云服务器Redis集群部署及客户端通过公网IP连接问题

谁践踏了优雅 2022-11-27 07:14 322阅读 0赞

### 目录 ###

*  1、配置文件
 *  2、启动服务并创建集群
 *   *  (1)启动6个Redis服务
     *  (2)通过客户端命令创建集群
 *  3、客户端连接
 *   *  (1)客户端配置
     *  (2)测试用例
     *  (3)错误日志分析
 *  4、问题解决
 *   *  (1)查redis.conf配置文件
     *  (2)修改配置文件
     *  (3)重新启动Redis服务并创建集群
 *  5、故障转移期间Lettuce客户端连接问题
 *   *  (1)测试用例
     *  (2)停掉其中一个master节点，模拟宕机
     *  (3)解决办法
     *   *  1）更换Redis客户端
         *  2）Lettuce客户端配置Redis集群拓扑刷新

# 1、配置文件 #

准备了6个配置文件：redis-6381.conf,redis-6382.conf,redis-6383.conf,redis-6384.conf,redis-6385.conf,  
redis-6386.conf。配置文件内容如下：

# 配置文件进行了精简，完整配置可自行和官方提供的完整conf文件进行对照。端口号自行对应修改
    #后台启动的意思
    daemonize yes 
    #端口号
    port 6381
    # IP绑定，redis不建议对公网开放，这里绑定了服务器私网IP及环回地址
    bind 172.17.0.13 127.0.0.1
    # redis数据文件存放的目录
    dir /redis/workingDir
    # 日志文件
    logfile "/redis/logs/cluster-node-6381.log"
    # 开启AOF
    appendonly yes
     # 开启集群
    cluster-enabled yes
    # 集群持久化配置文件，内容包含其它节点的状态，持久化变量等，会自动生成在上面配置的dir目录下
    cluster-config-file cluster-node-6381.conf
    # 集群节点不可用的最大时间(毫秒)，如果主节点在指定时间内不可达，那么会进行故障转移
    cluster-node-timeout 5000

备注：Redis版本为6.0.4。

# 2、启动服务并创建集群 #

## (1)启动6个Redis服务 ##

> `redis-server redis-6381.conf`  
> `redis-server redis-6382.conf`  
> `redis-server redis-6383.conf`  
> `redis-server redis-6384.conf`  
> `redis-server redis-6385.conf`  
> `redis-server redis-6386.conf`

## (2)通过客户端命令创建集群 ##

创建集群，每个master节点分配一个从节点：

> `redis-cli --cluster create \`  
> `172.17.0.13:6381 172.17.0.13:6382 172.17.0.13:6383 \`  
> `172.17.0.13:6384 172.17.0.13:6385 172.17.0.13:6386 \`  
> `--cluster-replicas 1`

# 3、客户端连接 #

## (1)客户端配置 ##

@Configuration
    public class RedisClusterConfig { 
    
    	@Bean
    	public RedisConnectionFactory redisConnectionFactory() { 
    		// 客户端读写分离配置
    		LettuceClientConfiguration clientConfig = LettuceClientConfiguration.builder()
    						.readFrom(ReadFrom.REPLICA_PREFERRED)
    						.build();
    		RedisClusterConfiguration redisClusterConfiguration = new RedisClusterConfiguration(Arrays.asList(
    						"122.51.151.130:6381",
    						"122.51.151.130:6382",
    						"122.51.151.130:6383",
    						"122.51.151.130:6384",
    						"122.51.151.130:6385",
    						"122.51.151.130:6386"));
    		return new LettuceConnectionFactory(redisClusterConfiguration, clientConfig);
    	}
    }

## (2)测试用例 ##

@RunWith(SpringRunner.class)
    @SpringBootTest(classes = Application.class)
    public class RedisClusterTest { 
    
    	@Autowired
    	private StringRedisTemplate stringRedisTemplate;
    
    	@Test
    	public void readFromReplicaWriteToMasterTest() { 
    		System.out.println("开始设置值...");
    		stringRedisTemplate.opsForValue().set("username", "Nick");
    		System.out.println("获取值：" + stringRedisTemplate.opsForValue().get("username"));
    	}
    }

## (3)错误日志分析 ##

2020-08-14 14:57:49.180  WARN 22012 --- [ioEventLoop-6-4] i.l.c.c.topology.ClusterTopologyRefresh  : Unable to connect to [172.17.0.13:6384]: connection timed out: /172.17.0.13:6384
    2020-08-14 14:57:49.180  WARN 22012 --- [ioEventLoop-6-3] i.l.c.c.topology.ClusterTopologyRefresh  : Unable to connect to [172.17.0.13:6383]: connection timed out: /172.17.0.13:6383
    2020-08-14 14:57:49.182  WARN 22012 --- [ioEventLoop-6-2] i.l.c.c.topology.ClusterTopologyRefresh  : Unable to connect to [172.17.0.13:6382]: connection timed out: /172.17.0.13:6382
    2020-08-14 14:57:49.182  WARN 22012 --- [ioEventLoop-6-1] i.l.c.c.topology.ClusterTopologyRefresh  : Unable to connect to [172.17.0.13:6381]: connection timed out: /172.17.0.13:6381
    2020-08-14 14:57:49.190  WARN 22012 --- [ioEventLoop-6-1] i.l.c.c.topology.ClusterTopologyRefresh  : Unable to connect to [172.17.0.13:6385]: connection timed out: /172.17.0.13:6385
    2020-08-14 14:57:49.191  WARN 22012 --- [ioEventLoop-6-2] i.l.c.c.topology.ClusterTopologyRefresh  : Unable to connect to [172.17.0.13:6386]: connection timed out: /172.17.0.13:6386
    2020-08-14 14:57:59.389  WARN 22012 --- [ioEventLoop-6-3] i.l.core.cluster.RedisClusterClient      : connection timed out: /172.17.0.13:6382
    2020-08-14 14:58:09.391  WARN 22012 --- [ioEventLoop-6-4] i.l.core.cluster.RedisClusterClient      : connection timed out: /172.17.0.13:6381
    2020-08-14 14:58:19.393  WARN 22012 --- [ioEventLoop-6-1] i.l.core.cluster.RedisClusterClient      : connection timed out: /172.17.0.13:6383
    2020-08-14 14:58:29.396  WARN 22012 --- [ioEventLoop-6-2] i.l.core.cluster.RedisClusterClient      : connection timed out: /172.17.0.13:6384
    2020-08-14 14:58:39.399  WARN 22012 --- [ioEventLoop-6-3] i.l.core.cluster.RedisClusterClient      : connection timed out: /172.17.0.13:6386
    2020-08-14 14:58:49.402  WARN 22012 --- [ioEventLoop-6-4] i.l.core.cluster.RedisClusterClient      : connection timed out: /172.17.0.13:6385

连接客户端我们用的是Lettuce，这里发现指定的公网ip竟然变成私网ip了，客户端获取的IP地址信息是从Redis集群获取的，所以我们得让集群返回给我们公网ip。

# 4、问题解决 #

## (1)查redis.conf配置文件 ##

让Redis暴露公网IP其实在redis.conf配置文件里是能找到的，下面这段配置主要针对docker这种特殊的部署，这里我们也可以手动指定Redis的公网IP、端口以及总线端口(默认服务端口加10000)。

########################## CLUSTER DOCKER/NAT support  ########################
    
    # In certain deployments, Redis Cluster nodes address discovery fails, because
    # addresses are NAT-ted or because ports are forwarded (the typical case is
    # Docker and other containers).
    #
    # In order to make Redis Cluster working in such environments, a static
    # configuration where each node knows its public address is needed. The
    # following two options are used for this scope, and are:
    #
    # * cluster-announce-ip
    # * cluster-announce-port
    # * cluster-announce-bus-port
    #
    # Each instruct the node about its address, client port, and cluster message
    # bus port. The information is then published in the header of the bus packets
    # so that other nodes will be able to correctly map the address of the node
    # publishing the information.
    #
    # If the above options are not used, the normal Redis Cluster auto-detection
    # will be used instead.
    #
    # Note that when remapped, the bus port may not be at the fixed offset of
    # clients port + 10000, so you can specify any port and bus-port depending
    # on how they get remapped. If the bus-port is not set, a fixed offset of
    # 10000 will be used as usually.
    #
    # Example:
    #
    # cluster-announce-ip 10.1.1.5
    # cluster-announce-port 6379
    # cluster-announce-bus-port 6380

## (2)修改配置文件 ##

手动指定了公网ip后，Redis集群中的节点会通过公网IP进行通信，也就是外网访问。因此相关的总线端口，如下面的16381等总线端口必须在云服务器中的安全组中放开，不然集群会处于fail状态。

## (3)重新启动Redis服务并创建集群 ##

这个时候我们可以查看一下节点配置文件cluster-node-6381.conf的内容前后有啥变化。

未指定公网IP前：

[universe@VM_0_13_centos workingDir]$ cat cluster-node-6381.conf 
    34287d78c1e9c4ff49880bb976707a0c17676f82 172.17.0.13:6384@16384 slave 1a206270f835a79e43e281df5f6f8215ab49d713 0 1597390563209 4 connected
    e306ae5e3ead5f2a837d3bdc0b95c0bd8e3cff99 172.17.0.13:6383@16383 master - 0 1597390565212 3 connected 10923-16383
    0932cc203a19f37a3f5ebca8278962f5b325c67e 172.17.0.13:6385@16385 slave 2cc1aed536ff5b48c2fdd94f16cd96cefc4fd4ef 0 1597390564711 5 connected
    2cc1aed536ff5b48c2fdd94f16cd96cefc4fd4ef 172.17.0.13:6382@16382 master - 0 1597390565000 2 connected 5461-10922
    1a206270f835a79e43e281df5f6f8215ab49d713 172.17.0.13:6381@16381 myself,master - 0 1597390564000 1 connected 0-5460
    0f63accb455594d0625cffa8d09aacc580d7e428 172.17.0.13:6386@16386 slave e306ae5e3ead5f2a837d3bdc0b95c0bd8e3cff99 0 1597390564210 6 connected

指定公网IP后：

[universe@VM_0_13_centos workingDir]$ cat cluster-node-6381.conf 
    e2691ffd4bf7d867bc91b3b91c7b233a5f1e5dd2 122.51.151.130:6384@16384 master - 0 1597389992286 7 connected 10923-16383
    511668874d39a7b1f701cc3df6f21d00510bfeae 122.51.151.130:6383@16383 slave e2691ffd4bf7d867bc91b3b91c7b233a5f1e5dd2 0 1597389991283 7 connected
    e77e540ef4115abe920fb191f354b81f42e7b4ed 122.51.151.130:6381@16381 myself,master - 0 1597389991000 1 connected 0-5460
    2a3ea359311b34cd59e10da7d2f1bba48403f0ee 122.51.151.130:6385@16385 slave e77e540ef4115abe920fb191f354b81f42e7b4ed 0 1597389990583 5 connected
    2bf4f01a4dba802eb1a50d9510947a4af0ac92ef 122.51.151.130:6382@16382 master - 0 1597389992789 2 connected 5461-10922
    2b7671e002143b329c9c6c969bfb825a86fb41b2 122.51.151.130:6386@16386 slave 2bf4f01a4dba802eb1a50d9510947a4af0ac92ef 0 1597389991784 6 connected
    vars currentEpoch 7 lastVoteEpoch 7

这里我们可以发现，各节点暴露的IP全是公网IP了，再次运行测试用例，一切正常。

# 5、故障转移期间Lettuce客户端连接问题 #

## (1)测试用例 ##

@RunWith(SpringRunner.class)
    @SpringBootTest(classes = Application.class)
    public class RedisClusterTest { 
    
    	@Autowired
    	private StringRedisTemplate stringRedisTemplate;
    
    	@Test
    	public void automaticFailoverTest() throws InterruptedException { 
    		int count = 0;
    		while (true) { 
    			try { 
    				stringRedisTemplate.opsForValue().set("count", String.valueOf(++count));
    				System.out.println("修改count的值：" + count);
    				System.out.println("获取count的值：" + stringRedisTemplate.opsForValue().get("count"));
    				Thread.sleep(2000);
    			} catch (Exception e) { 
    				System.out.println("可能发生切主，重新操作...");
    				Thread.sleep(3000);
    			}
    		}
    	}
    }

## (2)停掉其中一个master节点，模拟宕机 ##

日志如下：

2020-08-20 19:33:25.118  INFO 13696 --- [xecutorLoop-1-1] i.l.core.protocol.ConnectionWatchdog     : Reconnecting, last destination was /122.51.151.130:6384
    2020-08-20 19:33:26.213  WARN 13696 --- [ioEventLoop-6-1] i.l.core.protocol.ConnectionWatchdog     : Cannot reconnect to [122.51.151.130:6384]: Connection refused: no further information: /122.51.151.130:6384
    2020-08-20 19:33:31.015  INFO 13696 --- [xecutorLoop-1-2] i.l.core.protocol.ConnectionWatchdog     : Reconnecting, last destination was 122.51.151.130:6384
    2020-08-20 19:33:32.107  WARN 13696 --- [ioEventLoop-6-2] i.l.core.protocol.ConnectionWatchdog     : Cannot reconnect to [122.51.151.130:6384]: Connection refused: no further information: /122.51.151.130:6384
    2020-08-20 19:33:36.616  INFO 13696 --- [xecutorLoop-1-2] i.l.core.protocol.ConnectionWatchdog     : Reconnecting, last destination was 122.51.151.130:6384
    2020-08-20 19:33:37.709  WARN 13696 --- [ioEventLoop-6-2] i.l.core.protocol.ConnectionWatchdog     : Cannot reconnect to [122.51.151.130:6384]: Connection refused: no further information: /122.51.151.130:6384
    2020-08-20 19:33:42.016  INFO 13696 --- [xecutorLoop-1-4] i.l.core.protocol.ConnectionWatchdog     : Reconnecting, last destination was 122.51.151.130:6384
    2020-08-20 19:33:43.110  WARN 13696 --- [ioEventLoop-6-4] i.l.core.protocol.ConnectionWatchdog     : Cannot reconnect to [122.51.151.130:6384]: Connection refused: no further information: /122.51.151.130:6384
    2020-08-20 19:33:47.216  INFO 13696 --- [xecutorLoop-1-1] i.l.core.protocol.ConnectionWatchdog     : Reconnecting, last destination was 122.51.151.130:6384
    2020-08-20 19:33:48.317  WARN 13696 --- [ioEventLoop-6-1] i.l.core.protocol.ConnectionWatchdog     : Cannot reconnect to [122.51.151.130:6384]: Connection refused: no further information: /122.51.151.130:6384
    2020-08-20 19:33:56.515  INFO 13696 --- [xecutorLoop-1-2] i.l.core.protocol.ConnectionWatchdog     : Reconnecting, last destination was 122.51.151.130:6384
    2020-08-20 19:33:57.605  WARN 13696 --- [ioEventLoop-6-2] i.l.core.protocol.ConnectionWatchdog     : Cannot reconnect to [122.51.151.130:6384]: Connection refused: no further information: /122.51.151.130:6384
    2020-08-20 19:34:14.016  INFO 13696 --- [xecutorLoop-1-3] i.l.core.protocol.ConnectionWatchdog     : Reconnecting, last destination was 122.51.151.130:6384
    2020-08-20 19:34:15.113  WARN 13696 --- [ioEventLoop-6-3] i.l.core.protocol.ConnectionWatchdog     : Cannot reconnect to [122.51.151.130:6384]: Connection refused: no further information: /122.51.151.130:6384
    可能发生切主，重新操作...
    2020-08-20 19:34:45.116  INFO 13696 --- [xecutorLoop-1-4] i.l.core.protocol.ConnectionWatchdog     : Reconnecting, last destination was 122.51.151.130:6384
    2020-08-20 19:34:46.212  WARN 13696 --- [ioEventLoop-6-4] i.l.core.protocol.ConnectionWatchdog     : Cannot reconnect to [122.51.151.130:6384]: Connection refused: no further information: /122.51.151.130:6384
    2020-08-20 19:35:16.216  INFO 13696 --- [xecutorLoop-1-1] i.l.core.protocol.ConnectionWatchdog     : Reconnecting, last destination was 122.51.151.130:6384
    2020-08-20 19:35:17.310  WARN 13696 --- [ioEventLoop-6-1] i.l.core.protocol.ConnectionWatchdog     : Cannot reconnect to [122.51.151.130:6384]: Connection refused: no further information: /122.51.151.130:6384
    可能发生切主，重新操作...

等了很长一段时间发现，发现客户端一致处于重连状态，这Lettuce客户端可能有毒。

## (3)解决办法 ##

### 1）更换Redis客户端 ###

将客户端换为Jedis后，再次模拟主节点宕机，发现过段时间后客户端连接恢复正常了。

@Configuration
    public class RedisClusterConfig { 
    
    	@Bean
    	public RedisConnectionFactory redisConnectionFactory() { 
    		RedisClusterConfiguration redisClusterConfiguration = new RedisClusterConfiguration(Arrays.asList(
    						"122.51.151.130:6381",
    						"122.51.151.130:6382",
    						"122.51.151.130:6383",
    						"122.51.151.130:6384",
    						"122.51.151.130:6385",
    						"122.51.151.130:6386"));
    		return new JedisConnectionFactory(redisClusterConfiguration);
    	}
    }

### 2）Lettuce客户端配置Redis集群拓扑刷新 ###

难道Lettuce客户端不支持主从切换后客户端重连么，那是不可能的。我们在github上找到了关于lettuce关于Redis集群的一些信息，相关地址如下：  
[https://github.com/lettuce-io/lettuce-core/wiki/Redis-Cluster][https_github.com_lettuce-io_lettuce-core_wiki_Redis-Cluster]  
[https://github.com/lettuce-io/lettuce-core/wiki/Client-options\#cluster-specific-options][https_github.com_lettuce-io_lettuce-core_wiki_Client-options_cluster-specific-options]

接下来按照文档上的提示修改客户端配置：

@Configuration
    public class RedisClusterConfig { 
    
    	@Bean
    	public RedisConnectionFactory redisConnectionFactory() { 
    		// 开启自适应集群拓扑刷新和周期拓扑刷新，不开启相应槽位主节点挂掉会出现服务不可用，直到挂掉节点重新恢复
    		ClusterTopologyRefreshOptions clusterTopologyRefreshOptions =  ClusterTopologyRefreshOptions.builder()
    						.enableAllAdaptiveRefreshTriggers() // 开启自适应刷新,自适应刷新不开启,Redis集群变更时将会导致连接异常
    						.adaptiveRefreshTriggersTimeout(Duration.ofSeconds(30)) //自适应刷新超时时间(默认30秒)，默认关闭开启后时间为30秒
    						.enablePeriodicRefresh(Duration.ofSeconds(20))  // 默认关闭开启后时间为60秒 ClusterTopologyRefreshOptions.DEFAULT_REFRESH_PERIOD 60 .enablePeriodicRefresh(Duration.ofSeconds(2)) = .enablePeriodicRefresh().refreshPeriod(Duration.ofSeconds(2))
    						.build();
    		ClientOptions clientOptions = ClusterClientOptions.builder()
    						.topologyRefreshOptions(clusterTopologyRefreshOptions)
    						.build();
    		// 客户端读写分离配置
    		LettuceClientConfiguration clientConfig = LettuceClientConfiguration.builder()
    						.clientOptions(clientOptions)
    						.build();
    		RedisClusterConfiguration redisClusterConfiguration = new RedisClusterConfiguration(Arrays.asList(
    						"122.51.151.130:6381",
    						"122.51.151.130:6382",
    						"122.51.151.130:6383",
    						"122.51.151.130:6384",
    						"122.51.151.130:6385",
    						"122.51.151.130:6386"));
    		return new LettuceConnectionFactory(redisClusterConfiguration, clientConfig);
    	}
    }

修改完配置后，再次运行测试用例，模拟主节点宕机，客户端再次重连。

[https_github.com_lettuce-io_lettuce-core_wiki_Redis-Cluster]: https://github.com/lettuce-io/lettuce-core/wiki/Redis-Cluster
[https_github.com_lettuce-io_lettuce-core_wiki_Client-options_cluster-specific-options]: https://github.com/lettuce-io/lettuce-core/wiki/Client-options#cluster-specific-options