFactory对目标sql进行解析。
3、针对特定类型sql操作(INSERT,UPDATE,DELETE,SELECT_FOR_UPDATE)等进行特殊解析。
4、执行sql并返回结果。
关键点在于特定类型执行器里面的execute方法（下面以InsertExecutor类的execute方法举例），他调用了他的父类BaseTransactionalExecutor的execute方法，
@Override
public T execute(Object... args) throws Throwable {
if (RootContext.inGlobalTransaction()) {
String xid = RootContext.getXID();
statementProxy.getConnectionProxy().bind(xid);
}
statementProxy.getConnectionProxy().setGlobalLockRequire(RootContext.requireGlobalLock());
return doExecute(args);
}
将XID绑定到connectionProxy中并调用了doExecute方法，这里又调用了它的子类的AbstractDMLBaseExecutor的doExecute方法。
@Override
public T doExecute(Object... args) throws Throwable {
AbstractConnectionProxy connectionProxy = statementProxy.getConnectionProxy();
if (connectionProxy.getAutoCommit()) {
return executeAutoCommitTrue(args);
} else {
return executeAutoCommitFalse(args);
}
}
executeAutoCommitTrue方法中也会将AutoCommit属性设置为false，对sql进行解析生成undolog，防止在undolog生成之前入库.
protected T executeAutoCommitTrue(Object[] args) throws Throwable {
ConnectionProxy connectionProxy = statementProxy.getConnectionProxy();
try {
connectionProxy.setAutoCommit(false);
return new LockRetryPolicy(connectionProxy).execute(() -> {
T result = executeAutoCommitFalse(args);
connectionProxy.commit();
return result;
});
} catch (Exception e) {
// when exception occur in finally,this exception will lost, so just print it here
LOGGER.error("execute executeAutoCommitTrue error:{}", e.getMessage(), e);
if (!LockRetryPolicy.isLockRetryPolicyBranchRollbackOnConflict()) {
connectionProxy.getTargetConnection().rollback();
}
throw e;
} finally {
connectionProxy.getContext().reset();
connectionProxy.setAutoCommit(true);
}
}
在将AutoCommit设置为false后会执行executeAutoCommitFalse(args)，
protected T executeAutoCommitFalse(Object[] args) throws Exception {
TableRecords beforeImage = beforeImage();
T result = statementCallback.execute(statementProxy.getTargetStatement(), args);
TableRecords afterImage = afterImage(beforeImage);
prepareUndoLog(beforeImage, afterImage);
return result;
}
executeAutoCommitFalse中主要分四步执行：
1、获取sql执行前镜像beforeImage；
2、执行sql；
3、获取sql执行后afterimage
4、根据beforeImage，afterImage生成undolog记录并添加到connectionProxy的上下文中。
分支事务注册与事务提交
业务sql和undolog执行完成后会在代理连接ConnectionProxy中执行commit操作，
@Override
public void commit() throws SQLException {
try {
LOCK_RETRY_POLICY.execute(() -> {
doCommit();
return null;
});
} catch (SQLException e) {
throw e;
} catch (Exception e) {
throw new SQLException(e);
}
}
private void doCommit() throws SQLException {
if (context.inGlobalTransaction()) {
processGlobalTransactionCommit();
} else if (context.isGlobalLockRequire()) {
processLocalCommitWithGlobalLocks();
} else {
targetConnection.commit();
}
}
判断：
如果处于全局事务中则调用processGlobalTransactionCommit处理全局事务提交；
如果加了全局锁注释调用 processLocalCommitWithGlobalLocks()加全局锁并提交；
否在直接进行事务提交。
private void processGlobalTransactionCommit() throws SQLException {
try {
register();
} catch (TransactionException e) {
recognizeLockKeyConflictException(e, context.buildLockKeys());
}
try {
UndoLogManagerFactory.getUndoLogManager(this.getDbType()).flushUndoLogs(this);
targetConnection.commit();
} catch (Throwable ex) {
LOGGER.error("process connectionProxy commit error: {}", ex.getMessage(), ex);
report(false);
throw new SQLException(ex);
}
if (IS_REPORT_SUCCESS_ENABLE) {
report(true);
}
context.reset();
}
processGlobalTransactionCommit方法有以下几个操作步骤：
1、注册分支事务，将branchId分支绑定在上下文中。
private void register() throws TransactionException {
if (!context.hasUndoLog() || context.getLockKeysBuffer().isEmpty()) {
return;
}
Long branchId = DefaultResourceManager.get().branchRegister(BranchType.AT, getDataSourceProxy().getResourceId(),
null, context.getXid(), null, context.buildLockKeys());
context.setBranchId(branchId);
}
2、如果包含undolog，则将之前绑定到上下文中的undolog进行入库；
3、提交本地事务；
4、如果操作失败，report()中通过RM提交第一阶段失败消息，如果成功，report()提交第一阶段成功消息
private void report(boolean commitDone) throws SQLException {
if (context.getBranchId() == null) {
return;
}
int retry = REPORT_RETRY_COUNT;
while (retry > 0) {
try {
DefaultResourceManager.get().branchReport(BranchType.AT, context.getXid(), context.getBranchId(),
commitDone ? BranchStatus.PhaseOne_Done : BranchStatus.PhaseOne_Failed, null);
return;
} catch (Throwable ex) {
LOGGER.error("Failed to report [" + context.getBranchId() + "/" + context.getXid() + "] commit done ["
+ commitDone + "] Retry Countdown: " + retry);
retry--;
if (retry == 0) {
throw new SQLException("Failed to report branch status " + commitDone, ex);
}
}
}
}
由于undolog入库和业务sql的执行调用了同一个connection，处于同一个事务中，这就保证了业务sql和undolog肯定是成对存在。
第二阶段
在第一阶段中RmClient的init过程中，seata会找到一个RmClient的回调处理器RMHandlerAT，该类正是负责接收二阶段seata-server发给RmClient的提交、回滚指令，并做出相应的提交、回滚操作。
RMHandlerAT集成自AbstractRMHandler类，该类中有两个handler方法对应事务的提交和回滚操作。
@Override
public BranchCommitResponse handle(BranchCommitRequest request) {
BranchCommitResponse response = new BranchCommitResponse();
exceptionHandleTemplate(new AbstractCallback<BranchCommitRequest, BranchCommitResponse>() {
@Override
public void execute(BranchCommitRequest request, BranchCommitResponse response)
throws TransactionException {
doBranchCommit(request, response);
}
}, request, response);
return response;
}
@Override
public BranchRollbackResponse handle(BranchRollbackRequest request) {
BranchRollbackResponse response = new BranchRollbackResponse();
exceptionHandleTemplate(new AbstractCallback<BranchRollbackRequest, BranchRollbackResponse>() {
@Override
public void execute(BranchRollbackRequest request, BranchRollbackResponse response)
throws TransactionException {
doBranchRollback(request, response);
}
}, request, response);
return response;
}
首先来看全局事务提交：doBranchCommit(request, response);
protected void doBranchCommit(BranchCommitRequest request, BranchCommitResponse response)
throws TransactionException {
String xid = request.getXid();
long branchId = request.getBranchId();
String resourceId = request.getResourceId();
String applicationData = request.getApplicationData();
if (LOGGER.isInfoEnabled()) {
LOGGER.info("Branch committing: " + xid + " " + branchId + " " + resourceId + " " + applicationData);
}
BranchStatus status = getResourceManager().branchCommit(request.getBranchType(), xid, branchId, resourceId,
applicationData);
response.setXid(xid);
response.setBranchId(branchId);
response.setBranchStatus(status);
if (LOGGER.isInfoEnabled()) {
LOGGER.info("Branch commit result: " + status);
}
}
该方法中调用了getResourceManager()结果的branchCommit（）方法。getgetResourceManager（）的结果是通过SPI机制找到的DataSourceManager类，执行该类中的branchCommit方法。
@Override
public BranchStatus branchCommit(BranchType branchType, String xid, long branchId, String resourceId,
String applicationData) throws TransactionException {
return asyncWorker.branchCommit(branchType, xid, branchId, resourceId, applicationData);
}
该方法中调用了asyncWorker异步提交，
@Override
public BranchStatus branchCommit(BranchType branchType, String xid, long branchId, String resourceId,
String applicationData) throws TransactionException {
if (!ASYNC_COMMIT_BUFFER.offer(new Phase2Context(branchType, xid, branchId, resourceId, applicationData))) {
LOGGER.warn("Async commit buffer is FULL. Rejected branch [{}/{}] will be handled by housekeeping later.", branchId, xid);
}
return BranchStatus.PhaseTwo_Committed;
}
该方法中在ASYNC_COMMIT_BUFFER缓存list中新增一个二阶段提交的context对象并为提交，真正的提交逻辑是在该类的init方法中：
public synchronized void init() {
LOGGER.info("Async Commit Buffer Limit: {}", ASYNC_COMMIT_BUFFER_LIMIT);
ScheduledExecutorService timerExecutor = new ScheduledThreadPoolExecutor(1, new NamedThreadFactory("AsyncWorker", 1, true));
timerExecutor.scheduleAtFixedRate(() -> {
try {
doBranchCommits();
} catch (Throwable e) {
LOGGER.info("Failed at async committing ... {}", e.getMessage());
}
}, 10, 1000 * 1, TimeUnit.MILLISECONDS);
}
该类在初始化时会起一个定时任务线程池，每秒执行一次doBranchCommits方法；
private void doBranchCommits() {
if (ASYNC_COMMIT_BUFFER.isEmpty()) {
return;
}
Map<String, List<Phase2Context>> mappedContexts = new HashMap<>(DEFAULT_RESOURCE_SIZE);
while (!ASYNC_COMMIT_BUFFER.isEmpty()) {
//按照resourceid分组
Phase2Context commitContext = ASYNC_COMMIT_BUFFER.poll();
List<Phase2Context> contextsGroupedByResourceId = mappedContexts.computeIfAbsent(commitContext.resourceId, k -> new ArrayList<>());
contextsGroupedByResourceId.add(commitContext);
}
//按照resourceId找到DataSourceProxy；
for (Map.Entry<String, List<Phase2Context>> entry : mappedContexts.entrySet()) {
Connection conn = null;
DataSourceProxy dataSourceProxy;
try {
try {
DataSourceManager resourceManager = (DataSourceManager) DefaultResourceManager.get()
.getResourceManager(BranchType.AT);
dataSourceProxy = resourceManager.get(entry.getKey());
if (dataSourceProxy == null) {
throw new ShouldNeverHappenException("Failed to find resource on " + entry.getKey());
}
conn = dataSourceProxy.getPlainConnection();
} catch (SQLException sqle) {
LOGGER.warn("Failed to get connection for async committing on " + entry.getKey(), sqle);
continue;
}
List<Phase2Context> contextsGroupedByResourceId = entry.getValue();
Set<String> xids = new LinkedHashSet<>(UNDOLOG_DELETE_LIMIT_SIZE);
Set<Long> branchIds = new LinkedHashSet<>(UNDOLOG_DELETE_LIMIT_SIZE);
//根据xid、branchid找到对应的undolog并删除
for (Phase2Context commitContext : contextsGroupedByResourceId) {
xids.add(commitContext.xid);
branchIds.add(commitContext.branchId);
int maxSize = Math.max(xids.size(), branchIds.size());
if (maxSize == UNDOLOG_DELETE_LIMIT_SIZE) {
try {
UndoLogManagerFactory.getUndoLogManager(dataSourceProxy.getDbType()).batchDeleteUndoLog(
xids, branchIds, conn);
} catch (Exception ex) {
LOGGER.warn("Failed to batch delete undo log [" + branchIds + "/" + xids + "]", ex);
}
xids.clear();
branchIds.clear();
}
}
if (CollectionUtils.isEmpty(xids) || CollectionUtils.isEmpty(branchIds)) {
return;
}
try {
UndoLogManagerFactory.getUndoLogManager(dataSourceProxy.getDbType()).batchDeleteUndoLog(xids,
branchIds, conn);
} catch (Exception ex) {
LOGGER.warn("Failed to batch delete undo log [" + branchIds + "/" + xids + "]", ex);
}
if (!conn.getAutoCommit()) {
conn.commit();
}
} catch (Throwable e) {
LOGGER.error(e.getMessage(), e);
try {
conn.rollback();
} catch (SQLException rollbackEx) {
LOGGER.warn("Failed to rollback JDBC resource while deleting undo_log ", rollbackEx);
}
} finally {
if (conn != null) {
try {
conn.close();
} catch (SQLException closeEx) {
LOGGER.warn("Failed to close JDBC resource while deleting undo_log ", closeEx);
}
}
}
}
}
综上事务提交可以分为如下几步，
1、先按照resourceid对提交分组，一个数据库的可以一批操作；
2、根据resourceid找到DataSourceProxy获取一个普通的数据库连接；
3、调用deleteUndoLog(commitContext.xid, commitContext.branchId, conn)删除undolog。
全局事务回滚，doBranchCommit(request, response);：
protected void doBranchRollback(BranchRollbackRequest request, BranchRollbackResponse response)
throws TransactionException {
String xid = request.getXid();
long branchId = request.getBranchId();
String resourceId = request.getResourceId();
String applicationData = request.getApplicationData();
if (LOGGER.isInfoEnabled()) {
LOGGER.info("Branch Rollbacking: " + xid + " " + branchId + " " + resourceId);
}
BranchStatus status = getResourceManager().branchRollback(request.getBranchType(), xid, branchId, resourceId,
applicationData);
response.setXid(xid);
response.setBranchId(branchId);
response.setBranchStatus(status);
if (LOGGER.isInfoEnabled()) {
LOGGER.info("Branch Rollbacked result: " + status);
}
}
该方法中调用了DataSourceManager的branchRollback方法；
@Override
public BranchStatus branchRollback(BranchType branchType, String xid, long branchId, String resourceId,
String applicationData) throws TransactionException {
DataSourceProxy dataSourceProxy = get(resourceId);
if (dataSourceProxy == null) {
throw new ShouldNeverHappenException();
}
try {
UndoLogManagerFactory.getUndoLogManager(dataSourceProxy.getDbType()).undo(dataSourceProxy, xid, branchId);
} catch (TransactionException te) {
StackTraceLogger.info(LOGGER, te,
"branchRollback failed. branchType:[{}], xid:[{}], branchId:[{}], resourceId:[{}], applicationData:[{}]. reason:[{}]",
new Object[]{branchType, xid, branchId, resourceId, applicationData, te.getMessage()});
if (te.getCode() == TransactionExceptionCode.BranchRollbackFailed_Unretriable) {
return BranchStatus.PhaseTwo_RollbackFailed_Unretryable;
} else {
return BranchStatus.PhaseTwo_RollbackFailed_Retryable;
}
}
return BranchStatus.PhaseTwo_Rollbacked;
}
该方法中又调用了UndoLogManager的undo方法；
public void undo(DataSourceProxy dataSourceProxy, String xid, long branchId) throws TransactionException {
Connection conn = null;
ResultSet rs = null;
PreparedStatement selectPST = null;
boolean originalAutoCommit = true;
for (; ; ) {
try {
conn = dataSourceProxy.getPlainConnection();
// The entire undo process should run in a local transaction.
if (originalAutoCommit = conn.getAutoCommit()) {
conn.setAutoCommit(false);
}
// Find UNDO LOG 查到到undolog
selectPST = conn.prepareStatement(SELECT_UNDO_LOG_SQL);
selectPST.setLong(1, branchId);
selectPST.setString(2, xid);
rs = selectPST.executeQuery();
boolean exists = false;
while (rs.next()) {
exists = true;
// It is possible that the server repeatedly sends a rollback request to roll back
// the same branch transaction to multiple processes,
// ensuring that only the undo_log in the normal state is processed.
int state = rs.getInt(ClientTableColumnsName.UNDO_LOG_LOG_STATUS);
//判断当前undolog记录的状态，如果当前状态为不可操作则不进行回滚，防止重复执行undolog操作
if (!canUndo(state)) {
if (LOGGER.isInfoEnabled()) {
LOGGER.info("xid {} branch {}, ignore {} undo_log", xid, branchId, state);
}
return;
}
String contextString = rs.getString(ClientTableColumnsName.UNDO_LOG_CONTEXT);
Map<String, String> context = parseContext(contextString);
byte[] rollbackInfo = getRollbackInfo(rs);
String serializer = context == null ? null : context.get(UndoLogConstants.SERIALIZER_KEY);
UndoLogParser parser = serializer == null ? UndoLogParserFactory.getInstance()
: UndoLogParserFactory.getInstance(serializer);
BranchUndoLog branchUndoLog = parser.decode(rollbackInfo);
try {
// put serializer name to local
setCurrentSerializer(parser.getName());
List<SQLUndoLog> sqlUndoLogs = branchUndoLog.getSqlUndoLogs();
if (sqlUndoLogs.size() > 1) {
Collections.reverse(sqlUndoLogs);
}
for (SQLUndoLog sqlUndoLog : sqlUndoLogs) {
TableMeta tableMeta = TableMetaCacheFactory.getTableMetaCache(dataSourceProxy.getDbType()).getTableMeta(
conn, sqlUndoLog.getTableName(), dataSourceProxy.getResourceId());
sqlUndoLog.setTableMeta(tableMeta);
AbstractUndoExecutor undoExecutor = UndoExecutorFactory.getUndoExecutor(
dataSourceProxy.getDbType(), sqlUndoLog);
undoExecutor.executeOn(conn);
}
} finally {
// remove serializer name
removeCurrentSerializer();
}
}
// If undo_log exists, it means that the branch transaction has completed the first phase,
// we can directly roll back and clean the undo_log
// Otherwise, it indicates that there is an exception in the branch transaction,
// causing undo_log not to be written to the database.
// For example, the business processing timeout, the global transaction is the initiator rolls back.
// To ensure data consistency, we can insert an undo_log with GlobalFinished state
// to prevent the local transaction of the first phase of other programs from being correctly submitted.
// See https://github.com/seata/seata/issues/489
if (exists) {
deleteUndoLog(xid, branchId, conn);
conn.commit();
if (LOGGER.isInfoEnabled()) {
LOGGER.info("xid {} branch {}, undo_log deleted with {}", xid, branchId,
State.GlobalFinished.name());
}
} else {
insertUndoLogWithGlobalFinished(xid, branchId, UndoLogParserFactory.getInstance(), conn);
conn.commit();
if (LOGGER.isInfoEnabled()) {
LOGGER.info("xid {} branch {}, undo_log added with {}", xid, branchId,
State.GlobalFinished.name());
}
}
return;
} catch (SQLIntegrityConstraintViolationException e) {
// Possible undo_log has been inserted into the database by other processes, retrying rollback undo_log
if (LOGGER.isInfoEnabled()) {
LOGGER.info("xid {} branch {}, undo_log inserted, retry rollback", xid, branchId);
}
} catch (Throwable e) {
if (conn != null) {
try {
conn.rollback();
} catch (SQLException rollbackEx) {
LOGGER.warn("Failed to close JDBC resource while undo ... ", rollbackEx);
}
}
throw new BranchTransactionException(BranchRollbackFailed_Retriable, String
.format("Branch session rollback failed and try again later xid = %s branchId = %s %s", xid,
branchId, e.getMessage()), e);
} finally {
try {
if (rs != null) {
rs.close();
}
if (selectPST != null) {
selectPST.close();
}
if (conn != null) {
if (originalAutoCommit) {
conn.setAutoCommit(true);
}
conn.close();
}
} catch (SQLException closeEx) {
LOGGER.warn("Failed to close JDBC resource while undo ... ", closeEx);
}
}
}
}
上述可以为以下几个步骤：
1、获取普通数据库连接，查询该分支事务中的的undolog记录，
2、判断当前undolog记录是否为可回滚记录，防止重复回滚；
3、调用UndoLogParser的decode方法解析undolog记录，并执行回滚操作；
4、删除undolog记录并统一提交。
整个AT模式两阶段完成
案例源码：
seata源码：
官方资料：
