学习淘淘商城第三十二课（Redis集群测试）

      上节课我们一起学习了怎样搭建一个六台设备的Redis集群，这节我们一起学习下如何测试集群。

      首先启动Redis集群使用命令/usr/local/redis/bin/redis-trib.rb create --replicas 1 192.168.156.15:6379 192.168.156.16:6379 192.168.156.17:6379 192.168.156.18:6379 192.168.156.19:6379 192.168.156.20:6379进行启动，如下所示。可以看到192.168.156.18、192.168.156.19、192.168.156.20这三个节点是主节点，另外三个节点是从节点。槽号为0-5460的卡槽被分配到了192.168.156.20上，5461-10922被分配到了192.168.156.19上，10923-16383被分配到了192.168.156.18上。

[root@redis1 ~]# /usr/local/redis/bin/redis-trib.rb create --replicas 1 192.168.156.15:6379 192.168.156.16:6379 192.168.156.17:6379 192.168.156.18:6379 192.168.156.19:6379 192.168.156.20:6379
>>> Creating cluster
Connecting to node 192.168.156.15:6379: OK
Connecting to node 192.168.156.16:6379: OK
Connecting to node 192.168.156.17:6379: OK
Connecting to node 192.168.156.18:6379: OK
Connecting to node 192.168.156.19:6379: OK
Connecting to node 192.168.156.20:6379: OK
>>> Performing hash slots allocation on 6 nodes...
Using 3 masters:
192.168.156.20:6379
192.168.156.19:6379
192.168.156.18:6379
Adding replica 192.168.156.17:6379 to 192.168.156.20:6379
Adding replica 192.168.156.16:6379 to 192.168.156.19:6379
Adding replica 192.168.156.15:6379 to 192.168.156.18:6379
S: 0cbe6fc270afa701156c5dc9c47dba522dd62737 192.168.156.15:6379
   replicates 423f334654c3dfdacd39927e0528a95c38aa933f
S: 4eecf66148e88917b4ecf4b36e794a6e83e4fe85 192.168.156.16:6379
   replicates b69de3b87aeefac1d4711dd644a8dd5815c56bec
S: 149c9e9a0e0febd581193df17d509ce4970e4163 192.168.156.17:6379
   replicates 6808fa38709cac3a6ec4b232fd1e45980cbb0c96
M: 423f334654c3dfdacd39927e0528a95c38aa933f 192.168.156.18:6379
   slots:10923-16383 (5461 slots) master
M: b69de3b87aeefac1d4711dd644a8dd5815c56bec 192.168.156.19:6379
   slots:5461-10922 (5462 slots) master
M: 6808fa38709cac3a6ec4b232fd1e45980cbb0c96 192.168.156.20:6379
   slots:0-5460 (5461 slots) master
Can I set the above configuration? (type 'yes' to accept): yes
>>> Nodes configuration updated
>>> Assign a different config epoch to each node
>>> Sending CLUSTER MEET messages to join the cluster
Waiting for the cluster to join....
>>> Performing Cluster Check (using node 192.168.156.15:6379)
M: 0cbe6fc270afa701156c5dc9c47dba522dd62737 192.168.156.15:6379
   slots: (0 slots) master
   replicates 423f334654c3dfdacd39927e0528a95c38aa933f
M: 4eecf66148e88917b4ecf4b36e794a6e83e4fe85 192.168.156.16:6379
   slots: (0 slots) master
   replicates b69de3b87aeefac1d4711dd644a8dd5815c56bec
M: 149c9e9a0e0febd581193df17d509ce4970e4163 192.168.156.17:6379
   slots: (0 slots) master
   replicates 6808fa38709cac3a6ec4b232fd1e45980cbb0c96
M: 423f334654c3dfdacd39927e0528a95c38aa933f 192.168.156.18:6379
   slots:10923-16383 (5461 slots) master
M: b69de3b87aeefac1d4711dd644a8dd5815c56bec 192.168.156.19:6379
   slots:5461-10922 (5462 slots) master
M: 6808fa38709cac3a6ec4b232fd1e45980cbb0c96 192.168.156.20:6379
   slots:0-5460 (5461 slots) master
[OK] All nodes agree about slots configuration.
>>> Check for open slots...
>>> Check slots coverage...

      既然集群已经启动好了，我们使用redis客户端随意连接一台设备进行测试，比如我们连接192.168.156.16这台设备（这是个从节点，从属于192.168.156.19），如下所示（注意：一定要加上参数"-c"否则没办法重定向 到其它节点）。

[root@redis2 bin]# ./redis-cli -h 192.168.156.16 -p 6379 -c
192.168.156.16:6379>

      连接上之后，我们先来保存key1，它的值是123，如下所示，可以看到，这个key1经过算法并对16384进行求余之后的数字是9189，这个槽号是在192.168.156.19上，因此key1被保存到了192.168.156.19上。

192.168.156.16:6379> set key1 123
-> Redirected to slot [9189] located at 192.168.156.19:6379
OK
192.168.156.19:6379>

      我们再接着保存key2、key3、key4，如下所示，可以看到key2所对应的槽号是4998，显然是在192.168.156.20上，key3所对应的槽号也在0-5640，因此set key2的时候根本就没有重定向，还是停留在192.168.156.20上。key4所对应的槽号是13120，显然是在192.168.156.18上因此被保存到了192.168.156.18上。

192.168.156.19:6379> set key2 111
-> Redirected to slot [4998] located at 192.168.156.20:6379
OK
192.168.156.20:6379> set key3 aaa
OK
192.168.156.20:6379> set key4 bbb
-> Redirected to slot [13120] located at 192.168.156.18:6379
OK
192.168.156.18:6379> 

       我们可以使用命令cluster info查看集群的某些信息，如下所示，可以看到当前集群的状态是OK的，所有的槽号都已分配完毕，没有失败的节点。所有的节点数是6，当前提供服务的有3台

192.168.156.18:6379> cluster info
cluster_state:ok
cluster_slots_assigned:16384
cluster_slots_ok:16384
cluster_slots_pfail:0
cluster_slots_fail:0
cluster_known_nodes:6
cluster_size:3
cluster_current_epoch:6
cluster_my_epoch:4
cluster_stats_messages_sent:10748
cluster_stats_messages_received:10748
192.168.156.18:6379>

        还可以通过命令cluster nodes来查看节点信息。

192.168.156.18:6379> cluster nodes
423f334654c3dfdacd39927e0528a95c38aa933f 192.168.156.18:6379 myself,master - 0 0 4 connected 10923-16383
4eecf66148e88917b4ecf4b36e794a6e83e4fe85 192.168.156.16:6379 slave b69de3b87aeefac1d4711dd644a8dd5815c56bec 0 1492869406355 5 connected
0cbe6fc270afa701156c5dc9c47dba522dd62737 192.168.156.15:6379 slave 423f334654c3dfdacd39927e0528a95c38aa933f 0 1492869405340 4 connected
149c9e9a0e0febd581193df17d509ce4970e4163 192.168.156.17:6379 slave 6808fa38709cac3a6ec4b232fd1e45980cbb0c96 0 1492869402275 6 connected
6808fa38709cac3a6ec4b232fd1e45980cbb0c96 192.168.156.20:6379 master - 0 1492869408383 6 connected 0-5460
b69de3b87aeefac1d4711dd644a8dd5815c56bec 192.168.156.19:6379 master - 0 1492869407370 5 connected 5461-10922
192.168.156.18:6379> 

      下面使用JedisCluster来测试集群，我们在taotao-content-service工程的测试类中再添加一个测试方法（与第三十课结合学习），如下所示。运行，发现是可以正常输出"hello jedis cluster"的。

@Test
	public void testJedisCluster(){
		//创建构造参数Set类型，集合中每个元素是HostAndPort类型
		Set<HostAndPort> nodes = new HashSet<>();
		//向集合中添加节点
		nodes.add(new HostAndPort("192.168.156.15", 6379));
		nodes.add(new HostAndPort("192.168.156.16", 6379));
		nodes.add(new HostAndPort("192.168.156.17", 6379));
		nodes.add(new HostAndPort("192.168.156.18", 6379));
		nodes.add(new HostAndPort("192.168.156.19", 6379));
		nodes.add(new HostAndPort("192.168.156.20", 6379));
		//创建JedisCluster对象
		JedisCluster jedisCluster = new JedisCluster(nodes);
		//直接使用jedisCluster，自带连接池，jedisCluster可以是单例的
		jedisCluster.set("jedis-cluster", "hello jedis cluster");
		String result = jedisCluster.get("jedis-cluster");
		System.out.println(result);
		//系统关闭前关闭jedisCluster
		jedisCluster.close();
	}