基于springsession构建一个session共享的模块。 这里，基于redis的集群（Redis-5.0.3版本），为了解决整个物联网平台的各个子系统之间共享session需求，且方便各个子系统使用，将这个模块设计为一个pom组件，直接在pom.xml文件里面配置为dependency。
今天的主题，就是redis 5.0.3环境的构建。
和我之前介绍过的redis 3.2.8 （https://www.cnblogs.com/shihuc/p/7882004.html）有些类似，只是redis 5里面，不再依赖ruby脚本进行集群配置，而是直接用c程序实现，直接基于redis-cli指令完成。这些就不做介绍，直接介绍如何配置。
首先，需要修改的配置项如下：
bind 10.95.200.12 protected-mode no port 7380 daemonize yes pidfile /var/run/redis_7380.pid dbfilename dump-7380.rdb appendonly yes appendfilename "appendonly-7380.aof" cluster-enabled yes cluster-config-file nodes-7380.conf cluster-node-timeout 15000 notify-keyspace-events "Ex"
我的配置环境，是3对主从，上面的配置项，是其中一个redis节点的信息，IP是10.95.200.12，端口为7380. 参照这个配置信息，将下面的几个节点都配置上，我这里一共有三台虚拟机，分别是IP：10.95.200.12,10.95.200.13, 10.95.200.14，每台机器上部署两个实例，分别为端口7380以及7381.
配置信息配置好后，需要将每一个实例启动起来。例如，下面启动10.95.200.12端口为7380的实例。
[tkiot@tkwh-kfcs-app2 redis]$ ./bin/redis-server redis-7380.conf 23941:C 10 Jul 2019 08:34:05.607 # oO0OoO0OoO0Oo Redis is starting oO0OoO0OoO0Oo 23941:C 10 Jul 2019 08:34:05.607 # Redis version=5.0.3, bits=64, commit=00000000, modified=0, pid=23941, just started 23941:C 10 Jul 2019 08:34:05.607 # Configuration loaded
这里有个小插曲：
每个服务器上的redis的配置文件，能够正常运行的配置文件名称如上面：redis-7380.conf，刚开始，我没有太注意，或者说是自己的马虎，将这个配置文件命名为nodes-7380.conf,和配置文件内容配置项cluster-config-file的值搞成一样的了，真是想骂娘的这个坑，启动过程同样显示上面的输出，和正确的程序启动一样的效果，但是呢，ps查看进程后，发现没有redis的进程。。。。这个坑，让我查了一上午。。。。
上面指令启动redis-7380.conf的操作结束后，可以查看下启动状态（cluster nodes）：
[tkiot@tkwh-kfcs-app2 redis]$ redis-cli -c -h 10.95.200.12 -p 7380 10.95.200.12:7380> cluster nodes 7f4cf1bffc7e42a0e2d15bcc5a0a5386711813e8 :7380@17380 myself,master - 0 0 0 connected
先看看cluster相关的指令都有哪些：
[tkiot@tkwh-kfcs-app2 redis]$ redis-cli --cluster help
Cluster Manager Commands:
create host1:port1 ... hostN:portN
--cluster-replicas <arg>
check host:port
--cluster-search-multiple-owners
info host:port
fix host:port
--cluster-search-multiple-owners
reshard host:port
--cluster-from <arg>
--cluster-to <arg>
--cluster-slots <arg>
--cluster-yes
--cluster-timeout <arg>
--cluster-pipeline <arg>
--cluster-replace
rebalance host:port
--cluster-weight <node1=w1...nodeN=wN>
--cluster-use-empty-masters
--cluster-timeout <arg>
--cluster-simulate
--cluster-pipeline <arg>
--cluster-threshold <arg>
--cluster-replace
add-node new_host:new_port existing_host:existing_port
--cluster-slave
--cluster-master-id <arg>
del-node host:port node_id
call host:port command arg arg .. arg
set-timeout host:port milliseconds
import host:port
--cluster-from <arg>
--cluster-copy
--cluster-replace
help
接下来，就是通过redis-cli创建集群，配置成为一个shell脚本
#!/bin/bash /u02/redis/bin/redis-cli --cluster create 10.95.200.12:7380 10.95.200.13:7380 10.95.200.14:7380 10.95.200.12:7381 10.95.200.13:7381 10.95.200.14:7381 --cluster-replicas 1
这里，脚本当中，一共6个节点，不能确定哪些是master，哪些是slave，主从关系，是在构建集群的时候，自动配置的。
集群查看节点常用指令:
CLUSTER INFO 打印集群的信息 CLUSTER NODES 列出集群当前已知的所有节点（node），以及这些节点的相关信息。
[tkiot@tkwh-kfcs-app2 redis]$ redis-cli -c -h 10.95.200.12 -p 7380 10.95.200.12:7380> 10.95.200.12:7380> cluster nodes ed309033dbefe2b0b64ad7fb643c4d2531e53b95 10.95.200.13:7380@17380 master - 0 1562720545267 2 connected 5461-10922 2cb8c6a7c86f512db2c6ca88eb67449197f5b885 10.95.200.12:7381@17381 slave cf6ca00cb36850762fdff1223684edf1fb9bd4ba 0 1562720546000 4 connected cf6ca00cb36850762fdff1223684edf1fb9bd4ba 10.95.200.14:7380@17380 master - 0 1562720546000 3 connected 10923-16383 26c3a7cd48fe4bddf9cf0c60a3c30b6b1f273135 10.95.200.12:7380@17380 myself,slave 467d4c7508d1cb371ed52c4c6574506cba40c328 0 1562720546000 1 connected 467d4c7508d1cb371ed52c4c6574506cba40c328 10.95.200.13:7381@17381 master - 0 1562720547271 8 connected 0-5460 fbece4571b50904d93a45afbcce66941d53a45b5 10.95.200.14:7381@17381 slave ed309033dbefe2b0b64ad7fb643c4d2531e53b95 0 1562720546269 6 connected
剔除掉一个指定的节点（通过nodeId进行指示要删除的节点，nodeId就是上面cluster nodes输出结果的最左边的那一列）
10.95.200.12:7380> cluster forget ed309033dbefe2b0b64ad7fb643c4d2531e53b95 OK 10.95.200.12:7380> 10.95.200.12:7380> cluster nodes 2cb8c6a7c86f512db2c6ca88eb67449197f5b885 10.95.200.12:7381@17381 slave cf6ca00cb36850762fdff1223684edf1fb9bd4ba 0 1562720579324 4 connected cf6ca00cb36850762fdff1223684edf1fb9bd4ba 10.95.200.14:7380@17380 master - 0 1562720578323 3 connected 10923-16383 26c3a7cd48fe4bddf9cf0c60a3c30b6b1f273135 10.95.200.12:7380@17380 myself,slave 467d4c7508d1cb371ed52c4c6574506cba40c328 0 1562720576000 1 connected 467d4c7508d1cb371ed52c4c6574506cba40c328 10.95.200.13:7381@17381 master - 0 1562720577320 8 connected 0-5460 fbece4571b50904d93a45afbcce66941d53a45b5 10.95.200.14:7381@17381 slave - 0 1562720577000 6 connected 10.95.200.12:7380>
不能自己把自己忘记哟：
10.95.200.12:7380> cluster forget 26c3a7cd48fe4bddf9cf0c60a3c30b6b1f273135 (error) ERR I tried hard but I can't forget myself... 10.95.200.12:7380>
也不能将自己的master忘记哟：
10.95.200.12:7380> cluster forget 467d4c7508d1cb371ed52c4c6574506cba40c328 (error) ERR Can't forget my master! 10.95.200.12:7380>
上面的forget指令是不能讲节点真实删除的，只是一段时间内forget遗忘而已。因为过一段时间后，再次执行cluster nodes指令，还是可以看到完整集群的节点。
删除集群节点，可以采用下面的指令：
redis-cli --cluster del-node <ip>:<port> <node_id>
下面看看我的操作，这里是一个错误的操作，结果如下：
[tkiot@tkwh-kfcs-app2 redis]$ redis-cli --cluster del-node 10.95.200.12:7380 ed309033dbefe2b0b64ad7fb643c4d2531e53b95 >>> Removing node ed309033dbefe2b0b64ad7fb643c4d2531e53b95 from cluster 10.95.200.12:7380 [ERR] Node 10.95.200.13:7380 is not empty! Reshard data away and try again.
1）查看Node 10.95.200.13:7380上是否有信息：
10.95.200.13:7380> keys * 1) "taikang#session:sessions:expires:b6cd8269-dff0-463e-aefb-03167a167292"
2）清除所有的内容：
10.95.200.13:7380> flushdb OK
到此，继续执行redis-cli --cluster del-node 10.95.200.12:7380 ed309033dbefe2b0b64ad7fb643c4d2531e53b95，最终还是失败，不成功的原因一样，不是空节点。为何？原因是没有依据redis的指令说明在使用del-node指令。
注意，删除节点，按照上面的操作，显然是不对的，提示Node不是空的，需要将数据重新分片到其他节点，这个提示，有点不是太好理解，依据redis的集群工作原理，是全对等（master之间）节点集群，删除节点有三个步骤：
A. 首先删除slave节点，命令：redis-cli --cluster del-node <ip>:<port> <node_id> B. 将被删除的slave对应的master节点的slot进行reshard到其他节点，命令：redis-cli --cluster reshard <master的ip:port> --cluster-from <同一个master的node_id> --cluster-to <接收slot的master的node_id> --cluster-slots <将要参与分片的slot数量，这里就是待删除节点的全部slot> --cluster-yes C. 将reshard后的master执行删除操作，命令：redis-cli --cluster del-node <ip>:<port> <node_id>
下面演示基本操作redis cluster
操作前的节点信息：
[tkiot@tkwh-kfcs-app2 redis]$ redis-cli -c -h 10.95.200.12 -p 7380 cluster nodes ed309033dbefe2b0b64ad7fb643c4d2531e53b95 10.95.200.13:7380@17380 master - 0 1563276949623 11 connected 0-5459 5461-16382 2cb8c6a7c86f512db2c6ca88eb67449197f5b885 10.95.200.12:7381@17381 slave cf6ca00cb36850762fdff1223684edf1fb9bd4ba 0 1563276951624 10 connected cf6ca00cb36850762fdff1223684edf1fb9bd4ba 10.95.200.14:7380@17380 master - 0 1563276950000 10 connected 16383 26c3a7cd48fe4bddf9cf0c60a3c30b6b1f273135 10.95.200.12:7380@17380 myself,slave 467d4c7508d1cb371ed52c4c6574506cba40c328 0 1563276950000 1 connected fbece4571b50904d93a45afbcce66941d53a45b5 10.95.200.14:7381@17381 slave ed309033dbefe2b0b64ad7fb643c4d2531e53b95 0 1563276950000 11 connected 467d4c7508d1cb371ed52c4c6574506cba40c328 10.95.200.13:7381@17381 master - 0 1563276950624 8 connected 5460
1. 删除节点：
1）先删除slave节点
[tkiot@tkwh-kfcs-app2 redis]$ redis-cli --cluster del-node 10.95.200.14:7381 fbece4571b50904d93a45afbcce66941d53a45b5 >>> Removing node fbece4571b50904d93a45afbcce66941d53a45b5 from cluster 10.95.200.14:7381 >>> Sending CLUSTER FORGET messages to the cluster... >>> SHUTDOWN the node.
slave节点删除后的结构：
[tkiot@tkwh-kfcs-app2 redis]$ redis-cli -c -h 10.95.200.12 -p 7380 cluster nodes ed309033dbefe2b0b64ad7fb643c4d2531e53b95 10.95.200.13:7380@17380 master - 0 1563277493737 11 connected 0-5459 5461-16382 2cb8c6a7c86f512db2c6ca88eb67449197f5b885 10.95.200.12:7381@17381 slave cf6ca00cb36850762fdff1223684edf1fb9bd4ba 0 1563277492000 10 connected cf6ca00cb36850762fdff1223684edf1fb9bd4ba 10.95.200.14:7380@17380 master - 0 1563277493000 10 connected 16383 26c3a7cd48fe4bddf9cf0c60a3c30b6b1f273135 10.95.200.12:7380@17380 myself,slave 467d4c7508d1cb371ed52c4c6574506cba40c328 0 1563277493000 1 connected 467d4c7508d1cb371ed52c4c6574506cba40c328 10.95.200.13:7381@17381 master - 0 1563277494738 8 connected 5460
2）再对刚才删除的slave的master进行slot的reshard：
[tkiot@tkwh-kfcs-app2 redis]$ redis-cli --cluster reshard 10.95.200.13:7381 --cluster-from 467d4c7508d1cb371ed52c4c6574506cba40c328 --cluster-to ed309033dbefe2b0b64ad7fb643c4d2531e53b95 --cluster-slots 5460 --cluster-yes >>> Performing Cluster Check (using node 10.95.200.13:7381)
M: 467d4c7508d1cb371ed52c4c6574506cba40c328 10.95.200.13:7381
slots:[5460] (1 slots) master
1 additional replica(s)
M: ed309033dbefe2b0b64ad7fb643c4d2531e53b95 10.95.200.13:7380
slots:[0-5459],[5461-16382] (16382 slots) master
S: 26c3a7cd48fe4bddf9cf0c60a3c30b6b1f273135 10.95.200.12:7380
slots: (0 slots) slave
replicates 467d4c7508d1cb371ed52c4c6574506cba40c328
M: cf6ca00cb36850762fdff1223684edf1fb9bd4ba 10.95.200.14:7380
slots:[16383] (1 slots) master
1 additional replica(s)
S: 2cb8c6a7c86f512db2c6ca88eb67449197f5b885 10.95.200.12:7381
slots: (0 slots) slave
replicates cf6ca00cb36850762fdff1223684edf1fb9bd4ba
[OK] All nodes agree about slots configuration.
>>> Check for open slots...
>>> Check slots coverage...
[OK] All 16384 slots covered.
Ready to move 5460 slots.
Source nodes:
M: 467d4c7508d1cb371ed52c4c6574506cba40c328 10.95.200.13:7381
slots:[5460] (1 slots) master
1 additional replica(s)
Destination node:
M: ed309033dbefe2b0b64ad7fb643c4d2531e53b95 10.95.200.13:7380
slots:[0-5459],[5461-16382] (16382 slots) master
Resharding plan:
Moving slot 5460 from 467d4c7508d1cb371ed52c4c6574506cba40c328
Moving slot 5460 from 10.95.200.13:7381 to 10.95.200.13:7380:
参数基本说明：
--cluster-from:表示分片的源头节点，即从这个标识指定的参数节点上将分片分配出去给别的节点，参数节点可以有多个，用逗号分隔。
--cluster-to:和--cluster-from相对，标识接收分片的节点，这个参数指定的节点只能有一个。
--cluster-slots:参与重新分片的slot号。
--cluster-yes:不用显示分片的过程信息，直接后台操作。
操作后的集群信息：
[tkiot@tkwh-kfcs-app2 redis]$ redis-cli -c -h 10.95.200.12 -p 7380 cluster nodes ed309033dbefe2b0b64ad7fb643c4d2531e53b95 10.95.200.13:7380@17380 master - 0 1563277825000 11 connected 0-16382 2cb8c6a7c86f512db2c6ca88eb67449197f5b885 10.95.200.12:7381@17381 slave cf6ca00cb36850762fdff1223684edf1fb9bd4ba 0 1563277825000 10 connected cf6ca00cb36850762fdff1223684edf1fb9bd4ba 10.95.200.14:7380@17380 master - 0 1563277825386 10 connected 16383 26c3a7cd48fe4bddf9cf0c60a3c30b6b1f273135 10.95.200.12:7380@17380 myself,slave ed309033dbefe2b0b64ad7fb643c4d2531e53b95 0 1563277824000 1 connected 467d4c7508d1cb371ed52c4c6574506cba40c328 10.95.200.13:7381@17381 master - 0 1563277826390 8 connected
标红的节点，已经connected的slot没有了，在reshard之前，他有5460
3）将master进行下线操作：
[tkiot@tkwh-kfcs-app2 redis]$ redis-cli --cluster del-node 10.95.200.13:7381 467d4c7508d1cb371ed52c4c6574506cba40c328 >>> Removing node 467d4c7508d1cb371ed52c4c6574506cba40c328 from cluster 10.95.200.13:7381 >>> Sending CLUSTER FORGET messages to the cluster... >>> SHUTDOWN the node.
操作后的集群信息（是不是从原来的6个节点变成了现在的4个了）：
[tkiot@tkwh-kfcs-app2 redis]$ redis-cli -c -h 10.95.200.12 -p 7380 cluster nodes ed309033dbefe2b0b64ad7fb643c4d2531e53b95 10.95.200.13:7380@17380 master - 0 1563278285304 11 connected 0-16382 2cb8c6a7c86f512db2c6ca88eb67449197f5b885 10.95.200.12:7381@17381 slave cf6ca00cb36850762fdff1223684edf1fb9bd4ba 0 1563278286304 10 connected cf6ca00cb36850762fdff1223684edf1fb9bd4ba 10.95.200.14:7380@17380 master - 0 1563278284000 10 connected 16383 26c3a7cd48fe4bddf9cf0c60a3c30b6b1f273135 10.95.200.12:7380@17380 myself,slave ed309033dbefe2b0b64ad7fb643c4d2531e53b95 0 1563278284000 1 connected
依据上面的操作，全部删除后，最后只剩下一个master了，就直接kill掉，停服务就好了。
[tkiot@tkwh-kfcs-app2 redis]$ redis-cli -c -h 10.95.200.13 -p 7380 cluster nodes ed309033dbefe2b0b64ad7fb643c4d2531e53b95 10.95.200.13:7380@17380 myself,master - 0 1563321180000 11 connected 0-16383
下面记录一下创建一组4个节点的集群的操作，并添加一主一从的过程。
[tkiot@tkwh-kfcs-app2 redis]$ redis-cli --cluster create 10.95.200.12:7380 10.95.200.13:7380 10.95.200.12:7381 10.95.200.13:7381 --cluster-replicas 1 *** ERROR: Invalid configuration for cluster creation. *** Redis Cluster requires at least 3 master nodes. *** This is not possible with 4 nodes and 1 replicas per node. *** At least 6 nodes are required.
这个过程失败了，呵呵。那就6个节点吧，然后删除再添加。
[tkiot@tkwh-kfcs-app2 redis]$ redis-cli --cluster create 10.95.200.12:7380 10.95.200.13:7380 10.95.200.14:7380 10.95.200.12:7381 10.95.200.13:7381 10.95.200.14:7381 --cluster-replicas 1 >>> Performing hash slots allocation on 6 nodes... Master[0] -> Slots 0 - 5460 Master[1] -> Slots 5461 - 10922 Master[2] -> Slots 10923 - 16383 Adding replica 10.95.200.13:7381 to 10.95.200.12:7380 Adding replica 10.95.200.12:7381 to 10.95.200.13:7380 Adding replica 10.95.200.14:7381 to 10.95.200.14:7380 >>> Trying to optimize slaves allocation for anti-affinity [OK] Perfect anti-affinity obtained! 。。。。。 [OK] All nodes agree about slots configuration. >>> Check for open slots... >>> Check slots coverage... [OK] All 16384 slots covered.
删除两个节点，10.95.200.13:7380 10.95.200.14:7381，删除的过程不多记录（当前是4个节点）。
[tkiot@tkwh-kfcs-app2 redis]$ redis-cli -c -h 10.95.200.12 -p 7381 cluster nodes f855a20340af517e1502f610d82e8280cc6cd803 10.95.200.12:7381@17381 myself,slave 1591155d7df58218a26974b16996eeeba88c84f1 0 1563331430000 4 connected 1cf9c2ef22a7ce7831cfee357aca36bfb6f275b5 10.95.200.13:7381@17381 slave d46ae46cf66445ddeba923d6af84b78ca5f789cb 0 1563331431232 7 connected d46ae46cf66445ddeba923d6af84b78ca5f789cb 10.95.200.12:7380@17380 master - 0 1563331430000 7 connected 0-10922 1591155d7df58218a26974b16996eeeba88c84f1 10.95.200.14:7380@17380 master - 0 1563331430229 3 connected 10923-16383
记得删除掉的节点，其实已经shutdown了，要将他们起起来后，再执行添加操作。
1）将新的节点10.95.200.13:7380添加到集群（选择一个已经在集群中的节点即可）
[tkiot@tkwh-kfcs-app2 redis]$ redis-cli --cluster add-node 10.95.200.13:7380 10.95.200.14:7380 >>> Adding node 10.95.200.13:7380 to cluster 10.95.200.14:7380 >>> Performing Cluster Check (using node 10.95.200.14:7380) M: 1591155d7df58218a26974b16996eeeba88c84f1 10.95.200.14:7380 slots:[10923-16383] (5461 slots) master 1 additional replica(s) S: 1cf9c2ef22a7ce7831cfee357aca36bfb6f275b5 10.95.200.13:7381 slots: (0 slots) slave replicates d46ae46cf66445ddeba923d6af84b78ca5f789cb S: f855a20340af517e1502f610d82e8280cc6cd803 10.95.200.12:7381 slots: (0 slots) slave replicates 1591155d7df58218a26974b16996eeeba88c84f1 M: d46ae46cf66445ddeba923d6af84b78ca5f789cb 10.95.200.12:7380 slots:[0-10922] (10923 slots) master 1 additional replica(s) [OK] All nodes agree about slots configuration. >>> Check for open slots... >>> Check slots coverage... [OK] All 16384 slots covered. >>> Send CLUSTER MEET to node 10.95.200.13:7380 to make it join the cluster. [OK] New node added correctly.
加完后，看看集群：
[tkiot@tkwh-kfcs-app2 redis]$ redis-cli -c -h 10.95.200.12 -p 7381 cluster nodes f855a20340af517e1502f610d82e8280cc6cd803 10.95.200.12:7381@17381 myself,slave 1591155d7df58218a26974b16996eeeba88c84f1 0 1563332042000 4 connected 1cf9c2ef22a7ce7831cfee357aca36bfb6f275b5 10.95.200.13:7381@17381 slave d46ae46cf66445ddeba923d6af84b78ca5f789cb 0 1563332044428 7 connected 4fd334f64bc5b121f5810da3b5800a29d4e8c3ee 10.95.200.13:7380@17380 master - 0 1563332044000 0 connected d46ae46cf66445ddeba923d6af84b78ca5f789cb 10.95.200.12:7380@17380 master - 0 1563332043426 7 connected 0-10922 1591155d7df58218a26974b16996eeeba88c84f1 10.95.200.14:7380@17380 master - 0 1563332043000 3 connected 10923-16383
2）新加入的一个节点默认会变成master，下面给这个master分配slot，就是一个reshard的过程
[tkiot@tkwh-kfcs-app2 redis]$ redis-cli --cluster reshard 10.95.200.13:7380 --cluster-from d46ae46cf66445ddeba923d6af84b78ca5f789cb,1591155d7df58218a26974b16996eeeba88c84f1 --cluster-to 4fd334f64bc5b121f5810da3b5800a29d4e8c3ee --cluster-slots 5642
添加完后，看看集群：
[tkiot@tkwh-kfcs-app2 redis]$ redis-cli -c -h 10.95.200.12 -p 7381 cluster nodes f855a20340af517e1502f610d82e8280cc6cd803 10.95.200.12:7381@17381 myself,slave 1591155d7df58218a26974b16996eeeba88c84f1 0 1563332267000 4 connected 1cf9c2ef22a7ce7831cfee357aca36bfb6f275b5 10.95.200.13:7381@17381 slave d46ae46cf66445ddeba923d6af84b78ca5f789cb 0 1563332267021 7 connected 4fd334f64bc5b121f5810da3b5800a29d4e8c3ee 10.95.200.13:7380@17380 master - 0 1563332268022 8 connected 0-3761 10923-12802 d46ae46cf66445ddeba923d6af84b78ca5f789cb 10.95.200.12:7380@17380 master - 0 1563332267521 7 connected 3762-10922 1591155d7df58218a26974b16996eeeba88c84f1 10.95.200.14:7380@17380 master - 0 1563332267000 3 connected 12803-16383
3）给刚才添加的master节点添加slave节点
[tkiot@tkwh-kfcs-app2 redis]$ redis-cli --cluster add-node 10.95.200.14:7381 10.95.200.13:7380 --cluster-slave --cluster-master-id 4fd334f64bc5b121f5810da3b5800a29d4e8c3ee >>> Adding node 10.95.200.14:7381 to cluster 10.95.200.13:7380 >>> Performing Cluster Check (using node 10.95.200.13:7380) M: 4fd334f64bc5b121f5810da3b5800a29d4e8c3ee 10.95.200.13:7380 slots:[0-3761],[10923-12802] (5642 slots) master 1 additional replica(s) S: 1cf9c2ef22a7ce7831cfee357aca36bfb6f275b5 10.95.200.13:7381 slots: (0 slots) slave replicates 4fd334f64bc5b121f5810da3b5800a29d4e8c3ee M: d46ae46cf66445ddeba923d6af84b78ca5f789cb 10.95.200.12:7380 slots:[3762-10922] (7161 slots) master M: 1591155d7df58218a26974b16996eeeba88c84f1 10.95.200.14:7380 slots:[12803-16383] (3581 slots) master 1 additional replica(s) S: f855a20340af517e1502f610d82e8280cc6cd803 10.95.200.12:7381 slots: (0 slots) slave replicates 1591155d7df58218a26974b16996eeeba88c84f1 [OK] All nodes agree about slots configuration. >>> Check for open slots... >>> Check slots coverage... [OK] All 16384 slots covered. >>> Send CLUSTER MEET to node 10.95.200.14:7381 to make it join the cluster. Waiting for the cluster to join >>> Configure node as replica of 10.95.200.13:7380. [OK] New node added correctly.
最后再看看集群：
[tkiot@tkwh-kfcs-app2 redis]$ redis-cli -c -h 10.95.200.12 -p 7381 cluster nodes f855a20340af517e1502f610d82e8280cc6cd803 10.95.200.12:7381@17381 myself,slave 1591155d7df58218a26974b16996eeeba88c84f1 0 1563332620000 4 connected f6093d443470ae37b8330407d20291ae959fc22f :0@0 slave,fail,noaddr d46ae46cf66445ddeba923d6af84b78ca5f789cb 1563332505335 1563332504432 7 disconnected 1cf9c2ef22a7ce7831cfee357aca36bfb6f275b5 10.95.200.13:7381@17381 slave d46ae46cf66445ddeba923d6af84b78ca5f789cb 0 1563332621533 8 connected 4fd334f64bc5b121f5810da3b5800a29d4e8c3ee 10.95.200.13:7380@17380 master - 0 1563332621633 8 connected 0-3761 10923-12802 d46ae46cf66445ddeba923d6af84b78ca5f789cb 10.95.200.12:7380@17380 master - 0 1563332621533 7 connected 3762-10922 1591155d7df58218a26974b16996eeeba88c84f1 10.95.200.14:7380@17380 master - 0 1563332621000 3 connected 12803-16383 9f069b1f62cc93b95cb432f3726bc5bbfb0c8c76 10.95.200.14:7381@17381 slave 4fd334f64bc5b121f5810da3b5800a29d4e8c3ee 0 1563332620533 8 connected
总结一下集群节点添加的过程：
A) 增加节点（前提是对应节点服务器已经运行了），默认成为master节点，命令：redis-cli --cluster add-node <new_ip:new_port> <existing_ip:existing_port>
B) 给新增加的这个master节点分配slot，命令：redis-cli --cluster reshard <master_ip:master_port> --cluster-from <集群已存在master的nodeId，多个需要逗号分隔> --cluster-to <当前master的nodeId> --cluster-slots <分配slot的数量>
C) 给这个master添加slave，slot就不用了，因为是slave，命令：redis-cli --cluster add-node <slave_ip:slave_port> <master_ip:master_port> --cluster-slave --cluster-master-id <mastr node_id>
最后，有一个问题，没有搞清楚，就是将节点从集群删除后，再添加进去的时候，总会遇到提示说该节点不是空节点，导致添加失败，采取的策略就是将待添加的节点对应的数据文件全部删除后，重新启动，再执行添加操作指令，就成功了。
来源：https://www.cnblogs.com/shihuc/p/11212280.html