一、准備工作
1.系統環境：centos6.4
2.服務器六台（1主5從）：
192.168.1.161（master）
192.168.1.162（slave）
192.168.1.163（slave）
192.168.1.141（slave）
192.168.1.142（slave）
192.168.1.143（slave）
2.redis版本：5.0.3
3.安裝：
進入到目錄：cd /usr/local
下載redis：wget http://download.redis.io/releases/redis-5.0.3.tar.gz
下載完成后解壓：tar zxvf redis-5.0.3.tar.gz
重命名為redis文件夾（這一步純屬個人喜好）：mv redis-5.0.3 redis
進入到redis文件夾：cd redis
編譯及安裝：make && make install
特別說明：官方文檔只給出了make（編譯），沒有給出make install（安裝）
二、配置
1.Master配置，把master的redis.conf拷貝了一份命名為redis_master.conf
設置：
### NETWORK 設置：
# bind 127.0.0.1 //綁定監控的網卡IP，注釋掉bind，任何ip均可訪問，
protected-mode no //保護模式關閉，這里使用密碼訪問
port 17000 //設置端口，建議測試時可以使用默認端口，我這里改掉了，建議生產環境均使用自定義端口
timeout 30 //Client 端空閑斷開連接的時間
### GENERAL 設置：
daemonize yes //后台模式運行
pidfile /var/run/redis_17000.pid //pid進程文件名
logfile /usr/local/redis/logs/redis.log //日志文件的位置
### SNAPSHOTTING 設置：
dir /usr/local/redis/datas //快照文件的路徑
### APPEND ONLY MODE 設置：
appendonly yes //默認值是No，意思是不使用AOF增量持久化的方式，使用RDB全量持久化的方式。把No值改成Yes，使用AOF增量持久化的方式
appendfsync always
###SECURITY 設置密碼:，生產環境一定要使用復雜密碼
requirepass 123456
2.Slave配置，我把master的redis.conf拷貝了一份命名為redis_slave.conf
### NETWORK 設置：
# bind 127.0.0.1 //注釋掉bind，任何ip均可訪問
port 17000 //設置端口
protected-mode no //保護模式關閉，使用密碼訪問
timeout 30 //Client 端空閑斷開連接的時間
### GENERAL 設置：
daemonize yes //后台模式運行
pidfile /var/run/redis_17000.pid
logfile /usr/local/redis/logs/redis.log //日志文件的位置
### SNAPSHOTTING 設置：
dir /usr/local/redis/datas //SNAPSHOTTING文件的路徑
### REPLICATION 設置：
replicaof 192.168.1.161 17000 //主服務器的Ip地址和Port端口號
replica-serve-stale-data no //如果slave 無法與master 同步，設置成slave不可讀，方便監控腳本發現問題。
masterauth 123456 //master的密碼
### APPEND ONLY MODE 設置：
appendonly yes //默認值是No，意思是不使用AOF增量持久化的方式，使用RDB全量持久化的方式。把No值改成Yes，使用AOF增量持久化的方式
appendfsync always
###SECURITY 設置密碼:，生產環境一定要使用復雜密碼
requirepass 123456
3.Sentinel（哨兵）配置，配置文件為sentinel.conf
port 16000 //哨兵端口號
protected-mode no //關閉保護模式
daemonize yes //守護進程
dir /usr/local/redis/sentinel/ //哨兵程序的工作路徑
sentinel auth-pass mymaster 123456 //master的訪問密碼
//Sentinel去監視一個名為mymaster的主redis實例，這個主實例的IP地址為本機地址192.168.1.161，端口號為17000，而將這個主實例判斷為失效至少需要1個 Sentinel進程的同意，只要同意Sentinel的數量不達標，自動failover就不會執行
sentinel monitor mymaster 192.168.1.161 17000 1
sentinel down-after-milliseconds mymaster 5000 //哨兵程序每5秒檢測一次Master是否正常
//指定了在執行故障轉移時，最多可以有多少個從Redis實例在同步新的主實例，在從Redis實例較多的情況下這個數字越小，同步的時間越長，完成故障轉移所需的時間就越長
sentinel parallel-syncs mymaster 2
sentinel failover-timeout mymaster 300000 //如果在該時間（ms）內未能完成failover操作，則認為該failover失敗，生產環境需要根據數據量設置該值
三、啟動Redis服務
1.在各服務器上先建立配置文件中需要的文件夾
mkdir -p /usr/local/redis/datas
mkdir -p /usr/local/redis/logs
mkdir -p /usr/local/redis/sentinel
2.啟動服務
先開啟master服務，也就是192.168.1.161服務器
進入到redis文件夾中：cd /usr/local/redis
### 啟動master:
src/redis-server redis_master.conf
注意： 這里指定了配置文件redis_master.conf
使用redis-cli訪問服務端，查看狀態
src/redis-cli -p 17000 //由於修改了端口，所以在使用redis-cli工具時，需要指定端口
看到127.0.0.1:17000> 表示連接到服務端
輸入info replication 查看集群狀態，這里報錯：NOAUTH Authentication required.
原因是：master和slave均設置了密碼，如果要查看replication，則需要輸入密碼授權：
AUTH 123456 --密碼就是123456（配置文件里配置的），看到OK表示授權通過，再次輸入info replication便可以顯示集群狀態：
[root@mongodb-161 redis]# src/redis-server redis_master.conf [root@mongodb-161 redis]# src/redis-cli -p 17000 127.0.0.1:17000> info replication NOAUTH Authentication required. 127.0.0.1:17000> AUTH 123456 OK 127.0.0.1:17000> info replication # Replication role:master connected_slaves:0 master_replid:96942e0268bff14dba3b6d1e22a9b55e9e77343b master_replid2:0000000000000000000000000000000000000000 master_repl_offset:0 second_repl_offset:-1 repl_backlog_active:0 repl_backlog_size:1048576 repl_backlog_first_byte_offset:0 repl_backlog_histlen:0 127.0.0.1:17000>
這里只是剛啟動了master一個服務，所以connected_slaves:0
再啟動192.168.1.162服務器上的redis
src/redis-server redis_slave.conf
注意：這里指定了配置文件redis_slave.conf
啟動完成后使用redis-cli -p 17000 查看狀態
[root@mongodb-162 redis]# src/redis-server redis_slave.conf [root@mongodb-162 redis]# src/redis-cli -p 17000 127.0.0.1:17000> AUTH 123456 OK 127.0.0.1:17000> info replication # Replication role:slave master_host:192.168.1.161 master_port:17000 master_link_status:up master_last_io_seconds_ago:9 master_sync_in_progress:0 slave_repl_offset:28 slave_priority:100 slave_read_only:1 connected_slaves:0 master_replid:b3f15ec093bb8ab2b1e908b41740f4bdabbfeba4 master_replid2:0000000000000000000000000000000000000000 master_repl_offset:28 second_repl_offset:-1 repl_backlog_active:1 repl_backlog_size:1048576 repl_backlog_first_byte_offset:1 repl_backlog_histlen:28 127.0.0.1:17000>
可以看到當前的狀態是slave，連接的master是192.168.1.161，這里再來查看192.168.1.161的master上的狀態
127.0.0.1:17000> info replication # Replication role:master connected_slaves:1 slave0:ip=192.168.1.162,port=17000,state=online,offset=14,lag=1 master_replid:b3f15ec093bb8ab2b1e908b41740f4bdabbfeba4 master_replid2:0000000000000000000000000000000000000000 master_repl_offset:14 second_repl_offset:-1 repl_backlog_active:1 repl_backlog_size:1048576 repl_backlog_first_byte_offset:1 repl_backlog_histlen:14 127.0.0.1:17000>
可以看到連接到master有一個slave了，繼續把剩余服務器上的redis都啟動起來，然后再查看master狀態
127.0.0.1:17000> info replication # Replication role:master connected_slaves:5 slave0:ip=192.168.1.162,port=17000,state=online,offset=308,lag=1 slave1:ip=192.168.1.163,port=17000,state=online,offset=308,lag=0 slave2:ip=192.168.1.141,port=17000,state=online,offset=308,lag=1 slave3:ip=192.168.1.142,port=17000,state=online,offset=308,lag=1 slave4:ip=192.168.1.143,port=17000,state=online,offset=308,lag=0 master_replid:b3f15ec093bb8ab2b1e908b41740f4bdabbfeba4 master_replid2:0000000000000000000000000000000000000000 master_repl_offset:308 second_repl_offset:-1 repl_backlog_active:1 repl_backlog_size:1048576 repl_backlog_first_byte_offset:1 repl_backlog_histlen:308 127.0.0.1:17000>
這里5個slave均已成功連接到master上了。
四、啟動Sentinel（哨兵）進程
哨兵進程不一定與redis數量一致，也不一定要放在redis服務器上，sentinel的作用是監控所有服務及與其它哨兵通信，若sentinel單獨放其它服務器上，則也需要安裝redis，sentinel只是redis軟件包中的一個服務
每台服務器上都放了一個sentinel進程
啟動命令：src/redis-sentinel sentinel.conf
啟動后查看日志：
[root@redis redis]# src/redis-sentinel sentinel.conf [root@redis redis]# cat sentinel/sentinel.log 2747:X 15 Mar 2019 16:18:52.042 # oO0OoO0OoO0Oo Redis is starting oO0OoO0OoO0Oo 2747:X 15 Mar 2019 16:18:52.042 # Redis version=5.0.3, bits=64, commit=00000000, modified=0, pid=2747, just started 2747:X 15 Mar 2019 16:18:52.042 # Configuration loaded 2748:X 15 Mar 2019 16:18:52.059 * Increased maximum number of open files to 10032 (it was originally set to 1024). 2748:X 15 Mar 2019 16:18:52.061 * Running mode=sentinel, port=16000. 2748:X 15 Mar 2019 16:18:52.061 # WARNING: The TCP backlog setting of 511 cannot be enforced because /proc/sys/net/core/somaxconn is set to the lower value of 128. 2748:X 15 Mar 2019 16:18:52.071 # Sentinel ID is 53e37f3442336b7b229fa42c655bdca9eaa6578c 2748:X 15 Mar 2019 16:18:52.072 # +monitor master mymaster 192.168.1.161 17000 quorum 1 2748:X 15 Mar 2019 16:18:52.073 * +slave slave 192.168.1.162:17000 192.168.1.162 17000 @ mymaster 192.168.1.161 17000 2748:X 15 Mar 2019 16:18:52.075 * +slave slave 192.168.1.163:17000 192.168.1.163 17000 @ mymaster 192.168.1.161 17000 2748:X 15 Mar 2019 16:18:52.076 * +slave slave 192.168.1.141:17000 192.168.1.141 17000 @ mymaster 192.168.1.161 17000 2748:X 15 Mar 2019 16:18:52.078 * +slave slave 192.168.1.142:17000 192.168.1.142 17000 @ mymaster 192.168.1.161 17000 2748:X 15 Mar 2019 16:18:52.079 * +slave slave 192.168.1.143:17000 192.168.1.143 17000 @ mymaster 192.168.1.161 17000 2748:X 15 Mar 2019 16:18:52.515 * +sentinel sentinel 4a32c88aa8bff64295e45f8bb9c25154f4533d2a 192.168.1.163 16000 @ mymaster 192.168.1.161 17000 2748:X 15 Mar 2019 16:18:54.057 * +sentinel sentinel 93da0f13c96f22e0263e4288c68390cce5e11cea 192.168.1.161 16000 @ mymaster 192.168.1.161 17000 2748:X 15 Mar 2019 16:18:54.078 * +sentinel sentinel 605bf637a218de833308aa590cb3346de0eaacaf 192.168.1.162 16000 @ mymaster 192.168.1.161 17000 2748:X 15 Mar 2019 16:18:55.520 * +sentinel sentinel 3607a3f65b224ee03e49a825604b40d1cc1fbeaf 192.168.1.142 16000 @ mymaster 192.168.1.161 17000 2748:X 15 Mar 2019 16:18:58.040 * +sentinel sentinel d8e333f5daec7809605b8f276b1f2b44c1711459 192.168.1.143 16000 @ mymaster 192.168.1.161 17000
這里有個警告是由於TCP backlog 設置了511，而系統目前僅支持128，可以先忽略。
目前所有master-slave服務均正常。
五、故障轉移
模擬master宕機，這里直接把master進程殺掉
查看redis進程：
[root@mongodb-161 redis]# ps -e | grep redis 2730 ? 00:00:01 redis-server 2748 ? 00:00:02 redis-sentinel
redis-server進程號為2730
按進程號殺掉：kill -9 2730
[root@mongodb-161 redis]# kill -9 2730 [root@mongodb-161 redis]# ps -e | grep redis 2748 ? 00:00:02 redis-sentinel [root@mongodb-161 redis]#
殺掉后再查看進程，已經沒有了redis-server進程了。
然后查看192.168.1.161上的sentinel日志
[root@mongodb-161 redis]# cat sentinel/sentinel.log 3129:X 15 Mar 2019 16:18:42.113 # oO0OoO0OoO0Oo Redis is starting oO0OoO0OoO0Oo 3129:X 15 Mar 2019 16:18:42.113 # Redis version=5.0.3, bits=64, commit=00000000, modified=0, pid=3129, just started 3129:X 15 Mar 2019 16:18:42.113 # Configuration loaded 3130:X 15 Mar 2019 16:18:42.127 * Increased maximum number of open files to 10032 (it was originally set to 1024). 3130:X 15 Mar 2019 16:18:42.129 * Running mode=sentinel, port=16000. 3130:X 15 Mar 2019 16:18:42.129 # WARNING: The TCP backlog setting of 511 cannot be enforced because /proc/sys/net/core/somaxconn is set to the lower value of 128. 3130:X 15 Mar 2019 16:18:42.139 # Sentinel ID is 93da0f13c96f22e0263e4288c68390cce5e11cea 3130:X 15 Mar 2019 16:18:42.139 # +monitor master mymaster 192.168.1.161 17000 quorum 1 3130:X 15 Mar 2019 16:18:42.141 * +slave slave 192.168.1.162:17000 192.168.1.162 17000 @ mymaster 192.168.1.161 17000 3130:X 15 Mar 2019 16:18:42.143 * +slave slave 192.168.1.163:17000 192.168.1.163 17000 @ mymaster 192.168.1.161 17000 3130:X 15 Mar 2019 16:18:42.144 * +slave slave 192.168.1.141:17000 192.168.1.141 17000 @ mymaster 192.168.1.161 17000 3130:X 15 Mar 2019 16:18:42.146 * +slave slave 192.168.1.142:17000 192.168.1.142 17000 @ mymaster 192.168.1.161 17000 3130:X 15 Mar 2019 16:18:42.147 * +slave slave 192.168.1.143:17000 192.168.1.143 17000 @ mymaster 192.168.1.161 17000 3130:X 15 Mar 2019 16:18:50.335 * +sentinel sentinel 605bf637a218de833308aa590cb3346de0eaacaf 192.168.1.162 16000 @ mymaster 192.168.1.161 17000 3130:X 15 Mar 2019 16:18:52.836 * +sentinel sentinel 4a32c88aa8bff64295e45f8bb9c25154f4533d2a 192.168.1.163 16000 @ mymaster 192.168.1.161 17000 3130:X 15 Mar 2019 16:18:54.374 * +sentinel sentinel 53e37f3442336b7b229fa42c655bdca9eaa6578c 192.168.1.141 16000 @ mymaster 192.168.1.161 17000 3130:X 15 Mar 2019 16:18:55.841 * +sentinel sentinel 3607a3f65b224ee03e49a825604b40d1cc1fbeaf 192.168.1.142 16000 @ mymaster 192.168.1.161 17000 3130:X 15 Mar 2019 16:18:58.361 * +sentinel sentinel d8e333f5daec7809605b8f276b1f2b44c1711459 192.168.1.143 16000 @ mymaster 192.168.1.161 17000 3130:X 15 Mar 2019 16:25:46.300 # +new-epoch 1 3130:X 15 Mar 2019 16:25:46.301 # +vote-for-leader 4a32c88aa8bff64295e45f8bb9c25154f4533d2a 1 3130:X 15 Mar 2019 16:25:46.336 # +sdown master mymaster 192.168.1.161 17000 3130:X 15 Mar 2019 16:25:46.336 # +odown master mymaster 192.168.1.161 17000 #quorum 1/1 3130:X 15 Mar 2019 16:25:46.336 # Next failover delay: I will not start a failover before Fri Mar 15 16:35:47 2019 3130:X 15 Mar 2019 16:25:46.635 # +config-update-from sentinel 4a32c88aa8bff64295e45f8bb9c25154f4533d2a 192.168.1.163 16000 @ mymaster 192.168.1.161 17000 3130:X 15 Mar 2019 16:25:46.635 # +switch-master mymaster 192.168.1.161 17000 192.168.1.162 17000 3130:X 15 Mar 2019 16:25:46.636 * +slave slave 192.168.1.163:17000 192.168.1.163 17000 @ mymaster 192.168.1.162 17000 3130:X 15 Mar 2019 16:25:46.636 * +slave slave 192.168.1.142:17000 192.168.1.142 17000 @ mymaster 192.168.1.162 17000 3130:X 15 Mar 2019 16:25:46.636 * +slave slave 192.168.1.143:17000 192.168.1.143 17000 @ mymaster 192.168.1.162 17000 3130:X 15 Mar 2019 16:25:46.636 * +slave slave 192.168.1.141:17000 192.168.1.141 17000 @ mymaster 192.168.1.162 17000 3130:X 15 Mar 2019 16:25:46.636 * +slave slave 192.168.1.161:17000 192.168.1.161 17000 @ mymaster 192.168.1.162 17000 3130:X 15 Mar 2019 16:25:51.638 # +sdown slave 192.168.1.161:17000 192.168.1.161 17000 @ mymaster 192.168.1.162 17000
可以看到時間在16:25:46時master宕機后，已經進行了故障轉移，新的master是192.168.1.162，查看192.168.1.162的sentinel日志：
[root@mongodb-162 redis]# cat sentinel/sentinel.log 3103:X 15 Mar 2019 16:18:48.245 # oO0OoO0OoO0Oo Redis is starting oO0OoO0OoO0Oo 3103:X 15 Mar 2019 16:18:48.245 # Redis version=5.0.3, bits=64, commit=00000000, modified=0, pid=3103, just started 3103:X 15 Mar 2019 16:18:48.245 # Configuration loaded 3104:X 15 Mar 2019 16:18:48.250 * Increased maximum number of open files to 10032 (it was originally set to 1024). 3104:X 15 Mar 2019 16:18:48.251 * Running mode=sentinel, port=16000. 3104:X 15 Mar 2019 16:18:48.252 # WARNING: The TCP backlog setting of 511 cannot be enforced because /proc/sys/net/core/somaxconn is set to the lower value of 128. 3104:X 15 Mar 2019 16:18:48.254 # Sentinel ID is 605bf637a218de833308aa590cb3346de0eaacaf 3104:X 15 Mar 2019 16:18:48.254 # +monitor master mymaster 192.168.1.161 17000 quorum 1 3104:X 15 Mar 2019 16:18:48.256 * +slave slave 192.168.1.162:17000 192.168.1.162 17000 @ mymaster 192.168.1.161 17000 3104:X 15 Mar 2019 16:18:48.257 * +slave slave 192.168.1.163:17000 192.168.1.163 17000 @ mymaster 192.168.1.161 17000 3104:X 15 Mar 2019 16:18:48.259 * +slave slave 192.168.1.141:17000 192.168.1.141 17000 @ mymaster 192.168.1.161 17000 3104:X 15 Mar 2019 16:18:48.261 * +slave slave 192.168.1.142:17000 192.168.1.142 17000 @ mymaster 192.168.1.161 17000 3104:X 15 Mar 2019 16:18:48.262 * +slave slave 192.168.1.143:17000 192.168.1.143 17000 @ mymaster 192.168.1.161 17000 3104:X 15 Mar 2019 16:18:48.270 * +sentinel sentinel 93da0f13c96f22e0263e4288c68390cce5e11cea 192.168.1.161 16000 @ mymaster 192.168.1.161 17000 3104:X 15 Mar 2019 16:18:52.815 * +sentinel sentinel 4a32c88aa8bff64295e45f8bb9c25154f4533d2a 192.168.1.163 16000 @ mymaster 192.168.1.161 17000 3104:X 15 Mar 2019 16:18:54.355 * +sentinel sentinel 53e37f3442336b7b229fa42c655bdca9eaa6578c 192.168.1.141 16000 @ mymaster 192.168.1.161 17000 3104:X 15 Mar 2019 16:18:55.820 * +sentinel sentinel 3607a3f65b224ee03e49a825604b40d1cc1fbeaf 192.168.1.142 16000 @ mymaster 192.168.1.161 17000 3104:X 15 Mar 2019 16:18:58.341 * +sentinel sentinel d8e333f5daec7809605b8f276b1f2b44c1711459 192.168.1.143 16000 @ mymaster 192.168.1.161 17000 3104:X 15 Mar 2019 16:25:46.277 # +new-epoch 1 3104:X 15 Mar 2019 16:25:46.279 # +vote-for-leader 4a32c88aa8bff64295e45f8bb9c25154f4533d2a 1 3104:X 15 Mar 2019 16:25:46.309 # +sdown master mymaster 192.168.1.161 17000 3104:X 15 Mar 2019 16:25:46.309 # +odown master mymaster 192.168.1.161 17000 #quorum 1/1 3104:X 15 Mar 2019 16:25:46.309 # Next failover delay: I will not start a failover before Fri Mar 15 16:35:46 2019 3104:X 15 Mar 2019 16:25:46.612 # +config-update-from sentinel 4a32c88aa8bff64295e45f8bb9c25154f4533d2a 192.168.1.163 16000 @ mymaster 192.168.1.161 17000 3104:X 15 Mar 2019 16:25:46.612 # +switch-master mymaster 192.168.1.161 17000 192.168.1.162 17000 3104:X 15 Mar 2019 16:25:46.614 * +slave slave 192.168.1.142:17000 192.168.1.142 17000 @ mymaster 192.168.1.162 17000 3104:X 15 Mar 2019 16:25:46.614 * +slave slave 192.168.1.143:17000 192.168.1.143 17000 @ mymaster 192.168.1.162 17000 3104:X 15 Mar 2019 16:25:46.614 * +slave slave 192.168.1.141:17000 192.168.1.141 17000 @ mymaster 192.168.1.162 17000 3104:X 15 Mar 2019 16:25:46.614 * +slave slave 192.168.1.163:17000 192.168.1.163 17000 @ mymaster 192.168.1.162 17000 3104:X 15 Mar 2019 16:25:46.614 * +slave slave 192.168.1.161:17000 192.168.1.161 17000 @ mymaster 192.168.1.162 17000 3104:X 15 Mar 2019 16:25:51.622 # +sdown slave 192.168.1.161:17000 192.168.1.161 17000 @ mymaster 192.168.1.162 17000
跟161上日志幾乎是一樣的，那么162是怎么被選舉出作為master的呢，繼續查看163日志：
[root@mongodb-163 redis]# cat sentinel/sentinel.log 3067:X 15 Mar 2019 16:18:50.731 # oO0OoO0OoO0Oo Redis is starting oO0OoO0OoO0Oo 3067:X 15 Mar 2019 16:18:50.731 # Redis version=5.0.3, bits=64, commit=00000000, modified=0, pid=3067, just started 3067:X 15 Mar 2019 16:18:50.731 # Configuration loaded 3068:X 15 Mar 2019 16:18:50.747 * Increased maximum number of open files to 10032 (it was originally set to 1024). 3068:X 15 Mar 2019 16:18:50.749 * Running mode=sentinel, port=16000. 3068:X 15 Mar 2019 16:18:50.749 # WARNING: The TCP backlog setting of 511 cannot be enforced because /proc/sys/net/core/somaxconn is set to the lower value of 128. 3068:X 15 Mar 2019 16:18:50.751 # Sentinel ID is 4a32c88aa8bff64295e45f8bb9c25154f4533d2a 3068:X 15 Mar 2019 16:18:50.751 # +monitor master mymaster 192.168.1.161 17000 quorum 1 3068:X 15 Mar 2019 16:18:50.754 * +slave slave 192.168.1.162:17000 192.168.1.162 17000 @ mymaster 192.168.1.161 17000 3068:X 15 Mar 2019 16:18:50.756 * +slave slave 192.168.1.163:17000 192.168.1.163 17000 @ mymaster 192.168.1.161 17000 3068:X 15 Mar 2019 16:18:50.757 * +slave slave 192.168.1.141:17000 192.168.1.141 17000 @ mymaster 192.168.1.161 17000 3068:X 15 Mar 2019 16:18:50.759 * +slave slave 192.168.1.142:17000 192.168.1.142 17000 @ mymaster 192.168.1.161 17000 3068:X 15 Mar 2019 16:18:50.760 * +slave slave 192.168.1.143:17000 192.168.1.143 17000 @ mymaster 192.168.1.161 17000 3068:X 15 Mar 2019 16:18:52.329 * +sentinel sentinel 93da0f13c96f22e0263e4288c68390cce5e11cea 192.168.1.161 16000 @ mymaster 192.168.1.161 17000 3068:X 15 Mar 2019 16:18:52.331 * +sentinel sentinel 605bf637a218de833308aa590cb3346de0eaacaf 192.168.1.162 16000 @ mymaster 192.168.1.161 17000 3068:X 15 Mar 2019 16:18:54.330 * +sentinel sentinel 53e37f3442336b7b229fa42c655bdca9eaa6578c 192.168.1.141 16000 @ mymaster 192.168.1.161 17000 3068:X 15 Mar 2019 16:18:55.796 * +sentinel sentinel 3607a3f65b224ee03e49a825604b40d1cc1fbeaf 192.168.1.142 16000 @ mymaster 192.168.1.161 17000 3068:X 15 Mar 2019 16:18:58.316 * +sentinel sentinel d8e333f5daec7809605b8f276b1f2b44c1711459 192.168.1.143 16000 @ mymaster 192.168.1.161 17000 3068:X 15 Mar 2019 16:25:46.247 # +sdown master mymaster 192.168.1.161 17000 3068:X 15 Mar 2019 16:25:46.248 # +odown master mymaster 192.168.1.161 17000 #quorum 1/1 3068:X 15 Mar 2019 16:25:46.248 # +new-epoch 1 3068:X 15 Mar 2019 16:25:46.248 # +try-failover master mymaster 192.168.1.161 17000 3068:X 15 Mar 2019 16:25:46.250 # +vote-for-leader 4a32c88aa8bff64295e45f8bb9c25154f4533d2a 1 3068:X 15 Mar 2019 16:25:46.253 # 53e37f3442336b7b229fa42c655bdca9eaa6578c voted for 53e37f3442336b7b229fa42c655bdca9eaa6578c 1 3068:X 15 Mar 2019 16:25:46.254 # 93da0f13c96f22e0263e4288c68390cce5e11cea voted for 4a32c88aa8bff64295e45f8bb9c25154f4533d2a 1 3068:X 15 Mar 2019 16:25:46.255 # 605bf637a218de833308aa590cb3346de0eaacaf voted for 4a32c88aa8bff64295e45f8bb9c25154f4533d2a 1 3068:X 15 Mar 2019 16:25:46.255 # 3607a3f65b224ee03e49a825604b40d1cc1fbeaf voted for 4a32c88aa8bff64295e45f8bb9c25154f4533d2a 1 3068:X 15 Mar 2019 16:25:46.255 # d8e333f5daec7809605b8f276b1f2b44c1711459 voted for 4a32c88aa8bff64295e45f8bb9c25154f4533d2a 1 3068:X 15 Mar 2019 16:25:46.316 # +elected-leader master mymaster 192.168.1.161 17000 3068:X 15 Mar 2019 16:25:46.316 # +failover-state-select-slave master mymaster 192.168.1.161 17000 3068:X 15 Mar 2019 16:25:46.372 # +selected-slave slave 192.168.1.162:17000 192.168.1.162 17000 @ mymaster 192.168.1.161 17000 3068:X 15 Mar 2019 16:25:46.372 * +failover-state-send-slaveof-noone slave 192.168.1.162:17000 192.168.1.162 17000 @ mymaster 192.168.1.161 17000 3068:X 15 Mar 2019 16:25:46.424 * +failover-state-wait-promotion slave 192.168.1.162:17000 192.168.1.162 17000 @ mymaster 192.168.1.161 17000 3068:X 15 Mar 2019 16:25:46.528 # +promoted-slave slave 192.168.1.162:17000 192.168.1.162 17000 @ mymaster 192.168.1.161 17000 3068:X 15 Mar 2019 16:25:46.528 # +failover-state-reconf-slaves master mymaster 192.168.1.161 17000 3068:X 15 Mar 2019 16:25:46.584 * +slave-reconf-sent slave 192.168.1.163:17000 192.168.1.163 17000 @ mymaster 192.168.1.161 17000 3068:X 15 Mar 2019 16:25:46.584 * +slave-reconf-sent slave 192.168.1.142:17000 192.168.1.142 17000 @ mymaster 192.168.1.161 17000 3068:X 15 Mar 2019 16:25:47.546 * +slave-reconf-inprog slave 192.168.1.163:17000 192.168.1.163 17000 @ mymaster 192.168.1.161 17000 3068:X 15 Mar 2019 16:25:47.546 * +slave-reconf-done slave 192.168.1.163:17000 192.168.1.163 17000 @ mymaster 192.168.1.161 17000 3068:X 15 Mar 2019 16:25:47.546 * +slave-reconf-inprog slave 192.168.1.142:17000 192.168.1.142 17000 @ mymaster 192.168.1.161 17000 3068:X 15 Mar 2019 16:25:47.546 * +slave-reconf-done slave 192.168.1.142:17000 192.168.1.142 17000 @ mymaster 192.168.1.161 17000 3068:X 15 Mar 2019 16:25:47.602 * +slave-reconf-sent slave 192.168.1.143:17000 192.168.1.143 17000 @ mymaster 192.168.1.161 17000 3068:X 15 Mar 2019 16:25:47.602 * +slave-reconf-sent slave 192.168.1.141:17000 192.168.1.141 17000 @ mymaster 192.168.1.161 17000 3068:X 15 Mar 2019 16:25:48.588 * +slave-reconf-inprog slave 192.168.1.141:17000 192.168.1.141 17000 @ mymaster 192.168.1.161 17000 3068:X 15 Mar 2019 16:25:48.588 * +slave-reconf-done slave 192.168.1.141:17000 192.168.1.141 17000 @ mymaster 192.168.1.161 17000 3068:X 15 Mar 2019 16:25:48.588 * +slave-reconf-inprog slave 192.168.1.143:17000 192.168.1.143 17000 @ mymaster 192.168.1.161 17000 3068:X 15 Mar 2019 16:25:49.661 * +slave-reconf-done slave 192.168.1.143:17000 192.168.1.143 17000 @ mymaster 192.168.1.161 17000 3068:X 15 Mar 2019 16:25:49.732 # +failover-end master mymaster 192.168.1.161 17000 3068:X 15 Mar 2019 16:25:49.732 # +switch-master mymaster 192.168.1.161 17000 192.168.1.162 17000 3068:X 15 Mar 2019 16:25:49.733 * +slave slave 192.168.1.163:17000 192.168.1.163 17000 @ mymaster 192.168.1.162 17000 3068:X 15 Mar 2019 16:25:49.733 * +slave slave 192.168.1.142:17000 192.168.1.142 17000 @ mymaster 192.168.1.162 17000 3068:X 15 Mar 2019 16:25:49.733 * +slave slave 192.168.1.143:17000 192.168.1.143 17000 @ mymaster 192.168.1.162 17000 3068:X 15 Mar 2019 16:25:49.733 * +slave slave 192.168.1.141:17000 192.168.1.141 17000 @ mymaster 192.168.1.162 17000 3068:X 15 Mar 2019 16:25:49.733 * +slave slave 192.168.1.161:17000 192.168.1.161 17000 @ mymaster 192.168.1.162 17000 3068:X 15 Mar 2019 16:25:54.743 # +sdown slave 192.168.1.161:17000 192.168.1.161 17000 @ mymaster 192.168.1.162 17000
在16:25:46.247時sentinel已將161進行了主觀下線（sdown，標記為不可用），然后緊接着進行了客觀下線（odown，sentinel認為確實不可用），這里說明一下，為什么主觀下線后緊跟着客觀下線，原因是配置了：
sentinel monitor mymaster 192.168.1.161 17000 1
最后這個1的意思是只要大於等於1個sentinel認為sdown，則認為可以進行odown了，所以日志中有一句：odown master mymaster 192.168.1.161 17000 #quorum 1/1，所以odown緊跟在sdown后面。
在163的sentinel的日志中：Sentinel ID is 4a32c88aa8bff64295e45f8bb9c25154f4533d2a 表示163上的sentinel的ID，后面會用到這個。
當sentinel設置161為odown（客觀下線）后，這些sentinel開始進行投票選舉，這里特別注意，選舉的不是redis_server服務，而是sentinel服務，意思就是先選舉一個sentinel，然后由它決定將哪個slave提升為master。
查看141服務器上的sentinel日志：
[root@redis redis]# cat sentinel/sentinel.log 2747:X 15 Mar 2019 16:18:52.042 # oO0OoO0OoO0Oo Redis is starting oO0OoO0OoO0Oo 2747:X 15 Mar 2019 16:18:52.042 # Redis version=5.0.3, bits=64, commit=00000000, modified=0, pid=2747, just started 2747:X 15 Mar 2019 16:18:52.042 # Configuration loaded 2748:X 15 Mar 2019 16:18:52.059 * Increased maximum number of open files to 10032 (it was originally set to 1024). 2748:X 15 Mar 2019 16:18:52.061 * Running mode=sentinel, port=16000. 2748:X 15 Mar 2019 16:18:52.061 # WARNING: The TCP backlog setting of 511 cannot be enforced because /proc/sys/net/core/somaxconn is set to the lower value of 128. 2748:X 15 Mar 2019 16:18:52.071 # Sentinel ID is 53e37f3442336b7b229fa42c655bdca9eaa6578c 2748:X 15 Mar 2019 16:18:52.072 # +monitor master mymaster 192.168.1.161 17000 quorum 1 2748:X 15 Mar 2019 16:18:52.073 * +slave slave 192.168.1.162:17000 192.168.1.162 17000 @ mymaster 192.168.1.161 17000 2748:X 15 Mar 2019 16:18:52.075 * +slave slave 192.168.1.163:17000 192.168.1.163 17000 @ mymaster 192.168.1.161 17000 2748:X 15 Mar 2019 16:18:52.076 * +slave slave 192.168.1.141:17000 192.168.1.141 17000 @ mymaster 192.168.1.161 17000 2748:X 15 Mar 2019 16:18:52.078 * +slave slave 192.168.1.142:17000 192.168.1.142 17000 @ mymaster 192.168.1.161 17000 2748:X 15 Mar 2019 16:18:52.079 * +slave slave 192.168.1.143:17000 192.168.1.143 17000 @ mymaster 192.168.1.161 17000 2748:X 15 Mar 2019 16:18:52.515 * +sentinel sentinel 4a32c88aa8bff64295e45f8bb9c25154f4533d2a 192.168.1.163 16000 @ mymaster 192.168.1.161 17000 2748:X 15 Mar 2019 16:18:54.057 * +sentinel sentinel 93da0f13c96f22e0263e4288c68390cce5e11cea 192.168.1.161 16000 @ mymaster 192.168.1.161 17000 2748:X 15 Mar 2019 16:18:54.078 * +sentinel sentinel 605bf637a218de833308aa590cb3346de0eaacaf 192.168.1.162 16000 @ mymaster 192.168.1.161 17000 2748:X 15 Mar 2019 16:18:55.520 * +sentinel sentinel 3607a3f65b224ee03e49a825604b40d1cc1fbeaf 192.168.1.142 16000 @ mymaster 192.168.1.161 17000 2748:X 15 Mar 2019 16:18:58.040 * +sentinel sentinel d8e333f5daec7809605b8f276b1f2b44c1711459 192.168.1.143 16000 @ mymaster 192.168.1.161 17000 2748:X 15 Mar 2019 16:25:45.962 # +sdown master mymaster 192.168.1.161 17000 2748:X 15 Mar 2019 16:25:45.963 # +odown master mymaster 192.168.1.161 17000 #quorum 1/1 2748:X 15 Mar 2019 16:25:45.963 # +new-epoch 1 2748:X 15 Mar 2019 16:25:45.963 # +try-failover master mymaster 192.168.1.161 17000 2748:X 15 Mar 2019 16:25:45.966 # +vote-for-leader 53e37f3442336b7b229fa42c655bdca9eaa6578c 1 2748:X 15 Mar 2019 16:25:45.966 # 4a32c88aa8bff64295e45f8bb9c25154f4533d2a voted for 4a32c88aa8bff64295e45f8bb9c25154f4533d2a 1 2748:X 15 Mar 2019 16:25:45.967 # 93da0f13c96f22e0263e4288c68390cce5e11cea voted for 4a32c88aa8bff64295e45f8bb9c25154f4533d2a 1 2748:X 15 Mar 2019 16:25:45.968 # 605bf637a218de833308aa590cb3346de0eaacaf voted for 4a32c88aa8bff64295e45f8bb9c25154f4533d2a 1 2748:X 15 Mar 2019 16:25:45.968 # 3607a3f65b224ee03e49a825604b40d1cc1fbeaf voted for 4a32c88aa8bff64295e45f8bb9c25154f4533d2a 1 2748:X 15 Mar 2019 16:25:45.968 # d8e333f5daec7809605b8f276b1f2b44c1711459 voted for 4a32c88aa8bff64295e45f8bb9c25154f4533d2a 1 2748:X 15 Mar 2019 16:25:46.301 # +config-update-from sentinel 4a32c88aa8bff64295e45f8bb9c25154f4533d2a 192.168.1.163 16000 @ mymaster 192.168.1.161 17000 2748:X 15 Mar 2019 16:25:46.301 # +switch-master mymaster 192.168.1.161 17000 192.168.1.162 17000 2748:X 15 Mar 2019 16:25:46.302 * +slave slave 192.168.1.142:17000 192.168.1.142 17000 @ mymaster 192.168.1.162 17000 2748:X 15 Mar 2019 16:25:46.302 * +slave slave 192.168.1.141:17000 192.168.1.141 17000 @ mymaster 192.168.1.162 17000 2748:X 15 Mar 2019 16:25:46.303 * +slave slave 192.168.1.163:17000 192.168.1.163 17000 @ mymaster 192.168.1.162 17000 2748:X 15 Mar 2019 16:25:46.303 * +slave slave 192.168.1.143:17000 192.168.1.143 17000 @ mymaster 192.168.1.162 17000 2748:X 15 Mar 2019 16:25:46.303 * +slave slave 192.168.1.161:17000 192.168.1.161 17000 @ mymaster 192.168.1.162 17000 2748:X 15 Mar 2019 16:25:51.320 # +sdown slave 192.168.1.161:17000 192.168.1.161 17000 @ mymaster 192.168.1.162 17000
由141上日志看出：選舉后，141上的sentinel並沒有進行操作failover，所以可以證實選舉是sentinel。
日志說明：
這段日志就是在進行選舉：
3068:X 15 Mar 2019 16:25:46.248 # +new-epoch 1 3068:X 15 Mar 2019 16:25:46.248 # +try-failover master mymaster 192.168.1.161 17000 3068:X 15 Mar 2019 16:25:46.250 # +vote-for-leader 4a32c88aa8bff64295e45f8bb9c25154f4533d2a 1 3068:X 15 Mar 2019 16:25:46.253 # 53e37f3442336b7b229fa42c655bdca9eaa6578c voted for 53e37f3442336b7b229fa42c655bdca9eaa6578c 1 3068:X 15 Mar 2019 16:25:46.254 # 93da0f13c96f22e0263e4288c68390cce5e11cea voted for 4a32c88aa8bff64295e45f8bb9c25154f4533d2a 1 3068:X 15 Mar 2019 16:25:46.255 # 605bf637a218de833308aa590cb3346de0eaacaf voted for 4a32c88aa8bff64295e45f8bb9c25154f4533d2a 1 3068:X 15 Mar 2019 16:25:46.255 # 3607a3f65b224ee03e49a825604b40d1cc1fbeaf voted for 4a32c88aa8bff64295e45f8bb9c25154f4533d2a 1 3068:X 15 Mar 2019 16:25:46.255 # d8e333f5daec7809605b8f276b1f2b44c1711459 voted for 4a32c88aa8bff64295e45f8bb9c25154f4533d2a 1
可以看出id為4a32c88aa8bff64295e45f8bb9c25154f4533d2a的sentinel有4票（共6個sentinel），它決定哪個slave升為master：
3068:X 15 Mar 2019 16:25:46.316 # +elected-leader master mymaster 192.168.1.161 17000 3068:X 15 Mar 2019 16:25:46.316 # +failover-state-select-slave master mymaster 192.168.1.161 17000 3068:X 15 Mar 2019 16:25:46.372 # +selected-slave slave 192.168.1.162:17000 192.168.1.162 17000 @ mymaster 192.168.1.161 17000 3068:X 15 Mar 2019 16:25:46.372 * +failover-state-send-slaveof-noone slave 192.168.1.162:17000 192.168.1.162 17000 @ mymaster 192.168.1.161 17000 3068:X 15 Mar 2019 16:25:46.424 * +failover-state-wait-promotion slave 192.168.1.162:17000 192.168.1.162 17000 @ mymaster 192.168.1.161 17000 3068:X 15 Mar 2019 16:25:46.528 # +promoted-slave slave 192.168.1.162:17000 192.168.1.162 17000 @ mymaster 192.168.1.161 17000 3068:X 15 Mar 2019 16:25:46.528 # +failover-state-reconf-slaves master mymaster 192.168.1.161 17000
最終163上的sentinel選擇了162為master，並進行了failover
其它各服務器則進行更新配置：
3068:X 15 Mar 2019 16:25:46.584 * +slave-reconf-sent slave 192.168.1.163:17000 192.168.1.163 17000 @ mymaster 192.168.1.161 17000 3068:X 15 Mar 2019 16:25:46.584 * +slave-reconf-sent slave 192.168.1.142:17000 192.168.1.142 17000 @ mymaster 192.168.1.161 17000 3068:X 15 Mar 2019 16:25:47.546 * +slave-reconf-inprog slave 192.168.1.163:17000 192.168.1.163 17000 @ mymaster 192.168.1.161 17000 3068:X 15 Mar 2019 16:25:47.546 * +slave-reconf-done slave 192.168.1.163:17000 192.168.1.163 17000 @ mymaster 192.168.1.161 17000 3068:X 15 Mar 2019 16:25:47.546 * +slave-reconf-inprog slave 192.168.1.142:17000 192.168.1.142 17000 @ mymaster 192.168.1.161 17000 3068:X 15 Mar 2019 16:25:47.546 * +slave-reconf-done slave 192.168.1.142:17000 192.168.1.142 17000 @ mymaster 192.168.1.161 17000 3068:X 15 Mar 2019 16:25:47.602 * +slave-reconf-sent slave 192.168.1.143:17000 192.168.1.143 17000 @ mymaster 192.168.1.161 17000 3068:X 15 Mar 2019 16:25:47.602 * +slave-reconf-sent slave 192.168.1.141:17000 192.168.1.141 17000 @ mymaster 192.168.1.161 17000 3068:X 15 Mar 2019 16:25:48.588 * +slave-reconf-inprog slave 192.168.1.141:17000 192.168.1.141 17000 @ mymaster 192.168.1.161 17000 3068:X 15 Mar 2019 16:25:48.588 * +slave-reconf-done slave 192.168.1.141:17000 192.168.1.141 17000 @ mymaster 192.168.1.161 17000 3068:X 15 Mar 2019 16:25:48.588 * +slave-reconf-inprog slave 192.168.1.143:17000 192.168.1.143 17000 @ mymaster 192.168.1.161 17000 3068:X 15 Mar 2019 16:25:49.661 * +slave-reconf-done slave 192.168.1.143:17000 192.168.1.143 17000 @ mymaster 192.168.1.161 17000
最后，形成新的master-slave集群：
3068:X 15 Mar 2019 16:25:49.733 * +slave slave 192.168.1.163:17000 192.168.1.163 17000 @ mymaster 192.168.1.162 17000 3068:X 15 Mar 2019 16:25:49.733 * +slave slave 192.168.1.142:17000 192.168.1.142 17000 @ mymaster 192.168.1.162 17000 3068:X 15 Mar 2019 16:25:49.733 * +slave slave 192.168.1.143:17000 192.168.1.143 17000 @ mymaster 192.168.1.162 17000 3068:X 15 Mar 2019 16:25:49.733 * +slave slave 192.168.1.141:17000 192.168.1.141 17000 @ mymaster 192.168.1.162 17000 3068:X 15 Mar 2019 16:25:49.733 * +slave slave 192.168.1.161:17000 192.168.1.161 17000 @ mymaster 192.168.1.162 17000 3068:X 15 Mar 2019 16:25:54.743 # +sdown slave 192.168.1.161:17000 192.168.1.161 17000 @ mymaster 192.168.1.162 17000
查看162上的replication信息：
[root@mongodb-162 redis]# src/redis-cli -p 17000 127.0.0.1:17000> AUTH 123456 OK 127.0.0.1:17000> info replication # Replication role:master connected_slaves:4 slave0:ip=192.168.1.163,port=17000,state=online,offset=947470,lag=1 slave1:ip=192.168.1.142,port=17000,state=online,offset=947470,lag=1 slave2:ip=192.168.1.141,port=17000,state=online,offset=947754,lag=1 slave3:ip=192.168.1.143,port=17000,state=online,offset=947470,lag=1 master_replid:b6c12c4d48aa953cb473480198359da669f171cc master_replid2:b3f15ec093bb8ab2b1e908b41740f4bdabbfeba4 master_repl_offset:948180 second_repl_offset:172534 repl_backlog_active:1 repl_backlog_size:1048576 repl_backlog_first_byte_offset:1 repl_backlog_histlen:948180 127.0.0.1:17000>
再查看163上集群配置：
[root@mongodb-163 redis]# src/redis-cli -p 17000 127.0.0.1:17000> AUTH 123456 OK 127.0.0.1:17000> info replication # Replication role:slave master_host:192.168.1.162 master_port:17000 master_link_status:up master_last_io_seconds_ago:1 master_sync_in_progress:0 slave_repl_offset:955166 slave_priority:100 slave_read_only:1 connected_slaves:0 master_replid:b6c12c4d48aa953cb473480198359da669f171cc master_replid2:b3f15ec093bb8ab2b1e908b41740f4bdabbfeba4 master_repl_offset:955166 second_repl_offset:172534 repl_backlog_active:1 repl_backlog_size:1048576 repl_backlog_first_byte_offset:253 repl_backlog_histlen:954914 127.0.0.1:17000>
最后，master宕機后，master-slave自動切換成功。
六、遇到的坑
1.slave不能連接到master，原因是防火牆開着的，解決方案是關閉防火牆，或把相應端口添加到防火牆中
2.master宕機后，不能自動選舉新的master，查看日志報如下錯誤：
2230:X 15 Mar 2019 09:11:16.006 # -failover-abort-not-elected master mymaster 192.168.1.161 17000
2230:X 15 Mar 2019 09:11:16.007 # Next failover delay: I will not start a failover before Fri Mar 15 09:17:05 2019
原因是：原先開啟了保護模式：protected-mode yes，而bind參數沒有設置正確，實際上redis_master.conf、redis_slave.conf、sentinel.conf均需要設置正確，綁定上正確的網卡IP，我這里嫌麻煩改每台服務器配置（配置文件是做好后上傳到服務器的），就把保護模式關閉了：protected-mode no。
官方配置中是默認開啟的，一定要注意這點。不過文檔中也建議，如果設置了訪問密碼，也可以關閉。