SpringDataRedis事务处理

序

本文主要讲述如何在java里头使用redis进行cas操作。其实呢，redis不像memcached那样显示地支持cas操作，不过它有事务的概念。

准备

• redis的docker搭建

• SpringBoot应用之分布式缓存

redis的乐观锁支持

Redis通过使用WATCH, MULTI, and EXEC组成的事务来实现乐观锁(注意没有用DISCARD),Redis事务没有回滚操作。在SpringDataRedis当中通过RedisTemplate的SessionCallback中来支持(否则事务不生效)。discard的话不需要自己代码处理，callback返回null，成的话，返回非null，依据这个来判断事务是否成功(没有抛异常)。

实例

@Test
    public void cas() throws InterruptedException, ExecutionException {
        String key = "test-cas-1";
        ValueOperations<String, String> strOps = redisTemplate.opsForValue();
        strOps.set(key, "hello");
        ExecutorService pool  = Executors.newCachedThreadPool();
        List<Callable<Object>> tasks = new ArrayList<>();
        for(int i=0;i<5;i++){
            final int idx = i;
            tasks.add(new Callable() {
                @Override
                public Object call() throws Exception {
                    return redisTemplate.execute(new SessionCallback() {
                        @Override
                        public Object execute(RedisOperations operations) throws DataAccessException {
                            operations.watch(key);
                            String origin = (String) operations.opsForValue().get(key);
                            operations.multi();
                            operations.opsForValue().set(key, origin + idx);
                            Object rs = operations.exec();
                            System.out.println("set:"+origin+idx+" rs:"+rs);
                            return rs;
                        }
                    });
                }
            });
        }
        List<Future<Object>> futures = pool.invokeAll(tasks);
        for(Future<Object> f:futures){
            System.out.println(f.get());
        }
        pool.shutdown();
        pool.awaitTermination(1000, TimeUnit.MILLISECONDS);
    }

输出

set:hello2 rs:null
set:hello3 rs:[]
set:hello1 rs:null
set:hello4 rs:null
set:hello0 rs:null

查看该值

127.0.0.1:6379> get test-cas-1
"\"hello3\""

坑

SessionCallback

没有在SessionCallback里头执行watch、multi、exec，而是自己单独写

与数据库事务的混淆

template.setEnableTransactionSupport(true);

这个应该是支持数据库的事务成功才执行的意思。

/**
     * Gets a Redis connection. Is aware of and will return any existing corresponding connections bound to the current
     * thread, for example when using a transaction manager. Will create a new Connection otherwise, if
     * {@code allowCreate} is <tt>true</tt>.
     * 
     * @param factory connection factory for creating the connection
     * @param allowCreate whether a new (unbound) connection should be created when no connection can be found for the
     *          current thread
     * @param bind binds the connection to the thread, in case one was created
     * @param enableTransactionSupport
     * @return an active Redis connection
     */
    public static RedisConnection doGetConnection(RedisConnectionFactory factory, boolean allowCreate, boolean bind,
            boolean enableTransactionSupport) {

        Assert.notNull(factory, "No RedisConnectionFactory specified");

        RedisConnectionHolder connHolder = (RedisConnectionHolder) TransactionSynchronizationManager.getResource(factory);

        if (connHolder != null) {
            if (enableTransactionSupport) {
                potentiallyRegisterTransactionSynchronisation(connHolder, factory);
            }
            return connHolder.getConnection();
        }

        if (!allowCreate) {
            throw new IllegalArgumentException("No connection found and allowCreate = false");
        }

        if (log.isDebugEnabled()) {
            log.debug("Opening RedisConnection");
        }

        RedisConnection conn = factory.getConnection();

        if (bind) {

            RedisConnection connectionToBind = conn;
            if (enableTransactionSupport && isActualNonReadonlyTransactionActive()) {
                connectionToBind = createConnectionProxy(conn, factory);
            }

            connHolder = new RedisConnectionHolder(connectionToBind);

            TransactionSynchronizationManager.bindResource(factory, connHolder);
            if (enableTransactionSupport) {
                potentiallyRegisterTransactionSynchronisation(connHolder, factory);
            }

            return connHolder.getConnection();
        }

        return conn;
    }

不要跟本文的乐观锁说的事务混淆在一起。

参考

• Redis-Transactions-via-Spring-Data-Redis

• Spring-data-redis 第二天（事务）