Fescar源码学习--资源管理者RM(服务提供方)

 之前我们已经在博客《分布式事务--Fescar》中了解学习到Fescar相关的架构,接下来我们分别用几篇博客分别来介绍一下Fescar的 TM、RM 和 TC之间的交互流程。

TM、RM和TC之间的交互流程图:

Fescar源码学习--资源管理者RM(服务提供方)_第1张图片

简单角色理解:

TC: Fesacr-server应用

TM:dubbo服务调用方

RM:dubbo服务提供方

一、RM 简介

1、简介

       在上一篇博客《Fescar源码学习--事物管理者TM(服务调用方)》中我们已经介绍了TM角色相关的处理操作,这篇博客我们通过示例和源码来分析一下在分布式事务架构中RM角色所承担的责任。

RM在启动到处理TM的RPC调用主要做了以下操作:

(1)启动并向TC(Fescar Server)注册RM服务

(2)接收TM方的 Dubbo RPC 调用

(3)代理数据库生成undo日志,在undo_log表中记录相关原始数据

(4)返回RPC调用结果

(5)TM根据RM返回的结果向TC发送commit或rollback操作,TC根据XId事物id将commit或rollback通知对应的RM,RM接收TC(Fescar Server)的commit和rollback指令,其中commit是异步清理undo_log表中的数据,rollback操作会根据undo_log表中的数据恢复数据(读已提交,可能会产生脏数据)。

2、示例

dubbo服务提供者:

public class StorageServiceImpl implements StorageService {

    private static final Logger LOGGER = LoggerFactory.getLogger(StorageService.class);

    private JdbcTemplate jdbcTemplate;

    public void setJdbcTemplate(JdbcTemplate jdbcTemplate) {
        this.jdbcTemplate = jdbcTemplate;
    }

    @Override
    public void deduct(String commodityCode, int count) {
        LOGGER.info("Storage Service Begin ... xid: " + RootContext.getXID());
        jdbcTemplate.update("update storage_tbl set count = count - ? where commodity_code = ?", new Object[] {count, commodityCode});
        LOGGER.info("Storage Service End ... ");

    }

    public static void main(String[] args) throws Throwable {
        ClassPathXmlApplicationContext context = new ClassPathXmlApplicationContext(new String[]{"dubbo-storage-service.xml"});
        context.getBean("service");
        JdbcTemplate jdbcTemplate = (JdbcTemplate) context.getBean("jdbcTemplate");
        jdbcTemplate.update("delete from storage_tbl where commodity_code = 'C00321'");
        jdbcTemplate.update("insert into storage_tbl(commodity_code, count) values ('C00321', 100)");
        new ApplicationKeeper(context).keep();
    }
}

xml配置:



    
        
        
        
        
        
        
        
        
        
        
        
        
        
        
        
        
        
        
    

    
        
    

    
        
    

    
    
    
    

    
        
    

    
        
        
    

二、执行流程

1、初始化

在工程启动时会初始化GlobalTransactionScanner类,在initClient方法中会初始化RM相关的操作。

private void initClient() {
		
		//RM服务与TC建立连接
        TMClient.init(applicationId, txServiceGroup);
        if (LOGGER.isInfoEnabled()) {
            LOGGER.info(
                "Transaction Manager Client is initialized. applicationId[" + applicationId + "] txServiceGroup["
                    + txServiceGroup + "]");
        }
        if ((AT_MODE & mode) > 0) {
			//初始化与TC连接,初始化线程AsyncWorker
            RMClientAT.init(applicationId, txServiceGroup);
            if (LOGGER.isInfoEnabled()) {
                LOGGER.info(
                    "Resource Manager for AT Client is initialized. applicationId[" + applicationId
                        + "] txServiceGroup["
                        + txServiceGroup + "]");
            }
        }
        
    }

2、RPC调用

TM(服务消费者)在调用RM(服务提供者)使用dubbo的rpc调用机制,会将全局事务xid传递到RM服务中,通过TX_XID键获取值

@Activate(group = { Constants.PROVIDER, Constants.CONSUMER }, order = 100)
public class TransactionPropagationFilter implements Filter {

    private static final Logger LOGGER = LoggerFactory.getLogger(TransactionPropagationFilter.class);

    @Override
    public Result invoke(Invoker invoker, Invocation invocation) throws RpcException {
        String xid = RootContext.getXID();
        String rpcXid = RpcContext.getContext().getAttachment(RootContext.KEY_XID);
        if (LOGGER.isDebugEnabled()) {
            LOGGER.debug("xid in RootContext[" + xid + "] xid in RpcContext[" + rpcXid + "]");
        }
        boolean bind = false;
        if (xid != null) {
            RpcContext.getContext().setAttachment(RootContext.KEY_XID, xid);
        } else {
            if (rpcXid != null) {
                RootContext.bind(rpcXid);
                bind = true;
                if (LOGGER.isDebugEnabled()) {
                    LOGGER.debug("bind[" + rpcXid + "] to RootContext");
                }
            }
        }
        try {
            return invoker.invoke(invocation);

        } finally {
            if (bind) {
                String unbindXid = RootContext.unbind();
                if (LOGGER.isDebugEnabled()) {
                    LOGGER.debug("unbind[" + unbindXid + "] from RootContext");
                }
                if (!rpcXid.equalsIgnoreCase(unbindXid)) {
                    LOGGER.warn("xid in change during RPC from " + rpcXid + " to " + unbindXid);
                    if (unbindXid != null) {
                        RootContext.bind(unbindXid);
                        LOGGER.warn("bind [" + unbindXid + "] back to RootContext");
                    }
                }
            }
        }
    }
}

3、方法调用

在经过第二步dubbo的Filter拦截获取TX_XID之后,设置为本地线程变量,接下来就是执行本地函数。

    @Override
    public void deduct(String commodityCode, int count) {
        LOGGER.info("Storage Service Begin ... xid: " + RootContext.getXID());
        jdbcTemplate.update("update storage_tbl set count = count - ? where commodity_code = ?", new Object[] {count, commodityCode});
        LOGGER.info("Storage Service End ... ");

    }
jdbcTemplate在执行SQL操作时最终是交由ConnectionProxy的commit方法处理,在将事务提交到数据库之前做了一些处理操作。

4、分片事务提交

在ConnectionProxy的commit方法中做了以下操作:

(1)向TC(事务协调器,Fescar Server)注册分片事务,并获取分片事务id

(2)根据分片事务id在undo_log表中生成恢复日志

(3)真正提交数据库事务

(4)事务如果提交失败则向TC报告失败

(5)事务如果提交成功则向TC报告成功

    @Override
    public void commit() throws SQLException {
        if (context.inGlobalTransaction()) {
            try {
                //向TC注册并获取分片事务id
                register();
            } catch (TransactionException e) {
                recognizeLockKeyConflictException(e);
            }

            try {
                //生成恢复日志到undo_log表中
                if (context.hasUndoLog()) {
                    UndoLogManager.flushUndoLogs(this);
                }
                //真正提交数据库事事务
                targetConnection.commit();
            } catch (Throwable ex) {
                //向TC提交事务执行失败
                report(false);
                if (ex instanceof SQLException) {
                    throw (SQLException) ex;
                } else {
                    throw new SQLException(ex);
                }

            }
            //向TC提交事务执行成功
            report(true);
            context.reset();

        } else {
            targetConnection.commit();
        }
    }

在执行完commit操作之后,分片执行成功或失败在TC中都存在记录,并且数据库中原有的数据也会生成恢复日志保存到undo_log中,以便全局事务回滚时将数据还原。

5、事务commit或rollback

在上一篇博客《Fescar源码学习--事物管理者TM(服务调用方)》中我们已经学习到全局事务的提交或者回滚由TM(服务调用方)发送指令到TC,TC会根据全局事务XID将分片事务commit或rollback的操作通知到每个RM。

RM分片事务操作:

(1)commit:RM直接删除本地undo_log的记录即可

(2)rollback:RM根据本地undo_log表中的记录还原数据,可能会产生脏数据(undo_log中记录的是某个时间点的原始数据,可能和当前数据已经不一致了)。

RM提供了RmMessageListener用于建立TC发送过来的commit或rollback操作指令

    @Override
    public void onMessage(long msgId, String serverAddress, Object msg, ClientMessageSender sender) {
        if (LOGGER.isInfoEnabled()) {
            LOGGER.info("onMessage:" + msg);
        }
        //commit 操作
        if (msg instanceof BranchCommitRequest) {
            handleBranchCommit(msgId, serverAddress, (BranchCommitRequest)msg, sender);
        //rollback 操作
        } else if (msg instanceof BranchRollbackRequest) {
            handleBranchRollback(msgId, serverAddress, (BranchRollbackRequest)msg, sender);
        }
    }

6、commit操作

对于commit的操作RM的处理是简单的,RM只需要保证能将undo_log表中相关的记录删除即可,不需要过多的处理操作,因此commit请求最终会提交给AsyncWorker,由线程定时异步删除记录即可。

(1)xid最终会记录到map中

(2)在AsyncWorker初始化时建立定时任务每秒执行一次doBranchCommits函数,删除undo_log表中的记录

(3)在doBranchCommits中根据全局事务xid和分片事务branchId调用UndoLogManager.deleteUndoLog删除记录。

DataSourceManager中提交到 AsyncWorker异步操作:

    @Override
    public BranchStatus branchCommit(String xid, long branchId, String resourceId, String applicationData) throws TransactionException {
        return asyncWorker.branchCommit(xid, branchId, resourceId, applicationData);
    }

 在AsyncWorker中异常删除记录操作:

    @Override
    public BranchStatus branchCommit(String xid, long branchId, String resourceId, String applicationData) throws TransactionException {
        //提交事务ID到map中记录即可
        if (ASYNC_COMMIT_BUFFER.size() < ASYNC_COMMIT_BUFFER_LIMIT) {
            ASYNC_COMMIT_BUFFER.add(new Phase2Context(xid, branchId, resourceId, applicationData));
        } else {
            LOGGER.warn("Async commit buffer is FULL. Rejected branch [" + branchId + "/" + xid + "] will be handled by housekeeping later.");
        }
        return BranchStatus.PhaseTwo_Committed;
    }

    public synchronized void init() {
        LOGGER.info("Async Commit Buffer Limit: " + ASYNC_COMMIT_BUFFER_LIMIT);
        timerExecutor = new ScheduledThreadPoolExecutor(1,
            new NamedThreadFactory("AsyncWorker", 1, true));
        //定时任务,定时处理commit操作
        timerExecutor.scheduleAtFixedRate(new Runnable() {
            @Override
            public void run() {
                try {

                    doBranchCommits();


                } catch (Throwable e) {
                    LOGGER.info("Failed at async committing ... " + e.getMessage());

                }
            }
        }, 10, 1000 * 1, TimeUnit.MILLISECONDS);
    }

    //调用 UndoLogManager.deleteUndoLog 删除记录即可
    private void doBranchCommits() {
        if (ASYNC_COMMIT_BUFFER.size() == 0) {
            return;
        }
        Map> mappedContexts = new HashMap<>();
        Iterator iterator = ASYNC_COMMIT_BUFFER.iterator();
        while (iterator.hasNext()) {
            Phase2Context commitContext = iterator.next();
            List contextsGroupedByResourceId = mappedContexts.get(commitContext.resourceId);
            if (contextsGroupedByResourceId == null) {
                contextsGroupedByResourceId = new ArrayList<>();
                mappedContexts.put(commitContext.resourceId, contextsGroupedByResourceId);
            }
            contextsGroupedByResourceId.add(commitContext);

            iterator.remove();

        }

        for (String resourceId : mappedContexts.keySet()) {
            Connection conn = null;
            try {
                try {
                    DataSourceProxy dataSourceProxy = DataSourceManager.get().get(resourceId);
                    conn = dataSourceProxy.getPlainConnection();
                } catch (SQLException sqle) {
                    LOGGER.warn("Failed to get connection for async committing on " + resourceId, sqle);
                    continue;
                }

                List contextsGroupedByResourceId = mappedContexts.get(resourceId);
                for (Phase2Context commitContext : contextsGroupedByResourceId) {
                    try {
                        UndoLogManager.deleteUndoLog(commitContext.xid, commitContext.branchId, conn);
                    } catch (Exception ex) {
                        LOGGER.warn("Failed to delete undo log [" + commitContext.branchId + "/" + commitContext.xid + "]", ex);
                    }
                }

            } finally {
                if (conn != null) {
                    try {
                        conn.close();
                    } catch (SQLException closeEx) {
                        LOGGER.warn("Failed to close JDBC resource while deleting undo_log ", closeEx);
                    }
                }
            }


        }


    }

7、rollback操作

在DataSourceManager中进行事务回滚操作

    @Override
    public BranchStatus branchRollback(String xid, long branchId, String resourceId, String applicationData) throws TransactionException {
        DataSourceProxy dataSourceProxy = get(resourceId);
        if (dataSourceProxy == null) {
            throw new ShouldNeverHappenException();
        }
        try {
            //根据表undo_log中的快照信息回滚数据
            UndoLogManager.undo(dataSourceProxy, xid, branchId);
        } catch (TransactionException te) {
            if (te.getCode() == TransactionExceptionCode.BranchRollbackFailed_Unretriable) {
                return BranchStatus.PhaseTwo_RollbackFailed_Unretriable;
            } else {
                return BranchStatus.PhaseTwo_RollbackFailed_Retriable;
            }
        }
        //返回回滚结果信息
        return BranchStatus.PhaseTwo_Rollbacked;

    }

在UndoLogManager中首先根据全局事务XID和分片事务branchId获取结果信息,其中在rollback_info中以二进制方式保存了快照信息。

分片事务回滚: 

public static void undo(DataSourceProxy dataSourceProxy, String xid, long branchId) throws TransactionException {
        assertDbSupport(dataSourceProxy.getTargetDataSource().getDbType());

        Connection conn = null;
        ResultSet rs = null;
        PreparedStatement selectPST = null;
        try {
            conn = dataSourceProxy.getPlainConnection();

            // The entire undo process should run in a local transaction.
            conn.setAutoCommit(false);

            // Find UNDO LOG
            //根据xid和branchId查找数据
            selectPST = conn.prepareStatement(SELECT_UNDO_LOG_SQL);
            selectPST.setLong(1, branchId);
            selectPST.setString(2, xid);
            rs = selectPST.executeQuery();
            
            while (rs.next()) {
                Blob b = rs.getBlob("rollback_info");
                //获取rollback_info字段数据
                String rollbackInfo = StringUtils.blob2string(b);
                //生成回滚sql
                BranchUndoLog branchUndoLog = UndoLogParserFactory.getInstance().decode(rollbackInfo);

                //执行回滚操作
                for (SQLUndoLog sqlUndoLog : branchUndoLog.getSqlUndoLogs()) {
                    TableMeta tableMeta = TableMetaCache.getTableMeta(dataSourceProxy, sqlUndoLog.getTableName());
                    sqlUndoLog.setTableMeta(tableMeta);
                    AbstractUndoExecutor undoExecutor = UndoExecutorFactory.getUndoExecutor(dataSourceProxy.getDbType(), sqlUndoLog);
                    undoExecutor.executeOn(conn);
                }

            }
            //删除记录
            deleteUndoLog(xid, branchId, conn);
            //提交事务
            conn.commit();

        } catch (Throwable e) {
            if (conn != null) {
                try {
                    conn.rollback();
                } catch (SQLException rollbackEx) {
                    LOGGER.warn("Failed to close JDBC resource while undo ... ", rollbackEx);
                }
            }
            throw new TransactionException(BranchRollbackFailed_Retriable, String.format("%s/%s", branchId, xid), e);

        } finally {
            try {
                if (rs != null) {
                    rs.close();
                }
                if (selectPST != null) {
                    selectPST.close();
                }
                if (conn != null) {
                    conn.close();
                }
            } catch (SQLException closeEx) {
                LOGGER.warn("Failed to close JDBC resource while undo ... ", closeEx);
            }
        }

    }

总结:

(1)Fescar通过dubbo远程调用传递全局事务id

(2)RM在执行本地数据库操作时首先会向TC申请分片事务

(3)根据分片事务id在undo_log生成回滚日志

(4)执行本地数据库操作,成功或失败都会向TC进行报告

(5)TM进行事务commit或rollback操作,将操作提交到TC,TC将操作发送到RM

(6)RM接收到TC的通知进行commit或rollback操作

(7)如果是commit操作则异步通过线程AsyncWorker进行删除本地undo_log即可

(8)如果是rollback操作则根据全局事务xid和分片事务branchId找到记录,根据字段rollback_info中的快照信息进行回滚数据操作。

你可能感兴趣的:(数据库及分布式事务)