官方对consumer的定义如下：
Similar to previously mentioned producer group, consumers of the exactly same role are grouped together and named Consumer Group.
Consumer Group is a great concept with which achieving goals of load-balance and fault-tolerance, in terms of message consuming, is super easy.
Warning: consumer instances of a consumer group must have exactly the same topic subscription(s).
大意是消费者准确按照相同角色来分组，分组的目的是负载均衡和失败转移，并且警告同个分组中的消费者一定要订阅相同的topic。
在代码中看到consumerGroup变量定义注释如下，大意差不多。
Consumers of the same role is required to have exactly same subscriptions and consumerGroup to correctly achieve load balance. It's required and needs to be globally unique.

通过上面的定义，可以准确知道同一组的consumer一定要订阅相同的topic，那么问题来了，订阅的时候除了topic还有tags，这个tag会有影响吗？没有找到相关资料，我就自己做了测试。

测试目标：测试在相同的消费组中的消费者，订阅相同的topic时，tag不同会不会影响消费和负载均衡。
测试计划：
1.创建生产者发送100条消息；
2.创建消费者

3.观察消息消费情况和队列分配情况。

发消息

public class SimpleProducer {
    public static void sendSync() throws Exception {
        ClientConfig clientConfig=new ClientConfig();
        clientConfig.setNamesrvAddr("localhost:9876");
        MQClientInstance clientInstance=MQClientManager.getInstance().getAndCreateMQClientInstance(clientConfig);
        DefaultMQProducer producer = clientInstance.getDefaultMQProducer();
        producer.setProducerGroup("GroupA");
        producer.start();
        for (int i = 0; i < 100; i++) {
            Message msg = new Message("TopicA", "TagA", ("Hello mq" + i).getBytes(RemotingHelper.DEFAULT_CHARSET));
            SendResult sendResult = producer.send(msg);
            System.out.println("send " + i + " , result:" + sendResult.getMsgId());
        }
        producer.shutdown();
    }
}

消费者

public class SimpleConsumer {
    public static void pushConsume(final String instanceName, final String group, final String topic, final String tag) throws Exception {
        DefaultMQPushConsumer consumer = new DefaultMQPushConsumer(group);
        consumer.setNamesrvAddr("localhost:9876");
        consumer.setConsumeFromWhere(ConsumeFromWhere.CONSUME_FROM_FIRST_OFFSET);
        consumer.subscribe(topic, tag);
        consumer.setInstanceName(instanceName);
        consumer.registerMessageListener(new MessageListenerOrderly() {
            @Override
            public ConsumeOrderlyStatus consumeMessage(List<MessageExt> msgs, ConsumeOrderlyContext context) {
                for (MessageExt msg : msgs) {
                    System.out.println("[" + instanceName + "," + group + "," + topic + "," + tag + "] consume: " + new String(msg.getBody()));
                }
                return ConsumeOrderlyStatus.SUCCESS;
            }
        });
        consumer.start();
    }
}

测试

public static void main(String[] args) throws Exception {
        
        try {
            SimpleProducer.sendSync();
        } catch (Exception e) {
            e.printStackTrace();
        }

        Thread t2 = new Thread() {
            @Override
            public void run() {
                try {
                    SimpleConsumer.pushConsume("A", "GroupA", "TopicA", "TagA");
                } catch (Exception e) {
                    e.printStackTrace();
                }
            }
        };
        t2.start();

        Thread t3 = new Thread() {
            @Override
            public void run() {
                try {
                    SimpleConsumer.pushConsume("B", "GroupA", "TopicA", "TagB");
                } catch (Exception e) {
                    e.printStackTrace();
                }
            }
        };
        t3.start();


        t2.join();
        t3.join();
    }

结果：

....
send 97 , result:AC1100013B2118B4AAC2808264CC0061
send 98 , result:AC1100013B2118B4AAC2808264CD0062
send 99 , result:AC1100013B2118B4AAC2808264CE0063
[A,GroupA,TopicA,TagA] consume: Hello mq1
[A,GroupA,TopicA,TagA] consume: Hello mq2
[A,GroupA,TopicA,TagA] consume: Hello mq5
....

查看队列分布情况：
消费者A

./bin/mqadmin consumerStatus -n "localhost:9876" -g "GroupA" -i "172.17.0.1@A"
#Consumer MQ Detail#
#Topic                            #Broker Name                      #QID  #ProcessQueueInfo   
%RETRY%GroupA                     mo-x                              0     ProcessQueueInfo [commitOffset=0, cachedMsgMinOffset=0, cachedMsgMaxOffset=0, cachedMsgCount=0, cachedMsgSizeInMiB=0, transactionMsgMinOffset=0, transactionMsgMaxOffset=0, transactionMsgCount=0, locked=true, tryUnlockTimes=0, lastLockTimestamp=20180625231130503, droped=false, lastPullTimestamp=20180625231132580, lastConsumeTimestamp=20180625231129554]
TopicA                            mo-x                              0     ProcessQueueInfo [commitOffset=50, cachedMsgMinOffset=0, cachedMsgMaxOffset=0, cachedMsgCount=0, cachedMsgSizeInMiB=0, transactionMsgMinOffset=0, transactionMsgMaxOffset=0, transactionMsgCount=0, locked=true, tryUnlockTimes=0, lastLockTimestamp=20180625231130503, droped=true, lastPullTimestamp=20180625231132552, lastConsumeTimestamp=20180625231129547]
TopicA                            mo-x                              1     ProcessQueueInfo [commitOffset=50, cachedMsgMinOffset=0, cachedMsgMaxOffset=0, cachedMsgCount=0, cachedMsgSizeInMiB=0, transactionMsgMinOffset=0, transactionMsgMaxOffset=0, transactionMsgCount=0, locked=true, tryUnlockTimes=0, lastLockTimestamp=20180625231130503, droped=true, lastPullTimestamp=20180625231132572, lastConsumeTimestamp=20180625231129549]

消费者B

./bin/mqadmin consumerStatus -n "localhost:9876" -g "GroupA" -i "172.17.0.1@B"
#Consumer MQ Detail#
#Topic                            #Broker Name                      #QID  #ProcessQueueInfo   
TopicA                            mo-x                              2     ProcessQueueInfo [commitOffset=25, cachedMsgMinOffset=0, cachedMsgMaxOffset=0, cachedMsgCount=0, cachedMsgSizeInMiB=0, transactionMsgMinOffset=0, transactionMsgMaxOffset=0, transactionMsgCount=0, locked=true, tryUnlockTimes=0, lastLockTimestamp=20180625231150500, droped=false, lastPullTimestamp=20180625231209016, lastConsumeTimestamp=20180625231149571]
TopicA                            mo-x                              3     ProcessQueueInfo [commitOffset=25, cachedMsgMinOffset=0, cachedMsgMaxOffset=0, cachedMsgCount=0, cachedMsgSizeInMiB=0, transactionMsgMinOffset=0, transactionMsgMaxOffset=0, transactionMsgCount=0, locked=true, tryUnlockTimes=0, lastLockTimestamp=20180625231150500, droped=false, lastPullTimestamp=20180625231209016, lastConsumeTimestamp=20180625231149565]

mo@mo-x:~/rocket-mq$ ./bin/mqadmin consumerProgress -n localhost:9876 -g "GroupA"
Java HotSpot(TM) 64-Bit Server VM warning: ignoring option PermSize=128m; support was removed in 8.0
Java HotSpot(TM) 64-Bit Server VM warning: ignoring option MaxPermSize=128m; support was removed in 8.0
#Topic                            #Broker Name                      #QID  #Broker Offset        #Consumer Offset      #Client IP           #Diff                 #LastTime
%RETRY%GroupA                     mo-x                              0     0                     0                     172.17.0.1           0                     1970-01-01 08:00:00
TopicA                            mo-x                              0     25                    25                    172.17.0.1           0                     2018-06-25 23:11:29
TopicA                            mo-x                              1     25                    25                    172.17.0.1           0                     2018-06-25 23:11:29
TopicA                            mo-x                              2     25                    25                    172.17.0.1           0                     2018-06-25 23:11:29
TopicA                            mo-x                              3     25                    25                    172.17.0.1           0                     2018-06-25 23:11:29

测试结果总结：
1.生产者发送了100条TagA消息到TopicA
2.消费者A和消费者B都在GroupA中，都订阅TopicA
3.消费者A订阅TagA，消费者B订阅TagB
4.消费者A收到了部分消息
5.消费者A分配到了两个GroupA-TopicA的队列
6.消费者B分配到了两个GroupA-TopicA的队列

总结：
Tag对同组同Topic的消费者有影响，当存在不同Tag的时候，会导致消费混乱，比如TagA的消息被TagB的消费者消费了。