3>添加slave节点

下面为201:6382master节点添加一个slave节点201:6383

shell>redis-trib.rbadd-node--slave--master-id901cade1bc2703ef2bb4ff49466b38762346755d192.168.0.201:6383192.168.0.201:6379

参数说明：1>–slave：表示添加从节点2>–master-id：master节点ID，这里用上边新添加的主节点ID（201:6382）3>192.168.0.201:6383：新slave节点的IP和端口4>192.168.0.201:6379：集群中任意旧节点，只要能连上就可以。这里以201:6379为例

添加slave节点成功后集群中的节点状态：

[root@localhostredis-3.2.0]#redis-cliclusternodes901cade1bc2703ef2bb4ff49466b38762346755d192.168.0.201:6382master-014683338426659connected0-3325461-579410923-11255459dbcccc1e5be021f19efde794b9e84474c0c02192.168.0.202:16379slaveb6814c717080de33907e6f769e1fd8dc4fe53c38014683338416617connected9080c5049615730c66731b54cb084999009eeb52192.168.0.201:6383slave901cade1bc2703ef2bb4ff49466b38762346755d014683338436699connected06023f8283389ad5580c782797f49a5996c108b3192.168.0.201:6379myself,master-001connected333-546091179b96d9b8abce85fa7a5748aa9b4114f0b260192.168.0.202:16380slave06023f8283389ad5580c782797f49a5996c108b3014683338446745connectedef8a49e5242c791551f1b317d5039144e38acde6192.168.0.201:6380slave13567f6f8787d3aa3f4ef3abd15c2cc500a1469b014683338426658connectedb6814c717080de33907e6f769e1fd8dc4fe53c38192.168.0.201:6381master-014683338456787connected5795-1092213567f6f8787d3aa3f4ef3abd15c2cc500a1469b192.168.0.202:16381master-014683338466858connected11256-16383

verygood~可以看出，201:6383以成功添加为一个slave节点，并将201:6382作为master节点。

命令列表的节点状态信息说明：

459dbcccc1e5be021f19efde794b9e84474c0c02192.168.0.202:16379slaveb6814c717080de33907e6f769e1fd8dc4fe53c38014683338416617connected

1>459dbcccc1e5be021f19efde794b9e84474c0c02：节点ID2>192.168.0.202:16379：节点IP和端口3>slave：节点类型，master、slave或myself(master/slave)4>b6814c717080de33907e6f769e1fd8dc4fe53c38：master节点ID，如果自己是master节点，这一列的值为05>0：集群最近一次向节点发送PING命令之后，过去了多长时间还没接到回复。6>1468333841661：节点最近一次返回PONG回复的时间7>7：节点的配置纪元8>connected：节点网络连接状态9>如果是主节点，节点状态后边显示slot分配的范围

将上边新加的slave节点（201:6383）的master，修改为201:6379（06023f8283389ad5580c782797f49a5996c108b3）

shell>redis-cli-h192.168.0.201-p6383clusternodes|grepmyself9080c5049615730c66731b54cb084999009eeb52192.168.0.201:6383myself,slave901cade1bc2703ef2bb4ff49466b38762346755d000connected

当前slave节点201:6383的master节点是901cade1bc2703ef2bb4ff49466b38762346755d

shell>redis-cli-h192.168.0.201-p6383192.168.0.201:6383>clusterreplicate06023f8283389ad5580c782797f49a5996c108b3

192.168.0.201:6383>CLUSTERSLAVES06023f8283389ad5580c782797f49a5996c108b31)"91179b96d9b8abce85fa7a5748aa9b4114f0b260192.168.0.202:16380slave06023f8283389ad5580c782797f49a5996c108b3014683361246001connected"2)"9080c5049615730c66731b54cb084999009eeb52192.168.0.201:6383myself,slave06023f8283389ad5580c782797f49a5996c108b3000connected"

可以看出201:6383slave节点已经是201:6379master节点的子节点了。

1>删除slave节点

[root@localhostredis-3.2.0]#redis-trib.rbdel-node192.168.0.201:63839080c5049615730c66731b54cb084999009eeb52>>>Removingnode9080c5049615730c66731b54cb084999009eeb52fromcluster192.168.0.201:6383>>>SendingCLUSTERFORGETmessagestothecluster...>>>SHUTDOWNthenode.

参数说明：1>del-node：删除节点参数2>192.168.0.201:6383：节点IP和端口3>9080c5049615730c66731b54cb084999009eeb52：节点ID节点成功删除后，会同时将进程杀死

从节点重新恢复后，先将从节点加入集群，然后设置自己的master

2、删除master节点

注意：1>如果master节点有slot，必须先将所有slot分配给其它master节点。否则会提示你先将slot移除再尝试删除，如下所示：

shell>redis-trib.rbdel-node192.168.0.201:6382901cade1bc2703ef2bb4ff49466b38762346755d>>>Removingnode901cade1bc2703ef2bb4ff49466b38762346755dfromcluster192.168.0.201:6382[ERR]Node192.168.0.201:6382isnotempty!Resharddataawayandtryagain.

2>如果master节点有slave节点，先将slave节点移到其它master下

新增master节点后，也进行了这一步操作，当时是分配，现在是去掉，反着的。