轉(zhuǎn)自：https://www.zybuluo.com/phper/note/195558

前言

redis 是我們目前大規(guī)模使用的緩存中間件菌湃，由于它強(qiáng)大高效而又便捷的功能，得到了廣泛的使用∩舛叮現(xiàn)在的2.x的穩(wěn)定版本是2.8.19,也是我們項(xiàng)目中普遍用到的版本。

redis在年初發(fā)布了3.0.0,官方支持了redis cluster,也就是集群辅肾。至此結(jié)束了redis 沒(méi)有官方集群的時(shí)代队萤，之前我們用redis cluster用的最多的應(yīng)該是twitter 發(fā)布的Twemproxy(https://github.com/twitter/twemproxy)

還有就是豌豆莢開發(fā)的codis (https://github.com/wandoulabs/codis)

這2個(gè)我會(huì)在后續(xù)去使用它們。下面的文字矫钓，我盡量用通熟易懂的方式來(lái)闡述要尔。

redis cluster 理論知識(shí)

截止寫這篇文章前，redis 3.x的最新版本是3.0.5新娜。今天的學(xué)習(xí)和實(shí)踐就是用這個(gè)版本赵辕。

Redis Cluster設(shè)計(jì)要點(diǎn)

redis cluster在設(shè)計(jì)的時(shí)候，就考慮到了去中心化概龄，去中間件匆帚，也就是說(shuō)，集群中的每個(gè)節(jié)點(diǎn)都是平等的關(guān)系旁钧，都是對(duì)等的吸重，每個(gè)節(jié)點(diǎn)都保存各自的數(shù)據(jù)和整個(gè)集群的狀態(tài)。每個(gè)節(jié)點(diǎn)都和其他所有節(jié)點(diǎn)連接歪今，而且這些連接保持活躍嚎幸，這樣就保證了我們只需要連接集群中的任意一個(gè)節(jié)點(diǎn)，就可以獲取到其他節(jié)點(diǎn)的數(shù)據(jù)寄猩。

那么redis 是如何合理分配這些節(jié)點(diǎn)和數(shù)據(jù)的呢嫉晶？

Redis 集群沒(méi)有并使用傳統(tǒng)的一致性哈希來(lái)分配數(shù)據(jù)，而是采用另外一種叫做哈希槽 (hash slot)的方式來(lái)分配的田篇。redis cluster 默認(rèn)分配了 16384 個(gè)slot替废，當(dāng)我們set一個(gè)key 時(shí)，會(huì)用CRC16算法來(lái)取模得到所屬的slot泊柬，然后將這個(gè)key 分到哈希槽區(qū)間的節(jié)點(diǎn)上椎镣，具體算法就是：CRC16(key) % 16384。

注意的是：必須要3個(gè)以后的主節(jié)點(diǎn)兽赁，否則在創(chuàng)建集群時(shí)會(huì)失敗状答，我們?cè)诤罄m(xù)會(huì)實(shí)踐到。

所以刀崖，我們假設(shè)現(xiàn)在有3個(gè)節(jié)點(diǎn)已經(jīng)組成了集群惊科，分別是：A, B, C 三個(gè)節(jié)點(diǎn)，它們可以是一臺(tái)機(jī)器上的三個(gè)端口亮钦，也可以是三臺(tái)不同的服務(wù)器馆截。那么，采用哈希槽 (hash slot)的方式來(lái)分配16384個(gè)slot 的話蜂莉，它們?nèi)齻€(gè)節(jié)點(diǎn)分別承擔(dān)的slot 區(qū)間是：

節(jié)點(diǎn)A覆蓋0－5460;
節(jié)點(diǎn)B覆蓋5461－10922;
節(jié)點(diǎn)C覆蓋10923－16383.

如下圖所示：

那么蜡娶，現(xiàn)在我想設(shè)置一個(gè)key ,比如叫my_name:

set my_name yangyi

按照redis cluster的哈希槽算法：CRC16('my_name')%16384 = 2412混卵。 那么就會(huì)把這個(gè)key 的存儲(chǔ)分配到 A 上了。

同樣翎蹈，當(dāng)我連接(A,B,C)任何一個(gè)節(jié)點(diǎn)想獲取my_name這個(gè)key時(shí)淮菠，也會(huì)這樣的算法，然后內(nèi)部跳轉(zhuǎn)到B節(jié)點(diǎn)上獲取數(shù)據(jù)荤堪。

這種哈希槽的分配方式有好也有壞合陵，好處就是很清晰，比如我想新增一個(gè)節(jié)點(diǎn)D澄阳，redis cluster的這種做法是從各個(gè)節(jié)點(diǎn)的前面各拿取一部分slot到D上拥知，我會(huì)在接下來(lái)的實(shí)踐中實(shí)驗(yàn)。大致就會(huì)變成這樣：

節(jié)點(diǎn)A覆蓋1365-5460
節(jié)點(diǎn)B覆蓋6827-10922
節(jié)點(diǎn)C覆蓋12288-16383
節(jié)點(diǎn)D覆蓋0-1364,5461-6826,10923-12287

同樣刪除一個(gè)節(jié)點(diǎn)也是類似碎赢，移動(dòng)完成后就可以刪除這個(gè)節(jié)點(diǎn)了低剔。

所以redis cluster 就是這樣的一個(gè)形狀：

Redis Cluster主從模式

redis cluster 為了保證數(shù)據(jù)的高可用性，加入了主從模式肮塞，一個(gè)主節(jié)點(diǎn)對(duì)應(yīng)一個(gè)或多個(gè)從節(jié)點(diǎn)襟齿，主節(jié)點(diǎn)提供數(shù)據(jù)存取，從節(jié)點(diǎn)則是從主節(jié)點(diǎn)拉取數(shù)據(jù)備份枕赵，當(dāng)這個(gè)主節(jié)點(diǎn)掛掉后猜欺，就會(huì)有這個(gè)從節(jié)點(diǎn)選取一個(gè)來(lái)充當(dāng)主節(jié)點(diǎn)，從而保證集群不會(huì)掛掉拷窜。

上面那個(gè)例子里, 集群有ABC三個(gè)主節(jié)點(diǎn), 如果這3個(gè)節(jié)點(diǎn)都沒(méi)有加入從節(jié)點(diǎn)开皿，如果B掛掉了，我們就無(wú)法訪問(wèn)整個(gè)集群了篮昧。A和C的slot也無(wú)法訪問(wèn)赋荆。

所以我們?cè)诩航⒌臅r(shí)候，一定要為每個(gè)主節(jié)點(diǎn)都添加了從節(jié)點(diǎn), 比如像這樣, 集群包含主節(jié)點(diǎn)A懊昨、B窄潭、C, 以及從節(jié)點(diǎn)A1、B1疚颊、C1, 那么即使B掛掉系統(tǒng)也可以繼續(xù)正確工作狈孔。

B1節(jié)點(diǎn)替代了B節(jié)點(diǎn)，所以Redis集群將會(huì)選擇B1節(jié)點(diǎn)作為新的主節(jié)點(diǎn)材义，集群將會(huì)繼續(xù)正確地提供服務(wù)。 當(dāng)B重新開啟后嫁赏，它就會(huì)變成B1的從節(jié)點(diǎn)其掂。

不過(guò)需要注意，如果節(jié)點(diǎn)B和B1同時(shí)掛了潦蝇，Redis集群就無(wú)法繼續(xù)正確地提供服務(wù)了款熬。

流程下圖所示：

redis cluster 動(dòng)手實(shí)踐

先下載官方的redis 版本(3.0.5) : http://download.redis.io/releases/redis-3.0.5.tar.gz

1深寥、解壓

[root@web3 ~]# tar zxvf redis-3.0.5.tar.gz

2、安裝

[root@web3 ~]# cd redis-3.0.5
[root@web3 ~]# make && make install

3贤牛、將redis-trib.rb 復(fù)制到/usr/local/bin

[root@web3 redis-3.0.5]# cd src/
[root@web3 src]# cp redis-trib.rb /usr/local/bin

4惋鹅、開始集群搭建

[root@web3 redis-3.0.5]# vi redis.conf
###### 修改以下地方
port 7000
cluster-enabled yes
cluster-config-file nodes.conf
cluster-node-timeout 5000
appendonly yes

新建6個(gè)節(jié)點(diǎn)：

[root@web3 redis-3.0.5]# mkdir -p /usr/local/cluster-test
[root@web3 redis-3.0.5]# cd /usr/local/cluster-test/
[root@web3 cluster-test]# mkdir 7000
[root@web3 cluster-test]# mkdir 7001
[root@web3 cluster-test]# mkdir 7002
[root@web3 cluster-test]# mkdir 7003
[root@web3 cluster-test]# mkdir 7004
[root@web3 cluster-test]# mkdir 7005

將redis.conf 分別拷貝到這6個(gè)文件夾中,并修改成對(duì)應(yīng)的端口號(hào)

###### 拷貝配置文件
[root@web3 cluster-test]# cp /root/redis-3.0.5/redis.conf  /usr/local/cluster-test/7000
[root@web3 cluster-test]# cp /root/redis-3.0.5/redis.conf  /usr/local/cluster-test/7001
[root@web3 cluster-test]# cp /root/redis-3.0.5/redis.conf  /usr/local/cluster-test/7002
[root@web3 cluster-test]# cp /root/redis-3.0.5/redis.conf  /usr/local/cluster-test/7003
[root@web3 cluster-test]# cp /root/redis-3.0.5/redis.conf  /usr/local/cluster-test/7004
[root@web3 cluster-test]# cp /root/redis-3.0.5/redis.conf  /usr/local/cluster-test/7005
##### 修改端口號(hào)
root@web3 cluster-test]# sed -i "s/7000/7001/g" /usr/local/cluster-test/7001/redis.conf
[root@web3 cluster-test]# sed -i "s/7000/7002/g" /usr/local/cluster-test/7002/redis.conf
[root@web3 cluster-test]# sed -i "s/7000/7003/g" /usr/local/cluster-test/7003/redis.conf
[root@web3 cluster-test]# sed -i "s/7000/7004/g" /usr/local/cluster-test/7004/redis.conf
[root@web3 cluster-test]# sed -i "s/7000/7005/g" /usr/local/cluster-test/7005/redis.conf

分別，啟動(dòng)這6個(gè)節(jié)點(diǎn)

[root@web3 cluster-test]# cd /usr/local/cluster-test/7000/
[root@web3 7000]# redis-server redis.conf
[root@web3 7000]# cd ../7001
[root@web3 7001]# redis-server redis.conf
[root@web3 7001]# cd ../7002
[root@web3 7002]# redis-server redis.conf
[root@web3 7002]# cd ../7003
[root@web3 7003]# redis-server redis.conf
[root@web3 7003]# cd ../7004
[root@web3 7004]# redis-server redis.conf
[root@web3 7004]# cd ../7005
[root@web3 7005]# redis-server redis.conf
[root@web3 7005]#

查看6個(gè)節(jié)點(diǎn)的啟動(dòng)進(jìn)程情況：

[root@web3 7005]# ps -ef|grep redis
root     11380     1  0 07:37 ?        00:00:00 redis-server *:7000 [cluster] 
root     11384     1  0 07:37 ?        00:00:00 redis-server *:7001 [cluster] 
root     11388     1  0 07:37 ?        00:00:00 redis-server *:7002 [cluster] 
root     11392     1  0 07:37 ?        00:00:00 redis-server *:7003 [cluster] 
root     11396     1  0 07:37 ?        00:00:00 redis-server *:7004 [cluster] 
root     11400     1  0 07:37 ?        00:00:00 redis-server *:7005 [cluster] 
root     11404  8259  0 07:38 pts/0    00:00:00 grep redis

將6個(gè)節(jié)點(diǎn)連在一起構(gòu)招成集群

需要用到的命令就是redis-trib.rb,這是官方的一個(gè)用ruby寫的一個(gè)操作redis cluster的命令殉簸，所以闰集，你的機(jī)器上需要安裝ruby。我們先試一下這個(gè)命令：

redis-trib.rb create --replicas 1 127.0.0.1:7000 127.0.0.1:7001 127.0.0.1:7002 127.0.0.1:7003 127.0.0.1:7004 127.0.0.1:7005

因?yàn)槲覀円陆? 所以這里使用create命令. --replicas 1 參數(shù)表示為每個(gè)主節(jié)點(diǎn)創(chuàng)建一個(gè)從節(jié)點(diǎn). 其他參數(shù)是實(shí)例的地址集合般卑。

由于我機(jī)子上沒(méi)安裝ruby武鲁，所以，會(huì)報(bào)錯(cuò)：

/usr/bin/env: ruby: No such file or directory

那先安裝ruby和rubygems：

[root@web3 7005]# yum install ruby ruby-devel rubygems rpm-build

然后蝠检，再執(zhí)行一次沐鼠，發(fā)現(xiàn)還是報(bào)錯(cuò)：

[root@web3 7005]# redis-trib.rb create --replicas 1 127.0.0.1:7000 127.0.0.1:7001 127.0.0.1:7002 127.0.0.1:7003 127.0.0.1:7004 127.0.0.1:7005
/usr/lib/ruby/site_ruby/1.8/rubygems/custom_require.rb:31:in `gem_original_require': no such file to load -- redis (LoadError)
    from /usr/lib/ruby/site_ruby/1.8/rubygems/custom_require.rb:31:in `require'
    from /usr/local/bin/redis-trib.rb:25
[root@web3 7005]#

原來(lái)是ruby和redis的連接沒(méi)安裝好：

[root@web3 7005]# gem install redis
Successfully installed redis-3.2.1
1 gem installed
Installing ri documentation for redis-3.2.1...
Installing RDoc documentation for redis-3.2.1...

再次執(zhí)行構(gòu)建集群的命令：

[root@web3 7005]# redis-trib.rb create --replicas 1 127.0.0.1:7000 127.0.0.1:7001 127.0.0.1:7002 127.0.0.1:7003 127.0.0.1:7004 127.0.0.1:7005
>>> Creating cluster
Connecting to node 127.0.0.1:7000: OK
Connecting to node 127.0.0.1:7001: OK
Connecting to node 127.0.0.1:7002: OK
Connecting to node 127.0.0.1:7003: OK
Connecting to node 127.0.0.1:7004: OK
Connecting to node 127.0.0.1:7005: OK
>>> Performing hash slots allocation on 6 nodes...
Using 3 masters:
127.0.0.1:7000
127.0.0.1:7001
127.0.0.1:7002
Adding replica 127.0.0.1:7003 to 127.0.0.1:7000
Adding replica 127.0.0.1:7004 to 127.0.0.1:7001
Adding replica 127.0.0.1:7005 to 127.0.0.1:7002
M: 3707debcbe7be66d4a1968eaf3a5ffaf4308efa4 127.0.0.1:7000
   slots:0-5460 (5461 slots) master
M: cb5c04b6160c3b7e18cad5d49d8e2987b27e0d6c 127.0.0.1:7001
   slots:5461-10922 (5462 slots) master
M: dfa0754c7854a874a6ebd2613b86140ad97701fc 127.0.0.1:7002
   slots:10923-16383 (5461 slots) master
S: d2237fdcfbba672de766b913d1186cebcb6e1761 127.0.0.1:7003
   replicates 3707debcbe7be66d4a1968eaf3a5ffaf4308efa4
S: 4b4aef8b48c427a3c903518339d53b6447c58b93 127.0.0.1:7004
   replicates cb5c04b6160c3b7e18cad5d49d8e2987b27e0d6c
S: 30858dbf483b61b9838d5c1f853a60beaa4e7afd 127.0.0.1:7005
   replicates dfa0754c7854a874a6ebd2613b86140ad97701fc
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 127.0.0.1:7000)
M: 3707debcbe7be66d4a1968eaf3a5ffaf4308efa4 127.0.0.1:7000
   slots:0-5460 (5461 slots) master
M: cb5c04b6160c3b7e18cad5d49d8e2987b27e0d6c 127.0.0.1:7001
   slots:5461-10922 (5462 slots) master
M: dfa0754c7854a874a6ebd2613b86140ad97701fc 127.0.0.1:7002
   slots:10923-16383 (5461 slots) master
M: d2237fdcfbba672de766b913d1186cebcb6e1761 127.0.0.1:7003
   slots: (0 slots) master
   replicates 3707debcbe7be66d4a1968eaf3a5ffaf4308efa4
M: 4b4aef8b48c427a3c903518339d53b6447c58b93 127.0.0.1:7004
   slots: (0 slots) master
   replicates cb5c04b6160c3b7e18cad5d49d8e2987b27e0d6c
M: 30858dbf483b61b9838d5c1f853a60beaa4e7afd 127.0.0.1:7005
   slots: (0 slots) master
   replicates dfa0754c7854a874a6ebd2613b86140ad97701fc
[OK] All nodes agree about slots configuration.
>>> Check for open slots...
>>> Check slots coverage...
[OK] All 16384 slots covered.

redis-trib會(huì)提示你做了什么配置, 輸入yes接受. 集群就被配置和加入了, 意思是, 實(shí)例會(huì)經(jīng)過(guò)互相交流后啟動(dòng)。

測(cè)試集群的狀態(tài)：

[root@web3 7000]# redis-trib.rb check 127.0.0.1:7000
Connecting to node 127.0.0.1:7000: OK
Connecting to node 127.0.0.1:7002: OK
Connecting to node 127.0.0.1:7003: OK
Connecting to node 127.0.0.1:7005: OK
Connecting to node 127.0.0.1:7001: OK
Connecting to node 127.0.0.1:7004: OK
>>> Performing Cluster Check (using node 127.0.0.1:7000)
M: 3707debcbe7be66d4a1968eaf3a5ffaf4308efa4 127.0.0.1:7000
   slots:0-5460 (5461 slots) master
   1 additional replica(s)
M: dfa0754c7854a874a6ebd2613b86140ad97701fc 127.0.0.1:7002
   slots:10923-16383 (5461 slots) master
   1 additional replica(s)
S: d2237fdcfbba672de766b913d1186cebcb6e1761 127.0.0.1:7003
   slots: (0 slots) slave
   replicates 3707debcbe7be66d4a1968eaf3a5ffaf4308efa4
S: 30858dbf483b61b9838d5c1f853a60beaa4e7afd 127.0.0.1:7005
   slots: (0 slots) slave
   replicates dfa0754c7854a874a6ebd2613b86140ad97701fc
M: cb5c04b6160c3b7e18cad5d49d8e2987b27e0d6c 127.0.0.1:7001
   slots:5461-10922 (5462 slots) master
   1 additional replica(s)
S: 4b4aef8b48c427a3c903518339d53b6447c58b93 127.0.0.1:7004
   slots: (0 slots) slave
   replicates cb5c04b6160c3b7e18cad5d49d8e2987b27e0d6c
[OK] All nodes agree about slots configuration.
>>> Check for open slots...
>>> Check slots coverage...
[OK] All 16384 slots covered.

可以看到有3個(gè)主節(jié)點(diǎn)叹谁，3個(gè)從節(jié)點(diǎn)饲梭。每個(gè)節(jié)點(diǎn)都是成功的連接狀態(tài)。

3個(gè)主節(jié)點(diǎn)［M］是：

7000 (3707debcbe7be66d4a1968eaf3a5ffaf4308efa4)
7001 (cb5c04b6160c3b7e18cad5d49d8e2987b27e0d6c)
7002 (dfa0754c7854a874a6ebd2613b86140ad97701fc)

3個(gè)從節(jié)點(diǎn)［S］是：

7003 (d2237fdcfbba672de766b913d1186cebcb6e1761)->7000
7004 (4b4aef8b48c427a3c903518339d53b6447c58b93)->7001
7005 (30858dbf483b61b9838d5c1f853a60beaa4e7afd) ->7002

5焰檩、測(cè)試連接集群

剛才集群搭建成功了憔涉。按照redis cluster的特點(diǎn)，它是去中心化锅尘，每個(gè)節(jié)點(diǎn)都是對(duì)等的监氢，所以，你連接哪個(gè)節(jié)點(diǎn)都可以獲取和設(shè)置數(shù)據(jù)藤违，我們來(lái)試一下浪腐。

redis-cli是redis默認(rèn)的客戶端工具，啟動(dòng)時(shí)加上｀-c｀參數(shù)顿乒，就可以連接到集群议街。

連接任意一個(gè)節(jié)點(diǎn)端口：

[root@web3 7000]# redis-cli -c -p 7000
127.0.0.1:7000>

設(shè)置一個(gè)值：

127.0.0.1:7000> set my_name yangyi
-> Redirected to slot [12803] located at 127.0.0.1:7002
OK
127.0.0.1:7002> get my_name
"yangyi"
127.0.0.1:7002>

前面理論知識(shí)我們知道了，分配key的時(shí)候璧榄，它會(huì)使用CRC16('my_name')%16384算法特漩，來(lái)計(jì)算，將這個(gè)key 放到哪個(gè)節(jié)點(diǎn)骨杂，這里分配到了12803slot 就分配到了7002(10923-16383)這個(gè)節(jié)點(diǎn)上涂身。

Redirected to slot [12803] located at 127.0.0.1:7002

redis cluster 采用的方式很直接，它直接跳轉(zhuǎn)到7002 節(jié)點(diǎn)了搓蚪，而不是還在自身的7000節(jié)點(diǎn)蛤售。

好，現(xiàn)在我們連接7005這個(gè)從節(jié)點(diǎn)：

[root@web3 7000]# redis-cli -c -p 7005
127.0.0.1:7005> get my_name
-> Redirected to slot [12803] located at 127.0.0.1:7002
"yangyi"
127.0.0.1:7002>

我們同樣是獲取my_name的值，它同樣也是跳轉(zhuǎn)到了7002上悴能。

我們?cè)賴L試一些其他的可以：

127.0.0.1:7002> set age 123
-> Redirected to slot [741] located at 127.0.0.1:7000
OK
127.0.0.1:7000> set height 565
-> Redirected to slot [8223] located at 127.0.0.1:7001
OK
127.0.0.1:7001> set sex 1
-> Redirected to slot [2584] located at 127.0.0.1:7000
OK
127.0.0.1:7000> set home china
-> Redirected to slot [10814] located at 127.0.0.1:7001
OK
127.0.0.1:7001> set city shanghai
-> Redirected to slot [11479] located at 127.0.0.1:7002
OK
127.0.0.1:7002> set citylocate shanghaipudong
OK
127.0.0.1:7001> set wsdwxzx  hhh
-> Redirected to slot [15487] located at 127.0.0.1:7002
OK
127.0.0.1:7002> zadd erew 333 rrr
-> Redirected to slot [10576] located at 127.0.0.1:7001
(integer) 1
127.0.0.1:7000> zrange erew 0 -1
-> Redirected to slot [10576] located at 127.0.0.1:7001
1) "rrr"
127.0.0.1:7001>

可以看出揣钦，數(shù)據(jù)都會(huì)在7000-7002 這3個(gè)主節(jié)點(diǎn)來(lái)跳轉(zhuǎn)存儲(chǔ)。

6漠酿、測(cè)試集群中的節(jié)點(diǎn)掛掉

上面我們建立來(lái)了一個(gè)集群冯凹。3個(gè)主節(jié)點(diǎn)［7000-7002］提供數(shù)據(jù)存粗和讀取霉颠，3個(gè)從節(jié)點(diǎn)［7003-7005］則是負(fù)責(zé)把［7000-7002］的數(shù)據(jù)同步到自己的節(jié)點(diǎn)上來(lái)熔号，我們來(lái)看一下［7003-7005］的appendonly.aof的內(nèi)容育勺〈⒕兀看看是不是不這樣：

[root@web3 7005]# cd /usr/local/cluster-test/7003
[root@web3 7003]# vi appendonly.aof
*2
$6
SELECT
$1
0
*3
$3
set
$3
age
$3
123
*3
$3
set
$3
sex
$1
1
*3
$3
set
$3
job
$3
php

我們看下条篷，的確是從7000節(jié)點(diǎn)上同步過(guò)來(lái)的數(shù)據(jù)纬凤，7004避矢，7005也是畜疾。

下面寸痢，我們先來(lái)模擬其中一臺(tái)Master主服務(wù)器掛掉的情況呀洲，那就7000掛掉吧：

[root@web3 7003]# ps -ef|grep redis
root     11380     1  0 07:37 ?        00:00:03 redis-server *:7000 [cluster]
root     11384     1  0 07:37 ?        00:00:03 redis-server *:7001 [cluster]
root     11388     1  0 07:37 ?        00:00:03 redis-server *:7002 [cluster]
root     11392     1  0 07:37 ?        00:00:03 redis-server *:7003 [cluster]
root     11396     1  0 07:37 ?        00:00:04 redis-server *:7004 [cluster]
root     11400     1  0 07:37 ?        00:00:03 redis-server *:7005 [cluster]
[root@web3 7003]# kill 11380
[root@web3 7003]#

好，安裝前面的理論啼止，7000主節(jié)點(diǎn)掛掉了道逗，那么這個(gè)時(shí)候，7000的從節(jié)點(diǎn)只有7003一個(gè)献烦，肯定7003就會(huì)被選舉稱Master節(jié)點(diǎn)了：

[root@web3 7003]# redis-trib.rb check 127.0.0.1:7000
Connecting to node 127.0.0.1:7000: [ERR] Sorry, can't connect to node 127.0.0.1:7000
[root@web3 7003]# redis-trib.rb check 127.0.0.1:7001
Connecting to node 127.0.0.1:7001: OK
Connecting to node 127.0.0.1:7004: OK
Connecting to node 127.0.0.1:7005: OK
Connecting to node 127.0.0.1:7003: OK
Connecting to node 127.0.0.1:7002: OK
>>> Performing Cluster Check (using node 127.0.0.1:7001)
M: cb5c04b6160c3b7e18cad5d49d8e2987b27e0d6c 127.0.0.1:7001
   slots:5461-10922 (5462 slots) master
   1 additional replica(s)
S: 4b4aef8b48c427a3c903518339d53b6447c58b93 127.0.0.1:7004
   slots: (0 slots) slave
   replicates cb5c04b6160c3b7e18cad5d49d8e2987b27e0d6c
S: 30858dbf483b61b9838d5c1f853a60beaa4e7afd 127.0.0.1:7005
   slots: (0 slots) slave
   replicates dfa0754c7854a874a6ebd2613b86140ad97701fc
M: d2237fdcfbba672de766b913d1186cebcb6e1761 127.0.0.1:7003
   slots:0-5460 (5461 slots) master
   0 additional replica(s)
M: dfa0754c7854a874a6ebd2613b86140ad97701fc 127.0.0.1:7002
   slots:10923-16383 (5461 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.
[root@web3 7003]#

看了下滓窍，上面有了三個(gè)M 節(jié)點(diǎn)了，果真巩那，7003被選取成了替代7000成為主節(jié)點(diǎn)了吏夯。那我們來(lái)獲取原先存在7000節(jié)點(diǎn)的數(shù)據(jù)：

[root@web3 7003]# redis-cli -c -p 7001
127.0.0.1:7001> get sex
-> Redirected to slot [2584] located at 127.0.0.1:7003
"1"
127.0.0.1:7003>

數(shù)據(jù)果真沒(méi)有丟失，而是從7003上面獲取了即横。

OK噪生。我們?cè)賮?lái)模擬 7000節(jié)點(diǎn)重新啟動(dòng)了的情況，那么它還會(huì)自動(dòng)加入到集群中嗎东囚？那么跺嗽，7000這個(gè)節(jié)點(diǎn)上充當(dāng)什么角色呢？ 我們?cè)囈幌拢?/p>

重新啟動(dòng) 7000 節(jié)點(diǎn)：

[root@web3 7003]# cd ../7000
[root@web3 7000]# ll
total 56
-rw-r--r-- 1 root root   114 Oct 17 08:16 appendonly.aof
-rw-r--r-- 1 root root    43 Oct 17 08:37 dump.rdb
-rw-r--r-- 1 root root   745 Oct 17 08:00 nodes.conf
-rw-r--r-- 1 root root 41550 Oct 17 07:37 redis.conf
[root@web3 7000]# redis-server redis.conf

啟動(dòng)好了页藻，現(xiàn)在桨嫁，再來(lái)檢查一下集群：

redis-trib.rb check 127.0.0.1:700`

[root@web3 7000]# redis-trib.rb check 127.0.0.1:7001
Connecting to node 127.0.0.1:7001: OK
Connecting to node 127.0.0.1:7004: OK
Connecting to node 127.0.0.1:7005: OK
Connecting to node 127.0.0.1:7003: OK
Connecting to node 127.0.0.1:7000: OK
Connecting to node 127.0.0.1:7002: OK
>>> Performing Cluster Check (using node 127.0.0.1:7001)
M: cb5c04b6160c3b7e18cad5d49d8e2987b27e0d6c 127.0.0.1:7001
   slots:5461-10922 (5462 slots) master
   1 additional replica(s)
S: 4b4aef8b48c427a3c903518339d53b6447c58b93 127.0.0.1:7004
   slots: (0 slots) slave
   replicates cb5c04b6160c3b7e18cad5d49d8e2987b27e0d6c
S: 30858dbf483b61b9838d5c1f853a60beaa4e7afd 127.0.0.1:7005
   slots: (0 slots) slave
   replicates dfa0754c7854a874a6ebd2613b86140ad97701fc
M: d2237fdcfbba672de766b913d1186cebcb6e1761 127.0.0.1:7003
   slots:0-5460 (5461 slots) master
   1 additional replica(s)
S: 3707debcbe7be66d4a1968eaf3a5ffaf4308efa4 127.0.0.1:7000
   slots: (0 slots) slave
   replicates d2237fdcfbba672de766b913d1186cebcb6e1761
M: dfa0754c7854a874a6ebd2613b86140ad97701fc 127.0.0.1:7002
   slots:10923-16383 (5461 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.

7000節(jié)點(diǎn)啟動(dòng)起來(lái)了，它卻作為了 70003 的從節(jié)點(diǎn)了份帐。驗(yàn)證了之前的這張圖：

一定要保證有3個(gè)master 節(jié)點(diǎn)璃吧，不然，集群就掛掉了废境。

7肚逸、集群中新加入節(jié)點(diǎn)

我們?cè)賮?lái)測(cè)試一下爷辙，新加入一個(gè)節(jié)點(diǎn)彬坏，分2種情況朦促，1是作為主節(jié)點(diǎn)，2是作為一個(gè)節(jié)點(diǎn)的從節(jié)點(diǎn)栓始。我們分別來(lái)試一下：

(1)新建一個(gè) 7006 節(jié)點(diǎn) 作為一個(gè)新的主節(jié)點(diǎn)加入：

首先就是新建一個(gè) 7006的文件夾和redis.conf:

[root@web3 cluster-test]# cd /usr/local/cluster-test/
[root@web3 cluster-test]# mkdir 7006
[root@web3 cluster-test]# cp 7005/redis.conf  7006/redis.conf
#修改端口
[root@web3 cluster-test]sed -i "s/7005/7006/g" /usr/local/cluster-test/7006/redis.conf

啟動(dòng) 7006

[root@web3 7006]# redis-server redis.conf
[root@web3 7006]# ps -ef|grep redis
root     11384     1  0 07:37 ?        00:00:05 redis-server *:7001 [cluster]
root     11388     1  0 07:37 ?        00:00:05 redis-server *:7002 [cluster]
root     11392     1  0 07:37 ?        00:00:05 redis-server *:7003 [cluster]
root     11396     1  0 07:37 ?        00:00:06 redis-server *:7004 [cluster]
root     11400     1  0 07:37 ?        00:00:05 redis-server *:7005 [cluster]
root     12100     1  0 08:42 ?        00:00:01 redis-server *:7000 [cluster]
root     12132     1  0 09:09 ?        00:00:00 redis-server *:7006 [cluster]
root     12136  8259  0 09:10 pts/0    00:00:00 grep redis

ok,7006 端口已經(jīng)啟動(dòng)好了务冕，并且進(jìn)程也存在了，下面就是加入到集群中幻赚，也是得用到redis-trib.rb命令：

redis-trib.rb add-node 127.0.0.1:7006 127.0.0.1:7000

add-node是加入指令禀忆，127.0.0.1:7006 表示新加入的節(jié)點(diǎn)，127.0.0.1:7000 表示加入的集群的一個(gè)節(jié)點(diǎn)落恼，用來(lái)辨識(shí)是哪個(gè)集群箩退，理論上哪個(gè)都可以。

[root@web3 7006]# redis-trib.rb add-node 127.0.0.1:7006 127.0.0.1:7000
>>> Adding node 127.0.0.1:7006 to cluster 127.0.0.1:7000
Connecting to node 127.0.0.1:7000: OK
Connecting to node 127.0.0.1:7004: OK
Connecting to node 127.0.0.1:7005: OK
Connecting to node 127.0.0.1:7003: OK
Connecting to node 127.0.0.1:7001: OK
Connecting to node 127.0.0.1:7002: OK
>>> Performing Cluster Check (using node 127.0.0.1:7000)
S: 3707debcbe7be66d4a1968eaf3a5ffaf4308efa4 127.0.0.1:7000
   slots: (0 slots) slave
   replicates d2237fdcfbba672de766b913d1186cebcb6e1761
S: 4b4aef8b48c427a3c903518339d53b6447c58b93 127.0.0.1:7004
   slots: (0 slots) slave
   replicates cb5c04b6160c3b7e18cad5d49d8e2987b27e0d6c
S: 30858dbf483b61b9838d5c1f853a60beaa4e7afd 127.0.0.1:7005
   slots: (0 slots) slave
   replicates dfa0754c7854a874a6ebd2613b86140ad97701fc
M: d2237fdcfbba672de766b913d1186cebcb6e1761 127.0.0.1:7003
   slots:0-5460 (5461 slots) master
   1 additional replica(s)
M: cb5c04b6160c3b7e18cad5d49d8e2987b27e0d6c 127.0.0.1:7001
   slots:5461-10922 (5462 slots) master
   1 additional replica(s)
M: dfa0754c7854a874a6ebd2613b86140ad97701fc 127.0.0.1:7002
   slots:10923-16383 (5461 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.
Connecting to node 127.0.0.1:7006: OK
>>> Send CLUSTER MEET to node 127.0.0.1:7006 to make it join the cluster.
[OK] New node added correctly.

這個(gè)命令執(zhí)行完成之后佳谦，它順便檢查來(lái)其他的6個(gè)節(jié)點(diǎn)都是成功的戴涝，最后幾句話：

Connecting to node 127.0.0.1:7006: OK
>>> Send CLUSTER MEET to node 127.0.0.1:7006 to make it join the cluster.
[OK] New node added correctly.

表示新的節(jié)點(diǎn)連接成功了，而且也已經(jīng)加入到集群了钻蔑，我們?cè)賮?lái)檢查一下：

[root@web3 7006]# redis-trib.rb check 127.0.0.1:7006
Connecting to node 127.0.0.1:7006: OK
Connecting to node 127.0.0.1:7001: OK
Connecting to node 127.0.0.1:7004: OK
Connecting to node 127.0.0.1:7000: OK
Connecting to node 127.0.0.1:7002: OK
Connecting to node 127.0.0.1:7005: OK
Connecting to node 127.0.0.1:7003: OK
>>> Performing Cluster Check (using node 127.0.0.1:7006)
M: efc3131fbdc6cf929720e0e0f7136cae85657481 127.0.0.1:7006
   slots: (0 slots) master
   0 additional replica(s)
M: cb5c04b6160c3b7e18cad5d49d8e2987b27e0d6c 127.0.0.1:7001
   slots:5461-10922 (5462 slots) master
   1 additional replica(s)
S: 4b4aef8b48c427a3c903518339d53b6447c58b93 127.0.0.1:7004
   slots: (0 slots) slave
   replicates cb5c04b6160c3b7e18cad5d49d8e2987b27e0d6c
S: 3707debcbe7be66d4a1968eaf3a5ffaf4308efa4 127.0.0.1:7000
   slots: (0 slots) slave
   replicates d2237fdcfbba672de766b913d1186cebcb6e1761
M: dfa0754c7854a874a6ebd2613b86140ad97701fc 127.0.0.1:7002
   slots:10923-16383 (5461 slots) master
   1 additional replica(s)
S: 30858dbf483b61b9838d5c1f853a60beaa4e7afd 127.0.0.1:7005
   slots: (0 slots) slave
   replicates dfa0754c7854a874a6ebd2613b86140ad97701fc
M: d2237fdcfbba672de766b913d1186cebcb6e1761 127.0.0.1:7003
   slots:0-5460 (5461 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.

可以看到了啥刻，7006 已經(jīng)成為了主節(jié)點(diǎn)。

我們也可以連接到客戶端后咪笑，來(lái)看這個(gè)集群節(jié)點(diǎn)情況：

[root@web3 7006]# redis-cli -c -p 7006
127.0.0.1:7006> cluster nodes
cb5c04b6160c3b7e18cad5d49d8e2987b27e0d6c 127.0.0.1:7001 master - 0 1445073797986 2 connected 5461-10922
4b4aef8b48c427a3c903518339d53b6447c58b93 127.0.0.1:7004 slave cb5c04b6160c3b7e18cad5d49d8e2987b27e0d6c 0 1445073799497 2 connected
efc3131fbdc6cf929720e0e0f7136cae85657481 127.0.0.1:7006 myself,master - 0 0 0 connected
3707debcbe7be66d4a1968eaf3a5ffaf4308efa4 127.0.0.1:7000 slave d2237fdcfbba672de766b913d1186cebcb6e1761 0 1445073797482 7 connected
dfa0754c7854a874a6ebd2613b86140ad97701fc 127.0.0.1:7002 master - 0 1445073798489 3 connected 10923-16383
30858dbf483b61b9838d5c1f853a60beaa4e7afd 127.0.0.1:7005 slave dfa0754c7854a874a6ebd2613b86140ad97701fc 0 1445073798993 3 connected
d2237fdcfbba672de766b913d1186cebcb6e1761 127.0.0.1:7003 master - 0 1445073799498 7 connected 0-5460
127.0.0.1:7006>

可以看到 有7個(gè)節(jié)點(diǎn)可帽。7006 也作為了master節(jié)點(diǎn)，但是窗怒，但是映跟，你看看后面的：

efc3131fbdc6cf929720e0e0f7136cae85657481 127.0.0.1:7006
   slots: (0 slots) master

可以看到 o slots 也就是說(shuō)，雖然它現(xiàn)在是主節(jié)點(diǎn)扬虚，但是努隙，缺沒(méi)有分配任何節(jié)點(diǎn)給它，也就是它現(xiàn)在還不負(fù)責(zé)數(shù)據(jù)存取孔轴。那加上去有毛用疤攴ā！Ｂ酚ァ４蕖！

看來(lái)晋柱，redis cluster 不是在新加節(jié)點(diǎn)的時(shí)候幫我們做好了遷移工作优构，需要我們手動(dòng)對(duì)集群進(jìn)行重新分片遷移，也是這個(gè)命令：

redis-trib.rb reshard 127.0.0.1:7000

這個(gè)命令是用來(lái)遷移slot節(jié)點(diǎn)的雁竞，后面的127.0.0.1:7000是表示是哪個(gè)集群钦椭，端口填［7000-7006］都可以拧额，我們運(yùn)行下：

[root@web3 7006]# redis-trib.rb reshard 127.0.0.1:7000
Connecting to node 127.0.0.1:7006: OK
Connecting to node 127.0.0.1:7001: OK
Connecting to node 127.0.0.1:7004: OK
Connecting to node 127.0.0.1:7000: OK
Connecting to node 127.0.0.1:7002: OK
Connecting to node 127.0.0.1:7005: OK
Connecting to node 127.0.0.1:7003: OK
>>> Performing Cluster Check (using node 127.0.0.1:7006)
M: efc3131fbdc6cf929720e0e0f7136cae85657481 127.0.0.1:7006
   slots: (0 slots) master
   0 additional replica(s)
M: cb5c04b6160c3b7e18cad5d49d8e2987b27e0d6c 127.0.0.1:7001
   slots:5461-10922 (5462 slots) master
   1 additional replica(s)
S: 4b4aef8b48c427a3c903518339d53b6447c58b93 127.0.0.1:7004
   slots: (0 slots) slave
   replicates cb5c04b6160c3b7e18cad5d49d8e2987b27e0d6c
S: 3707debcbe7be66d4a1968eaf3a5ffaf4308efa4 127.0.0.1:7000
   slots: (0 slots) slave
   replicates d2237fdcfbba672de766b913d1186cebcb6e1761
M: dfa0754c7854a874a6ebd2613b86140ad97701fc 127.0.0.1:7002
   slots:10923-16383 (5461 slots) master
   1 additional replica(s)
S: 30858dbf483b61b9838d5c1f853a60beaa4e7afd 127.0.0.1:7005
   slots: (0 slots) slave
   replicates dfa0754c7854a874a6ebd2613b86140ad97701fc
M: d2237fdcfbba672de766b913d1186cebcb6e1761 127.0.0.1:7003
   slots:0-5460 (5461 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.
How many slots do you want to move (from 1 to 16384)?

它提示我們需要遷移多少slot到7006上，我們可以算一下：16384/4 = 4096彪腔，也就是說(shuō)侥锦，為了平衡分配起見，我們需要移動(dòng)4096個(gè)槽點(diǎn)到7006上德挣。

好恭垦，那輸入4096:

How many slots do you want to move (from 1 to 16384)? 4096
What is the receiving node ID?

它又提示我們，接受的node ID是多少格嗅，7006的id 我們通過(guò)上面就可以看到是efc3131fbdc6cf929720e0e0f7136cae85657481 :

What is the receiving node ID? efc3131fbdc6cf929720e0e0f7136cae85657481
Please enter all the source node IDs.
  Type 'all' to use all the nodes as source nodes for the hash slots.
  Type 'done' once you entered all the source nodes IDs.
Source node #1:

接著番挺， redis-trib 會(huì)向你詢問(wèn)重新分片的源節(jié)點(diǎn)（source node）， 也即是屯掖， 要從哪個(gè)節(jié)點(diǎn)中取出 4096 個(gè)哈希槽玄柏， 并將這些槽移動(dòng)到7006節(jié)點(diǎn)上面。

如果我們不打算從特定的節(jié)點(diǎn)上取出指定數(shù)量的哈希槽贴铜， 那么可以向 redis-trib 輸入 all 粪摘， 這樣的話， 集群中的所有主節(jié)點(diǎn)都會(huì)成為源節(jié)點(diǎn)阀湿， redis-trib 將從各個(gè)源節(jié)點(diǎn)中各取出一部分哈希槽赶熟， 湊夠 4096 個(gè)， 然后移動(dòng)到7006節(jié)點(diǎn)上：

Source node #1:all

接下來(lái)就開始遷移了陷嘴，并且會(huì)詢問(wèn)你是否確認(rèn)：

Moving slot 1359 from d2237fdcfbba672de766b913d1186cebcb6e1761
    Moving slot 1360 from d2237fdcfbba672de766b913d1186cebcb6e1761
    Moving slot 1361 from d2237fdcfbba672de766b913d1186cebcb6e1761
    Moving slot 1362 from d2237fdcfbba672de766b913d1186cebcb6e1761
    Moving slot 1363 from d2237fdcfbba672de766b913d1186cebcb6e1761
    Moving slot 1364 from d2237fdcfbba672de766b913d1186cebcb6e1761
Do you want to proceed with the proposed reshard plan (yes/no)?

輸入 yes 并使用按下回車之后映砖， redis-trib 就會(huì)正式開始執(zhí)行重新分片操作， 將指定的哈希槽從源節(jié)點(diǎn)一個(gè)個(gè)地移動(dòng)到7006節(jié)點(diǎn)上面灾挨。

遷移完畢之后邑退，我們來(lái)檢查下：

[root@web3 7006]# redis-trib.rb check 127.0.0.1:7000
Connecting to node 127.0.0.1:7000: OK
Connecting to node 127.0.0.1:7006: OK
Connecting to node 127.0.0.1:7004: OK
Connecting to node 127.0.0.1:7005: OK
Connecting to node 127.0.0.1:7003: OK
Connecting to node 127.0.0.1:7001: OK
Connecting to node 127.0.0.1:7002: OK
>>> Performing Cluster Check (using node 127.0.0.1:7000)
S: 3707debcbe7be66d4a1968eaf3a5ffaf4308efa4 127.0.0.1:7000
   slots: (0 slots) slave
   replicates d2237fdcfbba672de766b913d1186cebcb6e1761
M: efc3131fbdc6cf929720e0e0f7136cae85657481 127.0.0.1:7006
   slots:0-1364,5461-6826,10923-12287 (4096 slots) master
   0 additional replica(s)
S: 4b4aef8b48c427a3c903518339d53b6447c58b93 127.0.0.1:7004
   slots: (0 slots) slave
   replicates cb5c04b6160c3b7e18cad5d49d8e2987b27e0d6c
S: 30858dbf483b61b9838d5c1f853a60beaa4e7afd 127.0.0.1:7005
   slots: (0 slots) slave
   replicates dfa0754c7854a874a6ebd2613b86140ad97701fc
M: d2237fdcfbba672de766b913d1186cebcb6e1761 127.0.0.1:7003
   slots:1365-5460 (4096 slots) master
   1 additional replica(s)
M: cb5c04b6160c3b7e18cad5d49d8e2987b27e0d6c 127.0.0.1:7001
   slots:6827-10922 (4096 slots) master
   1 additional replica(s)
M: dfa0754c7854a874a6ebd2613b86140ad97701fc 127.0.0.1:7002
   slots:12288-16383 (4096 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.

我們著重看7006:

0-1364,5461-6826,10923-12287 (4096 slots)

這些原來(lái)在其他節(jié)點(diǎn)上的slot 杯遷移到了7006上。原來(lái)劳澄，它只是間隔的移動(dòng)地技，并不是銜接的整體移動(dòng)，我們看下有數(shù)據(jù)了沒(méi)秒拔？

[root@web3 7006]# redis-cli -c -p 7006
127.0.0.1:7006> keys *
1) "city"
2) "age"
3) "citylocate"
127.0.0.1:7006> get city
"shanghai"

非常贊莫矗，已經(jīng)有數(shù)據(jù)了。

(2)新建一個(gè) 7007從節(jié)點(diǎn)砂缩，作為7006的從節(jié)點(diǎn)

我們?cè)傩陆ㄒ粋€(gè)節(jié)點(diǎn)7007作谚，步驟類似，就先省略了庵芭。建好后妹懒，啟動(dòng)起來(lái)，我們看如何把它加入到集群中的從節(jié)點(diǎn)中：

[root@web3 7007]# redis-trib.rb add-node --slave 127.0.0.1:7007 127.0.0.1:7000

add-node的時(shí)候加上--slave表示是加入到從節(jié)點(diǎn)中双吆，但是這樣加眨唬，是隨機(jī)的会前。這里的命令行完全像我們?cè)谔砑右粋€(gè)新主服務(wù)器時(shí)使用的一樣，所以我們沒(méi)有指定要給哪個(gè)主服 務(wù)器添加副本匾竿。這種情況下瓦宜，redis-trib 會(huì)將7007作為一個(gè)具有較少副本的隨機(jī)的主服務(wù)器的副本。

它會(huì)作為誰(shuí)的從節(jié)點(diǎn)搂橙，應(yīng)該是7006歉提，因?yàn)?006還沒(méi)有從節(jié)點(diǎn)。我們運(yùn)行下

[root@web3 7007]# redis-trib.rb add-node --slave 127.0.0.1:7007 127.0.0.1:7000
...
...
[OK] All 16384 slots covered.
Automatically selected master 127.0.0.1:7006
Connecting to node 127.0.0.1:7007: OK
>>> Send CLUSTER MEET to node 127.0.0.1:7007 to make it join the cluster.
Waiting for the cluster to join.
>>> Configure node as replica of 127.0.0.1:7006.
[OK] New node added correctly.

上面提示說(shuō)区转，自動(dòng)選擇了7006作為master節(jié)點(diǎn)。并且成功了版扩。我們檢查下：

[root@web3 7007]# redis-trib.rb check 127.0.0.1:7000
Connecting to node 127.0.0.1:7000: OK
Connecting to node 127.0.0.1:7006: OK
Connecting to node 127.0.0.1:7004: OK
Connecting to node 127.0.0.1:7005: OK
Connecting to node 127.0.0.1:7003: OK
Connecting to node 127.0.0.1:7001: OK
Connecting to node 127.0.0.1:7007: OK
Connecting to node 127.0.0.1:7002: OK
>>> Performing Cluster Check (using node 127.0.0.1:7000)
S: 3707debcbe7be66d4a1968eaf3a5ffaf4308efa4 127.0.0.1:7000
   slots: (0 slots) slave
   replicates d2237fdcfbba672de766b913d1186cebcb6e1761
M: efc3131fbdc6cf929720e0e0f7136cae85657481 127.0.0.1:7006
   slots:0-1364,5461-6826,10923-12287 (4096 slots) master
   1 additional replica(s)
S: 4b4aef8b48c427a3c903518339d53b6447c58b93 127.0.0.1:7004
   slots: (0 slots) slave
   replicates cb5c04b6160c3b7e18cad5d49d8e2987b27e0d6c
S: 30858dbf483b61b9838d5c1f853a60beaa4e7afd 127.0.0.1:7005
   slots: (0 slots) slave
   replicates dfa0754c7854a874a6ebd2613b86140ad97701fc
M: d2237fdcfbba672de766b913d1186cebcb6e1761 127.0.0.1:7003
   slots:1365-5460 (4096 slots) master
   1 additional replica(s)
M: cb5c04b6160c3b7e18cad5d49d8e2987b27e0d6c 127.0.0.1:7001
   slots:6827-10922 (4096 slots) master
   1 additional replica(s)
S: 86d05e7c2b197dc182b5e71069e791d033cf899e 127.0.0.1:7007
   slots: (0 slots) slave
   replicates efc3131fbdc6cf929720e0e0f7136cae85657481
M: dfa0754c7854a874a6ebd2613b86140ad97701fc 127.0.0.1:7002
   slots:12288-16383 (4096 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.

果然废离，7007加入到了7006的從節(jié)點(diǎn)當(dāng)中礁芦。

你說(shuō)蜻韭，我如果想指定一個(gè)主節(jié)點(diǎn)行不行？當(dāng)然可以柿扣。我們?cè)俳ㄒ粋€(gè)7008節(jié)點(diǎn)肖方。

redis-trib.rb add-node --slave --master-id efc3131fbdc6cf929720e0e0f7136cae85657481 127.0.0.1:7008 127.0.0.1:7000

--master-id 表示指定的主節(jié)點(diǎn)node id。這里指定的是 7006 這個(gè)主節(jié)點(diǎn)未状。

Waiting for the cluster to join.
>>> Configure node as replica of 127.0.0.1:7006.
[OK] New node added correctly.
[root@web3 7008]#

提示我們已經(jīng)作為7006的附屬品俯画，也就是加入到7006的從節(jié)點(diǎn)來(lái)了，照這么說(shuō)司草，7006就有2個(gè)從節(jié)點(diǎn)了艰垂，我們看一下：

redis-cli -c -p 7008 cluster nodes |grep efc3131fbdc6cf929720e0e0f7136cae85657481
86d05e7c2b197dc182b5e71069e791d033cf899e 127.0.0.1:7007 slave efc3131fbdc6cf929720e0e0f7136cae85657481 0 1445089507786 8 connected
efc3131fbdc6cf929720e0e0f7136cae85657481 127.0.0.1:7006 master - 0 1445089508289 8 connected 0-1364 5461-6826 10923-12287
44321e7d619410dc4e0a8745366610a0d06d2395 127.0.0.1:7008 myself,slave efc3131fbdc6cf929720e0e0f7136cae85657481 0 0 0 connected

我們過(guò)濾了下看結(jié)果，果真埋虹，7007和7008是7006的從節(jié)點(diǎn)了猜憎。

剛好，我們?cè)僮鲆粋€(gè)實(shí)驗(yàn)搔课，我把7006的殺掉胰柑，看7007和7008誰(shuí)會(huì)變成主節(jié)點(diǎn)：

[root@web3 7008]# ps -ef|grep redis
root     11384     1  0 09:56 ?        00:00:16 redis-server *:7001 [cluster]
root     11388     1  0 09:56 ?        00:00:16 redis-server *:7002 [cluster]
root     11392     1  0 09:56 ?        00:00:16 redis-server *:7003 [cluster]
root     11396     1  0 09:56 ?        00:00:15 redis-server *:7004 [cluster]
root     11400     1  0 09:56 ?        00:00:15 redis-server *:7005 [cluster]
root     12100     1  0 11:01 ?        00:00:11 redis-server *:7000 [cluster]
root     12132     1  0 11:28 ?        00:00:11 redis-server *:7006 [cluster]
root     12202     1  0 13:14 ?        00:00:02 redis-server *:7007 [cluster]
root     12219     1  0 13:39 ?        00:00:00 redis-server *:7008 [cluster]
root     12239  8259  0 13:49 pts/0    00:00:00 grep redis
[root@web3 7008]# kill 12132
[root@web3 7008]# redis-cli -c -p 7008
127.0.0.1:7008> get ss5rtr
-> Redirected to slot [1188] located at 127.0.0.1:7007
"66"
127.0.0.1:7007> cluster nodes
efc3131fbdc6cf929720e0e0f7136cae85657481 127.0.0.1:7006 master,fail - 1445089780668 1445089779963 8 disconnected
d2237fdcfbba672de766b913d1186cebcb6e1761 127.0.0.1:7003 master - 0 1445089812195 7 connected 1365-5460
30858dbf483b61b9838d5c1f853a60beaa4e7afd 127.0.0.1:7005 slave dfa0754c7854a874a6ebd2613b86140ad97701fc 0 1445089813710 3 connected
86d05e7c2b197dc182b5e71069e791d033cf899e 127.0.0.1:7007 myself,master - 0 0 10 connected 0-1364 5461-6826 10923-12287
cb5c04b6160c3b7e18cad5d49d8e2987b27e0d6c 127.0.0.1:7001 master - 0 1445089814214 2 connected 6827-10922
4b4aef8b48c427a3c903518339d53b6447c58b93 127.0.0.1:7004 slave cb5c04b6160c3b7e18cad5d49d8e2987b27e0d6c 0 1445089812701 2 connected
44321e7d619410dc4e0a8745366610a0d06d2395 127.0.0.1:7008 slave 86d05e7c2b197dc182b5e71069e791d033cf899e 0 1445089814214 10 connected
3707debcbe7be66d4a1968eaf3a5ffaf4308efa4 127.0.0.1:7000 slave d2237fdcfbba672de766b913d1186cebcb6e1761 0 1445089813204 7 connected
dfa0754c7854a874a6ebd2613b86140ad97701fc 127.0.0.1:7002 master - 0 1445089813204 3 connected 12288-16383
127.0.0.1:7007>

看，7007獲得了成為主節(jié)點(diǎn)的機(jī)會(huì)爬泥，7008就變成了7007的從節(jié)點(diǎn)柬讨。

那么這個(gè)時(shí)候，重啟7006節(jié)點(diǎn)急灭，那么他就會(huì)變成了一個(gè)7007的從節(jié)點(diǎn)了姐浮。

27.0.0.1:7008> cluster nodes
30858dbf483b61b9838d5c1f853a60beaa4e7afd 127.0.0.1:7005 slave dfa0754c7854a874a6ebd2613b86140ad97701fc 0 1445089986148 3 connected
86d05e7c2b197dc182b5e71069e791d033cf899e 127.0.0.1:7007 master - 0 1445089986652 10 connected 0-1364 5461-6826 10923-12287
d2237fdcfbba672de766b913d1186cebcb6e1761 127.0.0.1:7003 master - 0 1445089986148 7 connected 1365-5460
cb5c04b6160c3b7e18cad5d49d8e2987b27e0d6c 127.0.0.1:7001 master - 0 1445089987155 2 connected 6827-10922
efc3131fbdc6cf929720e0e0f7136cae85657481 127.0.0.1:7006 slave 86d05e7c2b197dc182b5e71069e791d033cf899e 0 1445089985644 10 connected
44321e7d619410dc4e0a8745366610a0d06d2395 127.0.0.1:7008 myself,slave 86d05e7c2b197dc182b5e71069e791d033cf899e 0 0 0 connected
dfa0754c7854a874a6ebd2613b86140ad97701fc 127.0.0.1:7002 master - 0 1445089986652 3 connected 12288-16383
4b4aef8b48c427a3c903518339d53b6447c58b93 127.0.0.1:7004 slave cb5c04b6160c3b7e18cad5d49d8e2987b27e0d6c 0 1445089987660 2 connected
3707debcbe7be66d4a1968eaf3a5ffaf4308efa4 127.0.0.1:7000 slave d2237fdcfbba672de766b913d1186cebcb6e1761 0 1445089985644 7 connected
127.0.0.1:7008>

(3)移除一個(gè)主節(jié)點(diǎn)

有加肯定有減，redis cluster同樣支持移除節(jié)點(diǎn)功能葬馋，同樣也是redis-trib.rb的用法：

redis-trib del-node 127.0.0.1:7000 `<node-id>`

和新加節(jié)點(diǎn)有點(diǎn)不同的是卖鲤，移除需要節(jié)點(diǎn)的node-id肾扰。那我們嘗試將7007這個(gè)主節(jié)點(diǎn)移除：

[root@web3 7006]# redis-trib.rb del-node 127.0.0.1:7000 86d05e7c2b197dc182b5e71069e791d033cf899e
>>> Removing node 86d05e7c2b197dc182b5e71069e791d033cf899e from cluster 127.0.0.1:7000
Connecting to node 127.0.0.1:7000: OK
Connecting to node 127.0.0.1:7006: OK
Connecting to node 127.0.0.1:7004: OK
Connecting to node 127.0.0.1:7001: OK
Connecting to node 127.0.0.1:7003: OK
Connecting to node 127.0.0.1:7007: OK
Connecting to node 127.0.0.1:7008: OK
Connecting to node 127.0.0.1:7005: OK
Connecting to node 127.0.0.1:7002: OK
[ERR] Node 127.0.0.1:7007 is not empty! Reshard data away and try again.

報(bào)錯(cuò)了，它提示我們說(shuō)蛋逾，由于7007里面已經(jīng)有數(shù)據(jù)了集晚，不能被移除，要先將它的數(shù)據(jù)轉(zhuǎn)移出去区匣。也就是說(shuō)得重新分片偷拔，用上面增加新節(jié)點(diǎn)后的分片方式一樣，用我們?cè)賮?lái)一遍：

redis-trib.rb reshard 127.0.0.1:7000

由于中間太多內(nèi)容亏钩，就省略不重要的莲绰，只簡(jiǎn)單說(shuō)明下關(guān)鍵的幾步：

M: 86d05e7c2b197dc182b5e71069e791d033cf899e 127.0.0.1:7007
   slots:0-1364,5461-6826,10923-12287 (4096 slots) master
How many slots do you want to move (from 1 to 16384)?

提示，我們要分多少個(gè)槽點(diǎn)姑丑，由于7007上有4096個(gè)槽點(diǎn)蛤签，所以這里填寫4096

How many slots do you want to move (from 1 to 16384)? 4096
What is the receiving node ID?

提示我們，需要移動(dòng)到哪個(gè)id上栅哀，那就填7001的吧：

What is the receiving node ID? cb5c04b6160c3b7e18cad5d49d8e2987b27e0d6c
Please enter all the source node IDs.
  Type 'all' to use all the nodes as source nodes for the hash slots.
  Type 'done' once you entered all the source nodes IDs.
Source node #1:

這里就是關(guān)鍵了震肮，它要我們從哪個(gè)節(jié)點(diǎn)去轉(zhuǎn)移數(shù)據(jù)到7001，因?yàn)槲覀兪且獎(jiǎng)h除7007的留拾，所以戳晌，我們就得7007的id了

Source node #1:86d05e7c2b197dc182b5e71069e791d033cf899e
Source node #2:done
Do you want to proceed with the proposed reshard plan (yes/no)? yes

ok,這樣就遷移好了。我們看看7007是否為空了：

[root@web3 7006]# redis-cli -c -p 7007
127.0.0.1:7007> keys *
(empty list or set)
127.0.0.1:7007> cluster nodes
86d05e7c2b197dc182b5e71069e791d033cf899e 127.0.0.1:7007 myself,master - 0 0 10 connected

果然為空了痴柔，好÷儋耍現(xiàn)在再進(jìn)行移除節(jié)點(diǎn)操作：

[root@web3 7006]# redis-trib.rb del-node 127.0.0.1:7000 86d05e7c2b197dc182b5e71069e791d033cf899e
>>> Removing node 86d05e7c2b197dc182b5e71069e791d033cf899e from cluster 127.0.0.1:7000
Connecting to node 127.0.0.1:7000: OK
Connecting to node 127.0.0.1:7006: OK
Connecting to node 127.0.0.1:7004: OK
Connecting to node 127.0.0.1:7001: OK
Connecting to node 127.0.0.1:7003: OK
Connecting to node 127.0.0.1:7007: OK
Connecting to node 127.0.0.1:7008: OK
Connecting to node 127.0.0.1:7005: OK
Connecting to node 127.0.0.1:7002: OK
>>> Sending CLUSTER FORGET messages to the cluster...
>>> 127.0.0.1:7006 as replica of 127.0.0.1:7001
>>> 127.0.0.1:7008 as replica of 127.0.0.1:7001
>>> SHUTDOWN the node.

刪除成功，而且還很人性化的將7006和70082個(gè)沒(méi)了爹的孩子送給了7001竞帽。好評(píng)??

我們?cè)贆z查一下：7007 已經(jīng)移除扛施，連不上了。

[root@web3 7006]# redis-trib.rb check 127.0.0.1:7007
Connecting to node 127.0.0.1:7007: [ERR] Sorry, can't connect to node 127.0.0.1:7007

[root@web3 7006]# redis-trib.rb check 127.0.0.1:7008
Connecting to node 127.0.0.1:7008: OK
Connecting to node 127.0.0.1:7005: OK
Connecting to node 127.0.0.1:7003: OK
Connecting to node 127.0.0.1:7001: OK
Connecting to node 127.0.0.1:7006: OK
Connecting to node 127.0.0.1:7002: OK
Connecting to node 127.0.0.1:7004: OK
Connecting to node 127.0.0.1:7000: OK
>>> Performing Cluster Check (using node 127.0.0.1:7008)
S: 44321e7d619410dc4e0a8745366610a0d06d2395 127.0.0.1:7008
   slots: (0 slots) slave
   replicates cb5c04b6160c3b7e18cad5d49d8e2987b27e0d6c
S: 30858dbf483b61b9838d5c1f853a60beaa4e7afd 127.0.0.1:7005
   slots: (0 slots) slave
   replicates dfa0754c7854a874a6ebd2613b86140ad97701fc
M: d2237fdcfbba672de766b913d1186cebcb6e1761 127.0.0.1:7003
   slots:1365-5460 (4096 slots) master
   1 additional replica(s)
M: cb5c04b6160c3b7e18cad5d49d8e2987b27e0d6c 127.0.0.1:7001
   slots:0-1364,5461-12287 (8192 slots) master
   3 additional replica(s)
S: efc3131fbdc6cf929720e0e0f7136cae85657481 127.0.0.1:7006
   slots: (0 slots) slave
   replicates cb5c04b6160c3b7e18cad5d49d8e2987b27e0d6c
M: dfa0754c7854a874a6ebd2613b86140ad97701fc 127.0.0.1:7002
   slots:12288-16383 (4096 slots) master
   1 additional replica(s)
S: 4b4aef8b48c427a3c903518339d53b6447c58b93 127.0.0.1:7004
   slots: (0 slots) slave
   replicates cb5c04b6160c3b7e18cad5d49d8e2987b27e0d6c
S: 3707debcbe7be66d4a1968eaf3a5ffaf4308efa4 127.0.0.1:7000
   slots: (0 slots) slave
   replicates d2237fdcfbba672de766b913d1186cebcb6e1761
[OK] All nodes agree about slots configuration.
>>> Check for open slots...
>>> Check slots coverage...
[OK] All 16384 slots covered.
[root@web3 7006]#

7006和7008果然也成功加入到了7001的大家庭屹篓。

(4)移除一個(gè)從節(jié)點(diǎn)

移除一個(gè)從節(jié)點(diǎn)就簡(jiǎn)單的多了疙渣，因?yàn)椴恍枰紤]數(shù)據(jù)的遷移，我們7008給移除：

[root@web3 7006]# redis-trib.rb del-node 127.0.0.1:7005 44321e7d619410dc4e0a8745366610a0d06d2395
>>> Removing node 44321e7d619410dc4e0a8745366610a0d06d2395 from cluster 127.0.0.1:7005
Connecting to node 127.0.0.1:7005: OK
Connecting to node 127.0.0.1:7001: OK
Connecting to node 127.0.0.1:7002: OK
Connecting to node 127.0.0.1:7004: OK
Connecting to node 127.0.0.1:7000: OK
Connecting to node 127.0.0.1:7006: OK
Connecting to node 127.0.0.1:7008: OK
Connecting to node 127.0.0.1:7003: OK
>>> Sending CLUSTER FORGET messages to the cluster...
>>> SHUTDOWN the node.
[root@web3 7006]# redis-trib.rb check 127.0.0.1:7008
Connecting to node 127.0.0.1:7008: [ERR] Sorry, can't connect to node 127.0.0.1:7008

移除成功堆巧。