Redis 哨兵搭建的實(shí)現(xiàn)示例
更新時間：2025年09月23日 11:28:38 作者：咚咚？
本文詳細(xì)介紹了Redis 7.2.5單節(jié)點(diǎn)和多節(jié)點(diǎn)哨兵的搭建流程,文中通過示例代碼介紹的非常詳細(xì),對大家的學(xué)習(xí)或者工作具有一定的參考學(xué)習(xí)價值,需要的朋友們下面隨著小編來一起學(xué)習(xí)學(xué)習(xí)吧
一、單節(jié)點(diǎn)哨兵
1. 環(huán)境介紹
| 操作系統(tǒng) | Centos 7 |
|---|---|
| 內(nèi)核版本 | Linux 3.10.0-957.el7.x86_64 |
| 主機(jī)名稱 | master |
| IP | 192.168.1.100 |
| 端口 | master:16370 slaver1:16371 slaver2: 16372 sentinel:16379 |
2. 環(huán)境前準(zhǔn)備工作
# 關(guān)閉防火墻 systemctl stop firewalld systemctl disable firewalld # 修改 hostname hostnamectl set-hostname xxxx # 修改后退出當(dāng)前終端重新連接即可 # 更新下軟件源 # 地址 https://developer.aliyun.com/mirror/ # 時區(qū)調(diào)整，時間校準(zhǔn) date -R timedatectl set-timezone Asia/Shanghai yum -y install ntp ntpdate ntp1.aliyun.com # 關(guān)閉 selinux： sed -i 's/enforcing/disabled/' /etc/selinux/config setenforce 0 # 修改內(nèi)存分配控制 vi /etc/sysctl.conf vm.overcommit_memory = 1 # 設(shè)置TPC積壓值 net.core.somaxconn= 1024 # 使其配置文件生效 sysctl -p /etc/sysctl.conf
3. 安裝 Redis 7.2.5
# 下載地址 http://download.redis.io/releases/ # 下載 wget http://download.redis.io/releases/redis-7.2.5.tar.gz # 解壓 tar -xf redis-7.2.5.tar.gz -C /opt/redis # 編譯 make MALLOC=libc # 安裝 注意如果安裝提示python3不存在 執(zhí)行 yum install -y python3 make install PREFIX=/usr/local/redis # 配置環(huán)境變量 vi /etc/profile.d/redis.sh #!/bin/bash export REDIS_HOME=/usr/local/redis export PATH=$PATH:$REDIS_HOME/bin # 使其配置生效 source /etc/profile.d/redis.sh
4. redis 配置修改并且啟動
4.1 修改配置文件
# 創(chuàng)建redis數(shù)據(jù)存儲目錄
mkdir -p /usr/local/redis/data/1637{0..2}
# 創(chuàng)建redis配置文件存儲目錄
mkdir -p /usr/local/redis/conf/1637{0..2}
# 創(chuàng)建redis日志存儲目錄
mkdir -p /usr/local/redis/log/1637{0..2}
# 創(chuàng)建redis pid存儲目錄
mkdir /usr/local/redis/run
# redis配置文件 16370
# 開啟保護(hù)模式
protected-mode yes
# 添加本機(jī)的ip
bind 192.168.1.100
# 端口
port 16370
# pid存儲目錄
pidfile /usr/local/redis/run/redis_16370.pid
# 日志存儲目錄
logfile /usr/local/redis/log/16370/redis_16370.log
# 數(shù)據(jù)存儲目錄，目錄要提前創(chuàng)建好
dir /usr/local/redis/data/16370
# 設(shè)置實(shí)例的驗(yàn)證口令
requirepass dongdong
# 以認(rèn)證的方式連接到master。 如果master中使用了“密碼保護(hù)”，slave必須交付正確的授權(quán)密碼，才能連接成功。
# 此配置項中值需要和master機(jī)器的“requirepass”保持一致
masterauth dongdong
# 配置RDB持久化策略
save 900 1
save 300 10
save 60 10000
# 配置AOF持久化
appendonly yes
appendfsync everysec
# 守護(hù)進(jìn)程
daemonize yes
# redis配置文件 16371
# 開啟保護(hù)模式
protected-mode yes
# 添加本機(jī)的ip
bind 192.168.1.100
# 端口
port 16371
# pid存儲目錄
pidfile /usr/local/redis/run/redis_16371.pid
# 日志存儲目錄
logfile /usr/local/redis/log/16370/redis_16371.log
# 數(shù)據(jù)存儲目錄，目錄要提前創(chuàng)建好
dir /usr/local/redis/data/16371
# 設(shè)置實(shí)例的驗(yàn)證口令
requirepass dongdong
# 以認(rèn)證的方式連接到master。 如果master中使用了“密碼保護(hù)”，slave必須交付正確的授權(quán)密碼，才能連接成功。
# 此配置項中值需要和master機(jī)器的“requirepass”保持一致
masterauth dongdong
# 配置RDB持久化策略
save 900 1
save 300 10
save 60 10000
# 配置AOF持久化
appendonly yes
appendfsync everysec
# 守護(hù)進(jìn)程
daemonize yes
# redis配置文件 16372
# 開啟保護(hù)模式
protected-mode yes
# 添加本機(jī)的ip
bind 192.168.1.100
# 端口
port 16372
# pid存儲目錄
pidfile /usr/local/redis/run/redis_16372.pid
# 日志存儲目錄
logfile /usr/local/redis/log/16370/redis_16372.log
# 數(shù)據(jù)存儲目錄，目錄要提前創(chuàng)建好
dir /usr/local/redis/data/16372
# 設(shè)置實(shí)例的驗(yàn)證口令
requirepass dongdong
# 以認(rèn)證的方式連接到master。 如果master中使用了“密碼保護(hù)”，slave必須交付正確的授權(quán)密碼，才能連接成功。
# 此配置項中值需要和master機(jī)器的“requirepass”保持一致
masterauth dongdong
# 配置RDB持久化策略
save 900 1
save 300 10
save 60 10000
# 配置AOF持久化
appendonly yes
appendfsync everysec
# 守護(hù)進(jìn)程
daemonize yes
# 復(fù)制下面命令即可
cat > /usr/local/redis/conf/redis_16370.conf <<EOF
# 開啟保護(hù)模式
protected-mode yes
# 添加本機(jī)的ip
bind 192.168.1.100
# 端口
port 16370
# pid存儲目錄
pidfile /usr/local/redis/run/redis_16370.pid
# 日志存儲目錄
logfile /usr/local/redis/log/16370/redis_16370.log
# 數(shù)據(jù)存儲目錄，目錄要提前創(chuàng)建好
dir /usr/local/redis/data/16370
# 設(shè)置實(shí)例的驗(yàn)證口令
requirepass dongdong
# 以認(rèn)證的方式連接到master。 如果master中使用了“密碼保護(hù)”，slave必須交付正確的授權(quán)密碼，才能連接成功。
# 此配置項中值需要和master機(jī)器的“requirepass”保持一致
masterauth dongdong
# 配置RDB持久化策略
save 900 1
save 300 10
save 60 10000
# 配置AOF持久化
appendonly yes
appendfsync everysec
# 守護(hù)進(jìn)程
daemonize yes
EOF
cat > /usr/local/redis/conf/redis_16371.conf <<EOF
# 開啟保護(hù)模式
protected-mode yes
# 添加本機(jī)的ip
bind 192.168.1.100
# 端口
port 16371
# pid存儲目錄
pidfile /usr/local/redis/run/redis_16371.pid
# 日志存儲目錄
logfile /usr/local/redis/log/16370/redis_16371.log
# 數(shù)據(jù)存儲目錄，目錄要提前創(chuàng)建好
dir /usr/local/redis/data/16371
# 設(shè)置實(shí)例的驗(yàn)證口令
requirepass dongdong
# 以認(rèn)證的方式連接到master。 如果master中使用了“密碼保護(hù)”，slave必須交付正確的授權(quán)密碼，才能連接成功。
# 此配置項中值需要和master機(jī)器的“requirepass”保持一致
masterauth dongdong
# 配置RDB持久化策略
save 900 1
save 300 10
save 60 10000
# 配置AOF持久化
appendonly yes
appendfsync everysec
# 守護(hù)進(jìn)程
daemonize yes
EOF
cat > /usr/local/redis/conf/redis_16372.conf <<EOF
# 開啟保護(hù)模式
protected-mode yes
# 添加本機(jī)的ip
bind 192.168.1.100
# 端口
port 16372
# pid存儲目錄
pidfile /usr/local/redis/run/redis_16372.pid
# 日志存儲目錄
logfile /usr/local/redis/log/16370/redis_16372.log
# 數(shù)據(jù)存儲目錄，目錄要提前創(chuàng)建好
dir /usr/local/redis/data/16372
# 設(shè)置實(shí)例的驗(yàn)證口令
requirepass dongdong
# 以認(rèn)證的方式連接到master。 如果master中使用了“密碼保護(hù)”，slave必須交付正確的授權(quán)密碼，才能連接成功。
# 此配置項中值需要和master機(jī)器的“requirepass”保持一致
masterauth dongdong
# 配置RDB持久化策略
save 900 1
save 300 10
save 60 10000
# 配置AOF持久化
appendonly yes
appendfsync everysec
# 守護(hù)進(jìn)程
daemonize yes
EOF
4.2 編寫啟動腳本
# 啟動腳本
cat > /usr/local/redis/bin/master-slave_start.sh << EOF
#!/bin/bash
redis-server /usr/local/redis/conf/redis_16370.conf
redis-server /usr/local/redis/conf/redis_16371.conf
redis-server /usr/local/redis/conf/redis_16372.conf
EOF
# 賦予可執(zhí)行權(quán)限
chmod +x /usr/local/redis/bin/master-slave_start.sh
# 關(guān)閉腳本
vi /usr/local/redis/bin/master-slave_shutdown.sh
#!/bin/bash
kill -9 $(ps -ef | grep 1637 | grep -v grep | awk '{print $2}')
# 賦予可執(zhí)行權(quán)限
chmod +x /usr/local/redis/bin/master-slave_shutdown.sh
5. 開啟主從
5.1 開啟
# 以"192.168.1.100:16370"實(shí)例為主庫，以"192.168.1.100:16371"實(shí)例和"192.168.1.100:16372"實(shí)例為從庫。 # -a 參數(shù)指定密碼 # 從庫實(shí)例開啟主從 redis-cli -h 192.168.1.100 -p 16371 -a dongdong slaveof 192.168.1.100 16370 redis-cli -h 192.168.1.100 -p 16372 -a dongdong slaveof 192.168.1.100 16370 # Warning: Using a password with '-a' or '-u' option on the command line interface may not be safe. # 警告：在命令行界面上使用帶有“-a”或“-u”選項的密碼可能不安全。 可以忽略 # 解決方式 # redis-cli -h 192.168.1.100 -p 16371 -a dongdong --no-auth-warning slaveof 192.168.1.100 16370
5.2 主庫實(shí)例查看主從信息
redis-cli -h 192.168.1.100 -p 16370 -a dongdong info replication [root@master-slave ~]# redis-cli -h 192.168.1.100 -p 16370 -a dongdong info replication Warning: Using a password with '-a' or '-u' option on the command line interface may not be safe. # Replication role:master connected_slaves:2 slave0:ip=192.168.1.100,port=16371,state=online,offset=8428,lag=0 slave1:ip=192.168.1.100,port=16372,state=online,offset=8428,lag=1 master_failover_state:no-failover master_replid:a292603e2c96a6f451055e6d84a60795f81f2928 master_replid2:0000000000000000000000000000000000000000 master_repl_offset:8428 second_repl_offset:-1 repl_backlog_active:1 repl_backlog_size:1048576 repl_backlog_first_byte_offset:1 repl_backlog_histlen:8428
6. 創(chuàng)建sentinel的配置文件并啟動
6.1 創(chuàng)建配置文件
# 創(chuàng)建redis數(shù)據(jù)存儲目錄
mkdir -p /usr/local/redis/data/16379
# 創(chuàng)建redis配置文件存儲目錄
mkdir -p /usr/local/redis/conf/16379
# 創(chuàng)建redis日志存儲目錄
mkdir -p /usr/local/redis/log/16379
# 創(chuàng)建redis pid存儲目錄
mkdir /usr/local/redis/run
cat > /usr/local/redis/conf/sentinel.conf <<EOF
bind 192.168.1.100
port 16379
dir /usr/local/redis/data/16379
sentinel monitor redis_master 192.168.1.100 16370 1
sentinel down-after-milliseconds redis_master 5000
sentinel auth-pass redis_master dongdong
EOF
配置文件參數(shù)說明如下:
sentinel monitor redis_master 192.168.1.100 16370 1:
指定Redis初始集群的master庫節(jié)點(diǎn)并命名為"redis_master"，注意這個"1"的含義，他表示sentinel需要最少的投票數(shù)。這在于多sentinel的場景下很有用，由于本案例只是用了1個sentinel，因此我就設(shè)置為1。
在生產(chǎn)環(huán)境中，建議配置多sentinel的模式，建議最少設(shè)置為3(建議設(shè)置奇數(shù))個sentinel物理節(jié)點(diǎn)，而后將此處的"1"改為"2"，表示3個sentinel中最少有2個節(jié)點(diǎn)認(rèn)為master庫宕機(jī)時，才會真正意義上認(rèn)為master庫宕機(jī)。
sentinel down-after-milliseconds redis_master 5000:
監(jiān)控sentinel集群時，如果超過了5000毫秒(即5秒)仍然沒有響應(yīng)，則sentinel會判定master庫宕機(jī)了。
sentinel auth-pass redis_master dongdong:
由于sentinel需要訪問Redis集群，因此我們要設(shè)置訪問整個集群的密碼，我這里指定的密碼為"dongdong"，這意味著所有的節(jié)點(diǎn)都使用相同的密碼。
6.2 啟動
# 啟動 sentinel進(jìn)程 redis-sentinel /usr/local/redis/conf/sentinel.conf &> /usr/local/redis/log/16379/sentinel_16379.log & root@master ~]# redis-sentin el /usr/local/redis/conf/sentinel.conf &> /usr/local/redis/log/16379/sentinel_16379.log & [1] 29961 # 查看日志 tail -100f /usr/local/redis/log/16379/sentinel_16379.log [root@master ~]# tail -100f /usr/local/redis/log/16379/sentinel_16379.log 30008:X 17 Jul 2024 19:51:15.356 * oO0OoO0OoO0Oo Redis is starting oO0OoO0OoO0Oo 30008:X 17 Jul 2024 19:51:15.356 * Redis version=7.2.5, bits=64, commit=00000000, modified=0, pid=30008, just started 30008:X 17 Jul 2024 19:51:15.356 * Configuration loaded 30008:X 17 Jul 2024 19:51:15.356 * Increased maximum number of open files to 10032 (it was originally set to 1024). 30008:X 17 Jul 2024 19:51:15.356 * monotonic clock: POSIX clock_gettime 30008:X 17 Jul 2024 19:51:15.357 * Running mode=sentinel, port=16379. 30008:X 17 Jul 2024 19:51:15.357 # WARNING: The TCP backlog setting of 511 cannot be enforced because /proc/sys/net/core/somaxconn is set to the lower value of 128. 30008:X 17 Jul 2024 19:51:15.357 * Sentinel ID is a91bf648c16246f3f7c7e1aa2b1fc6fad38f772b 30008:X 17 Jul 2024 19:51:15.357 # +monitor master redis_master 192.168.1.100 16370 quorum 1 # 通過redis-cli 進(jìn)入 查看 redis-cli -h 192.168.1.100 -p 16379 [root@master ~]# redis-cli -h 192.168.1.100 -p 16379 192.168.1.100:16379> ping PONG 192.168.1.100:16379>
6.3 手動停止主庫運(yùn)行，模擬主庫宕機(jī)
# 切換前主從狀態(tài)正常
[root@master ~]# ss -npl | grep redis
tcp LISTEN 0 128 192.168.1.100:16370 *:* users:(("redis-server",pid=29871,fd=6))
tcp LISTEN 0 128 192.168.1.100:16371 *:* users:(("redis-server",pid=29873,fd=6))
tcp LISTEN 0 128 192.168.1.100:16372 *:* users:(("redis-server",pid=29881,fd=6))
tcp LISTEN 0 128 192.168.1.100:16379 *:* users:(("redis-sentinel",pid=30008,fd=6))
# 查看集群狀態(tài) master
redis-cli -h 192.168.1.100 -p 16370 -a dongdong info replication
[root@master ~]# redis-cli -h 192.168.1.100 -p 16370 -a dongdong info replication
Warning: Using a password with '-a' or '-u' option on the command line interface may not be safe.
# Replication
role:master
connected_slaves:2
slave0:ip=192.168.1.100,port=16371,state=online,offset=111085,lag=0
slave1:ip=192.168.1.100,port=16372,state=online,offset=111085,lag=1
master_failover_state:no-failover
master_replid:208227bbd586d90841dcf753ba669dbde246bba7
master_replid2:0000000000000000000000000000000000000000
master_repl_offset:111085
second_repl_offset:-1
repl_backlog_active:1
repl_backlog_size:1048576
repl_backlog_first_byte_offset:1
repl_backlog_histlen:111085
# 查看集群狀態(tài) slaver
redis-cli -h 192.168.1.100 -p 16371 -a dongdong info replication
[root@master ~]# redis-cli -h 192.168.1.100 -p 16371 -a dongdong info replication
Warning: Using a password with '-a' or '-u' option on the command line interface may not be safe.
# Replication
role:slave
master_host:192.168.1.100
master_port:16370
master_link_status:up
master_last_io_seconds_ago:2
master_sync_in_progress:0
slave_read_repl_offset:118087
slave_repl_offset:118087
slave_priority:100
slave_read_only:1
replica_announced:1
connected_slaves:0
master_failover_state:no-failover
master_replid:208227bbd586d90841dcf753ba669dbde246bba7
master_replid2:0000000000000000000000000000000000000000
master_repl_offset:118087
second_repl_offset:-1
repl_backlog_active:1
repl_backlog_size:1048576
repl_backlog_first_byte_offset:15
repl_backlog_histlen:118073
# 查看集群狀態(tài) slaver
redis-cli -h 192.168.1.100 -p 16372 -a dongdong info replication
[root@master ~]# redis-cli -h 192.168.1.100 -p 16372 -a dongdong info replication
Warning: Using a password with '-a' or '-u' option on the command line interface may not be safe.
# Replication
role:slave
master_host:192.168.1.100
master_port:16370
master_link_status:up
master_last_io_seconds_ago:1
master_sync_in_progress:0
slave_read_repl_offset:126869
slave_repl_offset:126869
slave_priority:100
slave_read_only:1
replica_announced:1
connected_slaves:0
master_failover_state:no-failover
master_replid:208227bbd586d90841dcf753ba669dbde246bba7
master_replid2:0000000000000000000000000000000000000000
master_repl_offset:126869
second_repl_offset:-1
repl_backlog_active:1
repl_backlog_size:1048576
repl_backlog_first_byte_offset:15
repl_backlog_histlen:126855
# 手動將主庫宕機(jī)，觀察從庫被sentinel實(shí)現(xiàn)了自動切換:
redis-cli -h 192.168.1.100 -p 16370 -a dongdong shutdown
[root@master ~]# redis-cli -h 192.168.1.100 -p 16370 -a dongdong shutdown
Warning: Using a password with '-a' or '-u' option on the command line interface may not be safe.
# 再查看集群狀態(tài) master
redis-cli -h 192.168.1.100 -p 16370 -a dongdong info replication # 由于將主庫服務(wù)停了，因此無法再次連接主庫
[root@master ~]# redis-cli -h 192.168.1.100 -p 16370 -a dongdong info replication
Warning: Using a password with '-a' or '-u' option on the command line interface may not be safe.
Could not connect to Redis at 192.168.1.100:16370: Connection refused
# 再查看集群狀態(tài) slaves 發(fā)現(xiàn)變成了master節(jié)點(diǎn)
redis-cli -h 192.168.1.100 -p 16371 -a dongdong info replication
[root@master ~]# redis-cli -h 192.168.1.100 -p 16371 -a dongdong info replication
Warning: Using a password with '-a' or '-u' option on the command line interface may not be safe.
# Replication
role:master
connected_slaves:1
slave0:ip=192.168.1.100,port=16372,state=online,offset=150234,lag=0
master_failover_state:no-failover
master_replid:23729dc7d7aeafa6f85a2a647442965c80b379ed
master_replid2:208227bbd586d90841dcf753ba669dbde246bba7
master_repl_offset:150234
second_repl_offset:134150
repl_backlog_active:1
repl_backlog_size:1048576
repl_backlog_first_byte_offset:15
repl_backlog_histlen:150220
# 查看集群狀態(tài) slaver
redis-cli -h 192.168.1.100 -p 16372 -a dongdong info replication
[root@master ~]# redis-cli -h 192.168.1.100 -p 16372 -a dongdong info replication
Warning: Using a password with '-a' or '-u' option on the command line interface may not be safe.
# Replication
role:slave
master_host:192.168.1.100
master_port:16371
master_link_status:up
master_last_io_seconds_ago:2
master_sync_in_progress:0
slave_read_repl_offset:154406
slave_repl_offset:154406
slave_priority:100
slave_read_only:1
replica_announced:1
connected_slaves:0
master_failover_state:no-failover
master_replid:23729dc7d7aeafa6f85a2a647442965c80b379ed
master_replid2:208227bbd586d90841dcf753ba669dbde246bba7
master_repl_offset:154406
second_repl_offset:134150
repl_backlog_active:1
repl_backlog_size:1048576
repl_backlog_first_byte_offset:15
repl_backlog_histlen:154392
# 查看sentinel的日志信息，不難發(fā)現(xiàn)有記錄切換的過程
tail -100f /usr/local/redis/log/16379/sentinel_16379.log
[root@master ~]# tail -100f /usr/local/redis/log/16379/sentinel_16379.log
30008:X 17 Jul 2024 19:51:15.356 * oO0OoO0OoO0Oo Redis is starting oO0OoO0OoO0Oo
30008:X 17 Jul 2024 19:51:15.356 * Redis version=7.2.5, bits=64, commit=00000000, modified=0, pid=30008, just started
30008:X 17 Jul 2024 19:51:15.356 * Configuration loaded
30008:X 17 Jul 2024 19:51:15.356 * Increased maximum number of open files to 10032 (it was originally set to 1024).
30008:X 17 Jul 2024 19:51:15.356 * monotonic clock: POSIX clock_gettime
30008:X 17 Jul 2024 19:51:15.357 * Running mode=sentinel, port=16379.
30008:X 17 Jul 2024 19:51:15.357 # WARNING: The TCP backlog setting of 511 cannot be enforced because /proc/sys/net/core/somaxconn is set to the lower value of 128.
30008:X 17 Jul 2024 19:51:15.357 * Sentinel ID is a91bf648c16246f3f7c7e1aa2b1fc6fad38f772b
30008:X 17 Jul 2024 19:51:15.357 # +monitor master redis_master 192.168.1.100 16370 quorum 1
30008:X 17 Jul 2024 20:15:40.962 # +sdown master redis_master 192.168.1.100 16370
30008:X 17 Jul 2024 20:15:40.962 # +odown master redis_master 192.168.1.100 16370 #quorum 1/1
30008:X 17 Jul 2024 20:15:40.962 # +new-epoch 1
30008:X 17 Jul 2024 20:15:40.962 # +try-failover master redis_master 192.168.1.100 16370
30008:X 17 Jul 2024 20:15:40.968 * Sentinel new configuration saved on disk
30008:X 17 Jul 2024 20:15:40.968 # +vote-for-leader a91bf648c16246f3f7c7e1aa2b1fc6fad38f772b 1
30008:X 17 Jul 2024 20:15:40.968 # +elected-leader master redis_master 192.168.1.100 16370
30008:X 17 Jul 2024 20:15:40.968 # +failover-state-select-slave master redis_master 192.168.1.100 16370
30008:X 17 Jul 2024 20:15:41.026 # +selected-slave slave 192.168.1.100:16371 192.168.1.100 16371 @ redis_master 192.168.1.100 16370
30008:X 17 Jul 2024 20:15:41.027 * +failover-state-send-slaveof-noone slave 192.168.1.100:16371 192.168.1.100 16371 @ redis_master 192.168.1.100 16370
30008:X 17 Jul 2024 20:15:41.093 * +failover-state-wait-promotion slave 192.168.1.100:16371 192.168.1.100 16371 @ redis_master 192.168.1.100 16370
30008:X 17 Jul 2024 20:15:42.015 * Sentinel new configuration saved on disk
30008:X 17 Jul 2024 20:15:42.015 # +promoted-slave slave 192.168.1.100:16371 192.168.1.100 16371 @ redis_master 192.168.1.100 16370
30008:X 17 Jul 2024 20:15:42.015 # +failover-state-reconf-slaves master redis_master 192.168.1.100 16370
30008:X 17 Jul 2024 20:15:42.096 * +slave-reconf-sent slave 192.168.1.100:16372 192.168.1.100 16372 @ redis_master 192.168.1.100 16370
30008:X 17 Jul 2024 20:15:43.040 * +slave-reconf-inprog slave 192.168.1.100:16372 192.168.1.100 16372 @ redis_master 192.168.1.100 16370
30008:X 17 Jul 2024 20:15:43.040 * +slave-reconf-done slave 192.168.1.100:16372 192.168.1.100 16372 @ redis_master 192.168.1.100 16370
30008:X 17 Jul 2024 20:15:43.094 # +failover-end master redis_master 192.168.1.100 16370
30008:X 17 Jul 2024 20:15:43.094 # +switch-master redis_master 192.168.1.100 16370 192.168.1.100 16371
30008:X 17 Jul 2024 20:15:43.095 * +slave slave 192.168.1.100:16372 192.168.1.100 16372 @ redis_master 192.168.1.100 16371
30008:X 17 Jul 2024 20:15:43.095 * +slave slave 192.168.1.100:16370 192.168.1.100 16370 @ redis_master 192.168.1.100 16371
30008:X 17 Jul 2024 20:15:43.100 * Sentinel new configuration saved on disk
30008:X 17 Jul 2024 20:15:48.121 # +sdown slave 192.168.1.100:16370 192.168.1.100 16370 @ redis_master 192.168.1.100 16371
6.4 手動修復(fù)宕機(jī)的主庫，sentinel會自動發(fā)現(xiàn)
# 修復(fù)之前，16370端口是連接不上的 redis-cli -h 192.168.1.100 -p 16370 -a dongdong info replication [root@master ~]# redis-cli -h 192.168.1.100 -p 16370 -a dongdong info replication Warning: Using a password with '-a' or '-u' option on the command line interface may not be safe. Could not connect to Redis at 192.168.1.100:16370: Connection refused [root@master ~]# redis-cli -h 192.168.1.100 -p 16371 -a dongdong info replication Warning: Using a password with '-a' or '-u' option on the command line interface may not be safe. # Replication role:master connected_slaves:1 slave0:ip=192.168.1.100,port=16372,state=online,offset=190264,lag=1 master_failover_state:no-failover master_replid:23729dc7d7aeafa6f85a2a647442965c80b379ed master_replid2:208227bbd586d90841dcf753ba669dbde246bba7 master_repl_offset:190410 second_repl_offset:134150 repl_backlog_active:1 repl_backlog_size:1048576 repl_backlog_first_byte_offset:15 repl_backlog_histlen:190396 [root@master ~]# redis-cli -h 192.168.1.100 -p 16372 -a dongdong info replication Warning: Using a password with '-a' or '-u' option on the command line interface may not be safe. # Replication role:slave master_host:192.168.1.100 master_port:16371 master_link_status:up master_last_io_seconds_ago:1 master_sync_in_progress:0 slave_read_repl_offset:192204 slave_repl_offset:192204 slave_priority:100 slave_read_only:1 replica_announced:1 connected_slaves:0 master_failover_state:no-failover master_replid:23729dc7d7aeafa6f85a2a647442965c80b379ed master_replid2:208227bbd586d90841dcf753ba669dbde246bba7 master_repl_offset:192204 second_repl_offset:134150 repl_backlog_active:1 repl_backlog_size:1048576 repl_backlog_first_byte_offset:15 repl_backlog_histlen:192190 # 修復(fù)之后，之前宕機(jī)的主庫會自動加入集群并成為slaves節(jié)點(diǎn) redis-server /usr/local/redis/conf/redis_16370.conf [root@master ~]# redis-server /usr/local/redis/conf/redis_16370.conf [root@master ~]# redis-cli -h 192.168.1.100 -p 16370 -a dongdong info replication Warning: Using a password with '-a' or '-u' option on the command line interface may not be safe. # Replication role:slave master_host:192.168.1.100 master_port:16371 master_link_status:up master_last_io_seconds_ago:1 master_sync_in_progress:0 slave_read_repl_offset:199215 slave_repl_offset:199215 slave_priority:100 slave_read_only:1 replica_announced:1 connected_slaves:0 master_failover_state:no-failover master_replid:23729dc7d7aeafa6f85a2a647442965c80b379ed master_replid2:0000000000000000000000000000000000000000 master_repl_offset:199215 second_repl_offset:-1 repl_backlog_active:1 repl_backlog_size:1048576 repl_backlog_first_byte_offset:199047 repl_backlog_histlen:169 # 查看 redis-cli -h 192.168.1.100 -p 16371 -a dongdong info replication redis-cli -h 192.168.1.100 -p 16372 -a dongdong info replication [root@master ~]# redis-cli -h 192.168.1.100 -p 16371 -a dongdong info replication Warning: Using a password with '-a' or '-u' option on the command line interface may not be safe. # Replication role:master connected_slaves:2 # slaves 變成兩個了 slave0:ip=192.168.1.100,port=16372,state=online,offset=203533,lag=1 slave1:ip=192.168.1.100,port=16370,state=online,offset=203533,lag=1 master_failover_state:no-failover master_replid:23729dc7d7aeafa6f85a2a647442965c80b379ed master_replid2:208227bbd586d90841dcf753ba669dbde246bba7 master_repl_offset:203533 second_repl_offset:134150 repl_backlog_active:1 repl_backlog_size:1048576 repl_backlog_first_byte_offset:15 repl_backlog_histlen:203519 [root@master ~]# redis-cli -h 192.168.1.100 -p 16372 -a dongdong info replication Warning: Using a password with '-a' or '-u' option on the command line interface may not be safe. # Replication role:slave master_host:192.168.1.100 master_port:16371 master_link_status:up master_last_io_seconds_ago:1 master_sync_in_progress:0 slave_read_repl_offset:207851 slave_repl_offset:207851 slave_priority:100 slave_read_only:1 replica_announced:1 connected_slaves:0 master_failover_state:no-failover master_replid:23729dc7d7aeafa6f85a2a647442965c80b379ed master_replid2:208227bbd586d90841dcf753ba669dbde246bba7 master_repl_offset:207851 second_repl_offset:134150 repl_backlog_active:1 repl_backlog_size:1048576 repl_backlog_first_byte_offset:15 repl_backlog_histlen:207837 # 注意觀察sentinel日志喲，會有新日志記錄 tail -100f /usr/local/redis/log/16379/sentinel_16379.log [root@master ~]# tail -100f /usr/local/redis/log/16379/sentinel_16379.log 30008:X 17 Jul 2024 19:51:15.356 * oO0OoO0OoO0Oo Redis is starting oO0OoO0OoO0Oo 30008:X 17 Jul 2024 19:51:15.356 * Redis version=7.2.5, bits=64, commit=00000000, modified=0, pid=30008, just started 30008:X 17 Jul 2024 19:51:15.356 * Configuration loaded 30008:X 17 Jul 2024 19:51:15.356 * Increased maximum number of open files to 10032 (it was originally set to 1024). 30008:X 17 Jul 2024 19:51:15.356 * monotonic clock: POSIX clock_gettime 30008:X 17 Jul 2024 19:51:15.357 * Running mode=sentinel, port=16379. 30008:X 17 Jul 2024 19:51:15.357 # WARNING: The TCP backlog setting of 511 cannot be enforced because /proc/sys/net/core/somaxconn is set to the lower value of 128. 30008:X 17 Jul 2024 19:51:15.357 * Sentinel ID is a91bf648c16246f3f7c7e1aa2b1fc6fad38f772b 30008:X 17 Jul 2024 19:51:15.357 # +monitor master redis_master 192.168.1.100 16370 quorum 1 30008:X 17 Jul 2024 20:15:40.962 # +sdown master redis_master 192.168.1.100 16370 30008:X 17 Jul 2024 20:15:40.962 # +odown master redis_master 192.168.1.100 16370 #quorum 1/1 30008:X 17 Jul 2024 20:15:40.962 # +new-epoch 1 30008:X 17 Jul 2024 20:15:40.962 # +try-failover master redis_master 192.168.1.100 16370 30008:X 17 Jul 2024 20:15:40.968 * Sentinel new configuration saved on disk 30008:X 17 Jul 2024 20:15:40.968 # +vote-for-leader a91bf648c16246f3f7c7e1aa2b1fc6fad38f772b 1 30008:X 17 Jul 2024 20:15:40.968 # +elected-leader master redis_master 192.168.1.100 16370 30008:X 17 Jul 2024 20:15:40.968 # +failover-state-select-slave master redis_master 192.168.1.100 16370 30008:X 17 Jul 2024 20:15:41.026 # +selected-slave slave 192.168.1.100:16371 192.168.1.100 16371 @ redis_master 192.168.1.100 16370 30008:X 17 Jul 2024 20:15:41.027 * +failover-state-send-slaveof-noone slave 192.168.1.100:16371 192.168.1.100 16371 @ redis_master 192.168.1.100 16370 30008:X 17 Jul 2024 20:15:41.093 * +failover-state-wait-promotion slave 192.168.1.100:16371 192.168.1.100 16371 @ redis_master 192.168.1.100 16370 30008:X 17 Jul 2024 20:15:42.015 * Sentinel new configuration saved on disk 30008:X 17 Jul 2024 20:15:42.015 # +promoted-slave slave 192.168.1.100:16371 192.168.1.100 16371 @ redis_master 192.168.1.100 16370 30008:X 17 Jul 2024 20:15:42.015 # +failover-state-reconf-slaves master redis_master 192.168.1.100 16370 30008:X 17 Jul 2024 20:15:42.096 * +slave-reconf-sent slave 192.168.1.100:16372 192.168.1.100 16372 @ redis_master 192.168.1.100 16370 30008:X 17 Jul 2024 20:15:43.040 * +slave-reconf-inprog slave 192.168.1.100:16372 192.168.1.100 16372 @ redis_master 192.168.1.100 16370 30008:X 17 Jul 2024 20:15:43.040 * +slave-reconf-done slave 192.168.1.100:16372 192.168.1.100 16372 @ redis_master 192.168.1.100 16370 30008:X 17 Jul 2024 20:15:43.094 # +failover-end master redis_master 192.168.1.100 16370 30008:X 17 Jul 2024 20:15:43.094 # +switch-master redis_master 192.168.1.100 16370 192.168.1.100 16371 30008:X 17 Jul 2024 20:15:43.095 * +slave slave 192.168.1.100:16372 192.168.1.100 16372 @ redis_master 192.168.1.100 16371 30008:X 17 Jul 2024 20:15:43.095 * +slave slave 192.168.1.100:16370 192.168.1.100 16370 @ redis_master 192.168.1.100 16371 30008:X 17 Jul 2024 20:15:43.100 * Sentinel new configuration saved on disk 30008:X 17 Jul 2024 20:15:48.121 # +sdown slave 192.168.1.100:16370 192.168.1.100 16370 @ redis_master 192.168.1.100 16371 30008:X 17 Jul 2024 20:30:15.529 # -sdown slave 192.168.1.100:16370 192.168.1.100 16370 @ redis_master 192.168.1.100 16371 30008:X 17 Jul 2024 20:30:25.504 * +convert-to-slave slave 192.168.1.100:16370 192.168.1.100 16370 @ redis_master 192.168.1.100 16371 # 除了提到上面的變化外，還有一點(diǎn)變化就是配置文件也修改了，每個"slave"節(jié)點(diǎn)都新增了 下面三行 本案例只有192.168.1.100:16370 和 192.168.1.100:16372是 slave # Generated by CONFIG REWRITE replicaof 192.168.1.100 16371 latency-tracking-info-percentiles 50 99 99.9 user default on sanitize-payload #162fffe9f429e1fa5d351c11f5a7f2f3031806c5125eeb84993f68738cbb6c84 ~* &* +@all # Generated by CONFIG REWRITE replicaof 192.168.1.100 16371 latency-tracking-info-percentiles 50 99 99.9 user default on sanitize-payload #162fffe9f429e1fa5d351c11f5a7f2f3031806c5125eeb84993f68738cbb6c84 ~* &* +@all
二、多節(jié)點(diǎn)哨兵
一主兩從三哨兵
一共三臺機(jī)器，每臺機(jī)器部署一個哨兵
注意： 后面啟動哨兵可能會出現(xiàn)兩個警告
WARNING Memory overcommit must be enabled! Without it, a background save or replication may fail under low memory condition. To fix this issue add ‘vm.overcommit_memory = 1’ to /etc/sysctl.conf and then reboot or run the command ‘sysctl vm.overcommit_memory=1’ for this to take effect.
警告：必須啟用內(nèi)存超量使用！沒有它，在內(nèi)存不足的情況下，后臺保存或復(fù)制可能會失敗。要解決此問題，請將“vm.overcommit_memory=1”添加到/etc/sysctl.conf中，然后重新啟動或運(yùn)行命令“sysctl vm.overcommit_memory=1”以使其生效。
修改內(nèi)存分配控制
vi /etc/sysctl.conf vm.overcommit_memory = 1
使其配置文件生效
sysctl -p /etc/sysctl.conf
WARNING: The TCP backlog setting of 511 cannot be enforced because /proc/sys/net/core/somaxconn is set to the lower value of 128.
警告：無法強(qiáng)制執(zhí)行TCP積壓設(shè)置511，因?yàn)?proc/sys/net/core/somaxconn設(shè)置為較低的值128。
vi /etc/sysctl.conf net.core.somaxconn= 1024 sysctl -p /etc/sysctl.conf
| 操作系統(tǒng) | Centos 7 | Centos 7 | Centos 7 |
|---|---|---|---|
| 內(nèi)核版本 | Linux 3.10.0-957.el7.x86_64 | Linux 3.10.0-957.el7.x86_64 | Linux 3.10.0-957.el7.x86_64 |
| 主機(jī)名稱 | redis_master | redis_salve1 | redis_salve2 |
| IP | 192.168.1.100 | 192.168.1.200 | 192.168.1.250 |
| 端口 | 16370、26370 | 16371、26370 | 16372、26370 |
安裝前準(zhǔn)備工作已經(jīng)Redis安裝,請參照上面步驟進(jìn)行
1.Redis配置文件修改并啟動
1.1 redis_master 節(jié)點(diǎn)
# 創(chuàng)建redis數(shù)據(jù)存儲目錄 mkdir -p /usr/local/redis/data/16370 # 創(chuàng)建redis配置文件存儲目錄 mkdir -p /usr/local/redis/conf/16370 # 創(chuàng)建redis日志存儲目錄 mkdir -p /usr/local/redis/log/16370 # 創(chuàng)建redis pid存儲目錄 mkdir /usr/local/redis/run/16370 # 修改配置文件 # 復(fù)制下面命令即可 cat > /usr/local/redis/conf/16370/redis_16370.conf <<EOF # 開啟保護(hù)模式 protected-mode yes # 添加本機(jī)的ip bind 192.168.1.100 # 端口 port 16370 # pid存儲目錄 pidfile /usr/local/redis/run/16370/redis_16370.pid # 日志存儲目錄 logfile /usr/local/redis/log/16370/redis_16370.log # 數(shù)據(jù)存儲目錄，目錄要提前創(chuàng)建好 dir /usr/local/redis/data/16370 # 設(shè)置實(shí)例的驗(yàn)證口令 requirepass dongdong # 以認(rèn)證的方式連接到master。 如果master中使用了“密碼保護(hù)”，slave必須交付正確的授權(quán)密碼，才能連接成功。 # 此配置項中值需要和master機(jī)器的“requirepass”保持一致 masterauth dongdong # 配置RDB持久化策略 save 900 1 save 300 10 save 60 10000 # 配置AOF持久化 appendonly yes appendfsync everysec # 守護(hù)進(jìn)程 daemonize yes EOF # 啟動Redis redis-server /usr/local/redis/conf/16370/redis_16370.conf
1.2 redis_slave1 節(jié)點(diǎn)
# 創(chuàng)建redis數(shù)據(jù)存儲目錄 mkdir -p /usr/local/redis/data/16371 # 創(chuàng)建redis配置文件存儲目錄 mkdir -p /usr/local/redis/conf/16371 # 創(chuàng)建redis日志存儲目錄 mkdir -p /usr/local/redis/log/16371 # 創(chuàng)建redis pid存儲目錄 mkdir /usr/local/redis/run/16371 # 修改配置文件 # 復(fù)制下面命令即可 cat > /usr/local/redis/conf/16371/redis_16371.conf <<EOF # 開啟保護(hù)模式 protected-mode yes # 添加本機(jī)的ip bind 192.168.1.200 # 端口 port 16371 # pid存儲目錄 pidfile /usr/local/redis/run/16371/redis_16371.pid # 日志存儲目錄 logfile /usr/local/redis/log/16371/redis_16371.log # 數(shù)據(jù)存儲目錄，目錄要提前創(chuàng)建好 dir /usr/local/redis/data/16371 # 設(shè)置實(shí)例的驗(yàn)證口令 requirepass dongdong # 以認(rèn)證的方式連接到master。 如果master中使用了“密碼保護(hù)”，slave必須交付正確的授權(quán)密碼，才能連接成功。 # 此配置項中值需要和master機(jī)器的“requirepass”保持一致 masterauth dongdong # 配置RDB持久化策略 save 900 1 save 300 10 save 60 10000 # 配置AOF持久化 appendonly yes appendfsync everysec # 守護(hù)進(jìn)程 daemonize yes EOF # 啟動Redis redis-server /usr/local/redis/conf/16371/redis_16371.conf
1.3 redis_slave2 節(jié)點(diǎn)
# 創(chuàng)建redis數(shù)據(jù)存儲目錄 mkdir -p /usr/local/redis/data/16372 # 創(chuàng)建redis配置文件存儲目錄 mkdir -p /usr/local/redis/conf/16372 # 創(chuàng)建redis日志存儲目錄 mkdir -p /usr/local/redis/log/16372 # 創(chuàng)建redis pid存儲目錄 mkdir /usr/local/redis/run/16372 # 修改配置文件 # 復(fù)制下面命令即可 cat > /usr/local/redis/conf/16372/redis_16372.conf <<EOF # 開啟保護(hù)模式 protected-mode yes # 添加本機(jī)的ip bind 192.168.1.250 # 端口 port 16372 # pid存儲目錄 pidfile /usr/local/redis/run/16372/redis_16372.pid # 日志存儲目錄 logfile /usr/local/redis/log/16372/redis_16372.log # 數(shù)據(jù)存儲目錄，目錄要提前創(chuàng)建好 dir /usr/local/redis/data/16372 # 設(shè)置實(shí)例的驗(yàn)證口令 requirepass dongdong # 以認(rèn)證的方式連接到master。 如果master中使用了“密碼保護(hù)”，slave必須交付正確的授權(quán)密碼，才能連接成功。 # 此配置項中值需要和master機(jī)器的“requirepass”保持一致 masterauth dongdong # 配置RDB持久化策略 save 900 1 save 300 10 save 60 10000 # 配置AOF持久化 appendonly yes appendfsync everysec # 守護(hù)進(jìn)程 daemonize yes EOF # 啟動Redis redis-server /usr/local/redis/conf/16372/redis_16372.conf
2. 開啟主從
2.1 redis_slave1 節(jié)點(diǎn)
# 從庫實(shí)例開啟主從 redis-cli -h 192.168.1.200 -p 16371 -a dongdong slaveof 192.168.1.100 16370 [root@redis_slave1 ~]# redis-cli -h 192.168.1.200 -p 16371 -a dongdong slaveof 192.168.1.100 16370 Warning: Using a password with '-a' or '-u' option on the command line interface may not be safe. OK # 從庫實(shí)例查看主從信息 redis-cli -h 192.168.1.200 -p 16371 -a dongdong info replication [root@redis_slave1 ~]# redis-cli -h 192.168.1.200 -p 16371 -a dongdong info replication Warning: Using a password with '-a' or '-u' option on the command line interface may not be safe. # Replication role:slave master_host:192.168.1.100 master_port:16370 master_link_status:up master_last_io_seconds_ago:9 master_sync_in_progress:0 slave_read_repl_offset:210 slave_repl_offset:210 slave_priority:100 slave_read_only:1 replica_announced:1 connected_slaves:0 master_failover_state:no-failover master_replid:466362240f16a7414b9236f931c6f97252ee00be master_replid2:0000000000000000000000000000000000000000 master_repl_offset:210 second_repl_offset:-1 repl_backlog_active:1 repl_backlog_size:1048576 repl_backlog_first_byte_offset:15 repl_backlog_histlen:196
2.2 redis_slave2 節(jié)點(diǎn)
# 從庫實(shí)例開啟主從 redis-cli -h 192.168.1.250 -p 16372 -a dongdong slaveof 192.168.1.100 16370 [root@redis_slave2 ~]# redis-cli -h 192.168.1.250 -p 16372 -a dongdong slaveof 192.168.1.100 16370 Warning: Using a password with '-a' or '-u' option on the command line interface may not be safe. OK # 從庫實(shí)例查看主從信息 redis-cli -h 192.168.1.250 -p 16372 -a dongdong info replication [root@redis_slave2 ~]# redis-cli -h 192.168.1.250 -p 16372 -a dongdong info replication Warning: Using a password with '-a' or '-u' option on the command line interface may not be safe. # Replication role:slave master_host:192.168.1.100 master_port:16370 master_link_status:up master_last_io_seconds_ago:10 master_sync_in_progress:0 slave_read_repl_offset:266 slave_repl_offset:266 slave_priority:100 slave_read_only:1 replica_announced:1 connected_slaves:0 master_failover_state:no-failover master_replid:466362240f16a7414b9236f931c6f97252ee00be master_replid2:0000000000000000000000000000000000000000 master_repl_offset:266 second_repl_offset:-1 repl_backlog_active:1 repl_backlog_size:1048576 repl_backlog_first_byte_offset:43 repl_backlog_histlen:224
2.3 redis_master 節(jié)點(diǎn)
# 主庫實(shí)例查看主從信息 redis-cli -h 192.168.1.100 -p 16370 -a dongdong info replication [root@redis_master ~]# redis-cli -h 192.168.1.100 -p 16370 -a dongdong info replication Warning: Using a password with '-a' or '-u' option on the command line interface may not be safe. # Replication role:master connected_slaves:2 slave0:ip=192.168.1.200,port=16371,state=online,offset=42,lag=0 slave1:ip=192.168.1.250,port=16372,state=online,offset=42,lag=1 master_failover_state:no-failover master_replid:466362240f16a7414b9236f931c6f97252ee00be master_replid2:0000000000000000000000000000000000000000 master_repl_offset:56 second_repl_offset:-1 repl_backlog_active:1 repl_backlog_size:1048576 repl_backlog_first_byte_offset:1 repl_backlog_histlen:56
3. 編寫sentinel配置文件并啟動
3.1 redis_master 節(jié)點(diǎn)
# 創(chuàng)建redis數(shù)據(jù)存儲目錄 mkdir -p /usr/local/redis/data/26370 # 創(chuàng)建redis配置文件存儲目錄 mkdir -p /usr/local/redis/conf/26370 # 創(chuàng)建redis日志存儲目錄 mkdir -p /usr/local/redis/log/26370 # 創(chuàng)建redis pid存儲目錄 mkdir /usr/local/redis/run/26370 cat > /usr/local/redis/conf/26370/sentinel.conf <<EOF bind 192.168.1.100 port 26370 dir /usr/local/redis/data/26370 sentinel monitor redis_master 192.168.1.100 16370 2 sentinel down-after-milliseconds redis_master 5000 sentinel auth-pass redis_master dongdong EOF # 啟動 sentinel進(jìn)程 redis-sentinel /usr/local/redis/conf/26370/sentinel.conf &> /usr/local/redis/log/26370/sentinel_26370.log & [root@redis_master ~]# redis-sentinel /usr/local/redis/conf/26370/sentinel.conf &> /usr/local/redis/log/26370/sentinel_26370.log & [1] 30569 # 查看redis-sentinel 進(jìn)程 [root@redis_master ~]# ps -ef | grep redis root 30552 1 0 10:17 ? 00:00:01 redis-server 192.168.1.100:16370 root 30569 30499 0 10:33 pts/1 00:00:00 redis-sentinel 192.168.1.100:26370 [sentinel] # 已經(jīng)成功啟動了 root 30574 30499 0 10:34 pts/1 00:00:00 grep --color=auto redis # 查看日志信息 tail -100f /usr/local/redis/log/26370/sentinel_26370.log [root@redis_master ~]# tail -100f /usr/local/redis/log/26370/sentinel_26370.log 30569:X 18 Jul 2024 10:33:50.636 * oO0OoO0OoO0Oo Redis is starting oO0OoO0OoO0Oo 30569:X 18 Jul 2024 10:33:50.636 * Redis version=7.2.5, bits=64, commit=00000000, modified=0, pid=30569, just started 30569:X 18 Jul 2024 10:33:50.636 * Configuration loaded 30569:X 18 Jul 2024 10:33:50.636 * Increased maximum number of open files to 10032 (it was originally set to 1024). 30569:X 18 Jul 2024 10:33:50.636 * monotonic clock: POSIX clock_gettime 30569:X 18 Jul 2024 10:33:50.637 * Running mode=sentinel, port=26370. 30569:X 18 Jul 2024 10:33:50.637 # WARNING: The TCP backlog setting of 511 cannot be enforced because /proc/sys/net/core/somaxconn is set to the lower value of 128. 30569:X 18 Jul 2024 10:33:50.638 * Sentinel new configuration saved on disk 30569:X 18 Jul 2024 10:33:50.639 * Sentinel ID is 088b671eb4ac4c8d8a57805c1be88e4f76cf7ff8 30569:X 18 Jul 2024 10:33:50.639 # +monitor master redis_master 192.168.1.100 26370 quorum 2 30569:X 18 Jul 2024 10:33:52.638 * +sentinel sentinel 088b671eb4ac4c8d8a57805c1be88e4f76cf7ff8 192.168.1.100 26370 @ redis_master 192.168.1.100 26370 30569:X 18 Jul 2024 10:33:52.644 * Sentinel new configuration saved on disk
3.2 redis_slave1 節(jié)點(diǎn)
# 創(chuàng)建redis數(shù)據(jù)存儲目錄 mkdir -p /usr/local/redis/data/26370 # 創(chuàng)建redis配置文件存儲目錄 mkdir -p /usr/local/redis/conf/26370 # 創(chuàng)建redis日志存儲目錄 mkdir -p /usr/local/redis/log/26370 # 創(chuàng)建redis pid存儲目錄 mkdir /usr/local/redis/run/26370 cat > /usr/local/redis/conf/26370/sentinel.conf <<EOF bind 192.168.1.200 port 26370 dir /usr/local/redis/data/26370 sentinel monitor redis_master 192.168.1.100 16370 2 sentinel down-after-milliseconds redis_master 5000 sentinel auth-pass redis_master dongdong EOF # 啟動 sentinel進(jìn)程 redis-sentinel /usr/local/redis/conf/26370/sentinel.conf &> /usr/local/redis/log/26370/sentinel_26370.log & [root@redis_slave1 ~]# redis-sentinel /usr/local/redis/conf/26370/sentinel.conf &> /usr/local/redis/log/26370/sentinel_26370.log & [1] 38894 # 查看redis-sentinel 進(jìn)程 [root@redis_slave1 ~]# ps -ef | grep redis root 38879 1 0 10:19 ? 00:00:01 redis-server 192.168.1.200:16371 root 38894 38840 0 10:36 pts/0 00:00:00 redis-sentinel 192.168.1.200:26370 [sentinel] root 38899 38840 0 10:37 pts/0 00:00:00 grep --color=auto redis # 查看日志信息 tail -100f /usr/local/redis/log/26370/sentinel_26370.log [root@redis_slave1 ~]# tail -100f /usr/local/redis/log/26370/sentinel_26370.log 38894:X 18 Jul 2024 10:36:54.115 # WARNING Memory overcommit must be enabled! Without it, a background save or replication may fail under low memory condition. To fix this issue add 'vm.overcommit_memory = 1' to /etc/sysctl.conf and then reboot or run the command 'sysctl vm.overcommit_memory=1' for this to take effect. 38894:X 18 Jul 2024 10:36:54.116 * oO0OoO0OoO0Oo Redis is starting oO0OoO0OoO0Oo 38894:X 18 Jul 2024 10:36:54.116 * Redis version=7.2.5, bits=64, commit=00000000, modified=0, pid=38894, just started 38894:X 18 Jul 2024 10:36:54.116 * Configuration loaded 38894:X 18 Jul 2024 10:36:54.116 * Increased maximum number of open files to 10032 (it was originally set to 1024). 38894:X 18 Jul 2024 10:36:54.116 * monotonic clock: POSIX clock_gettime 38894:X 18 Jul 2024 10:36:54.116 * Running mode=sentinel, port=26370. 38894:X 18 Jul 2024 10:36:54.116 # WARNING: The TCP backlog setting of 511 cannot be enforced because /proc/sys/net/core/somaxconn is set to the lower value of 128. 38894:X 18 Jul 2024 10:36:54.119 * Sentinel new configuration saved on disk 38894:X 18 Jul 2024 10:36:54.119 * Sentinel ID is 93a83bf4770f1f8eda48b96ead065382ab2b5fe6 38894:X 18 Jul 2024 10:36:54.119 # +monitor master redis_slave1 192.168.1.200 26370 quorum 2 38894:X 18 Jul 2024 10:36:56.164 * +sentinel sentinel 93a83bf4770f1f8eda48b96ead065382ab2b5fe6 192.168.1.200 26370 @ redis_slave1 192.168.1.200 26370 38894:X 18 Jul 2024 10:36:56.170 * Sentinel new configuration saved on disk
3.3 redis_slave2 節(jié)點(diǎn)
# 創(chuàng)建redis數(shù)據(jù)存儲目錄 mkdir -p /usr/local/redis/data/26370 # 創(chuàng)建redis配置文件存儲目錄 mkdir -p /usr/local/redis/conf/26370 # 創(chuàng)建redis日志存儲目錄 mkdir -p /usr/local/redis/log/26370 # 創(chuàng)建redis pid存儲目錄 mkdir /usr/local/redis/run/26370 cat > /usr/local/redis/conf/26370/sentinel.conf <<EOF bind 192.168.1.250 port 26370 dir /usr/local/redis/data/26370 sentinel monitor redis_master 192.168.1.100 16370 2 sentinel down-after-milliseconds redis_master 5000 sentinel auth-pass redis_master dongdong EOF # 啟動 sentinel進(jìn)程 redis-sentinel /usr/local/redis/conf/26370/sentinel.conf &> /usr/local/redis/log/26370/sentinel_26370.log & [root@redis_slave2 ~]# redis-sentinel /usr/local/redis/conf/26370/sentinel.conf &> /usr/local/redis/log/26370/sentinel_26370.log & [1] 39310 # 查看redis-sentinel 進(jìn)程 [root@redis_slave2 ~]# ps -ef | grep redis root 39290 1 0 10:20 ? 00:00:03 redis-server 192.168.1.250:16372 root 39310 39251 0 10:50 pts/0 00:00:00 redis-sentinel 192.168.1.250:26370 [sentinel] root 39316 39251 0 10:51 pts/0 00:00:00 grep --color=auto redis # 查看日志信息 tail -100f /usr/local/redis/log/26370/sentinel_26370.log [root@redis_slave2 ~]# tail -100f /usr/local/redis/log/26370/sentinel_26370.log 39310:X 18 Jul 2024 10:50:57.403 # WARNING Memory overcommit must be enabled! Without it, a background save or replication may fail under low memory condition. To fix this issue add 'vm.overcommit_memory = 1' to /etc/sysctl.conf and then reboot or run the command 'sysctl vm.overcommit_memory=1' for this to take effect. 39310:X 18 Jul 2024 10:50:57.403 * oO0OoO0OoO0Oo Redis is starting oO0OoO0OoO0Oo 39310:X 18 Jul 2024 10:50:57.403 * Redis version=7.2.5, bits=64, commit=00000000, modified=0, pid=39310, just started 39310:X 18 Jul 2024 10:50:57.403 * Configuration loaded 39310:X 18 Jul 2024 10:50:57.404 * Increased maximum number of open files to 10032 (it was originally set to 1024). 39310:X 18 Jul 2024 10:50:57.404 * monotonic clock: POSIX clock_gettime 39310:X 18 Jul 2024 10:50:57.404 * Running mode=sentinel, port=26370. 39310:X 18 Jul 2024 10:50:57.405 # WARNING: The TCP backlog setting of 511 cannot be enforced because /proc/sys/net/core/somaxconn is set to the lower value of 128. 39310:X 18 Jul 2024 10:50:57.407 * Sentinel new configuration saved on disk 39310:X 18 Jul 2024 10:50:57.407 * Sentinel ID is 9227635399651de294a88fa16281232b9cbbccaf 39310:X 18 Jul 2024 10:50:57.407 # +monitor master redis_slave2 192.168.1.250 26370 quorum 2 39310:X 18 Jul 2024 10:50:59.427 * +sentinel sentinel 9227635399651de294a88fa16281232b9cbbccaf 192.168.1.250 26370 @ redis_slave2 192.168.1.250 26370 39310:X 18 Jul 2024 10:50:59.434 * Sentinel new configuration saved on disk
4. 驗(yàn)證sentinel能夠自動切換
4.1 手動停止主庫運(yùn)行,模擬主庫宕機(jī)
# 查看主庫的狀態(tài)
redis-cli -h 192.168.1.100 -p 16370 -a dongdong info replication
Warning: Using a password with '-a' or '-u' option on the command line interface may not be safe.
# Replication
role:master
connected_slaves:2
slave0:ip=192.168.1.200,port=16371,state=online,offset=2898,lag=0
slave1:ip=192.168.1.250,port=16372,state=online,offset=2898,lag=0
master_failover_state:no-failover
master_replid:466362240f16a7414b9236f931c6f97252ee00be
master_replid2:0000000000000000000000000000000000000000
master_repl_offset:2898
second_repl_offset:-1
repl_backlog_active:1
repl_backlog_size:1048576
repl_backlog_first_byte_offset:1
repl_backlog_histlen:2898
# 查看從庫狀態(tài)
redis-cli -h 192.168.1.200 -p 16371 -a dongdong info replication
[root@redis_slave1 ~]# redis-cli -h 192.168.1.200 -p 16371 -a dongdong info replication
Warning: Using a password with '-a' or '-u' option on the command line interface may not be safe.
# Replication
role:slave
master_host:192.168.1.100
master_port:16370
master_link_status:up
master_last_io_seconds_ago:4
master_sync_in_progress:0
slave_read_repl_offset:3066
slave_repl_offset:3066
slave_priority:100
slave_read_only:1
replica_announced:1
connected_slaves:0
master_failover_state:no-failover
master_replid:466362240f16a7414b9236f931c6f97252ee00be
master_replid2:0000000000000000000000000000000000000000
master_repl_offset:3066
second_repl_offset:-1
repl_backlog_active:1
repl_backlog_size:1048576
repl_backlog_first_byte_offset:15
repl_backlog_histlen:3052
redis-cli -h 192.168.1.250 -p 16372 -a dongdong info replication
[root@redis_slave2 ~]# redis-cli -h 192.168.1.250 -p 16372 -a dongdong info replication
Warning: Using a password with '-a' or '-u' option on the command line interface may not be safe.
# Replication
role:slave
master_host:192.168.1.100
master_port:16370
master_link_status:up
master_last_io_seconds_ago:5
master_sync_in_progress:0
slave_read_repl_offset:3094
slave_repl_offset:3094
slave_priority:100
slave_read_only:1
replica_announced:1
connected_slaves:0
master_failover_state:no-failover
master_replid:466362240f16a7414b9236f931c6f97252ee00be
master_replid2:0000000000000000000000000000000000000000
master_repl_offset:3094
second_repl_offset:-1
repl_backlog_active:1
repl_backlog_size:1048576
repl_backlog_first_byte_offset:43
repl_backlog_histlen:3052
# 各節(jié)點(diǎn)運(yùn)行redis進(jìn)程情況
root@redis_master ~]# ss -npl | grep redis
tcp LISTEN 0 128 192.168.1.100:16370 *:* users:(("redis-server",pid=30552,fd=6))
tcp LISTEN 0 128 192.168.1.100:26370 *:* users:(("redis-sentinel",pid=30569,fd=6))
[root@redis_slave1 ~]# ss -npl | grep redis
tcp LISTEN 0 128 192.168.1.200:16371 *:* users:(("redis-server",pid=38879,fd=6))
tcp LISTEN 0 511 192.168.1.200:26370 *:* users:(("redis-sentinel",pid=38908,fd=6))
[root@redis_slave2 ~]# ss -npl | grep redis
tcp LISTEN 0 128 192.168.1.250:16372 *:* users:(("redis-server",pid=39290,fd=6))
tcp LISTEN 0 128 192.168.1.250:26370 *:* users:(("redis-sentinel",pid=39310,fd=6))
# 一切正常
# 手動模擬主庫宕機(jī) 再來查看情況
redis-cli -h 192.168.1.100 -p 16370 -a dongdong shutdown
[root@redis_master ~]# redis-cli -h 192.168.1.100 -p 16370 -a dongdong shutdown
Warning: Using a password with '-a' or '-u' option on the command line interface may not be safe.
# 再次執(zhí)行命令
[root@redis_master ~]# redis-cli -h 192.168.1.100 -p 16370 -a dongdong shutdown
Warning: Using a password with '-a' or '-u' option on the command line interface may not be safe.
Could not connect to Redis at 192.168.1.100:16370: Connection refused # 發(fā)現(xiàn)連接不上 因?yàn)槲覀儼阎鲙戾戳?
4.2 當(dāng)16370 實(shí)例宕機(jī)后,查看集群情況
# 查看集群狀態(tài) redis-cli -h 192.168.1.200 -p 16371 -a dongdong info replication [root@redis_master ~]# redis-cli -h 192.168.1.200 -p 16371 -a dongdong info replication Warning: Using a password with '-a' or '-u' option on the command line interface may not be safe. # Replication role:master # 此時的slave1主機(jī)變成了master 節(jié)點(diǎn) connected_slaves:1 # slave 也變成了1個 slave0:ip=192.168.1.250,port=16372,state=online,offset=14837,lag=1 master_failover_state:no-failover master_replid:c6e88908dbe348999f5406625612e76ee4d6ffd1 master_replid2:b23a9a971ad73ed23fb81186529c671e54acbac4 master_repl_offset:14837 second_repl_offset:14231 repl_backlog_active:1 repl_backlog_size:1048576 repl_backlog_first_byte_offset:1 repl_backlog_histlen:14837 # 查看集群狀態(tài) redis-cli -h 192.168.1.250 -p 16372 -a dongdong info replication [root@redis_slave2 ~]# redis-cli -h 192.168.1.250 -p 16372 -a dongdong info replication Warning: Using a password with '-a' or '-u' option on the command line interface may not be safe. # Replication role:slave master_host:192.168.1.200 master_port:16371 master_link_status:up master_last_io_seconds_ago:1 master_sync_in_progress:0 slave_read_repl_offset:87953 slave_repl_offset:87953 slave_priority:100 slave_read_only:1 replica_announced:1 connected_slaves:0 master_failover_state:no-failover master_replid:c6e88908dbe348999f5406625612e76ee4d6ffd1 master_replid2:b23a9a971ad73ed23fb81186529c671e54acbac4 master_repl_offset:87953 second_repl_offset:14231 repl_backlog_active:1 repl_backlog_size:1048576 repl_backlog_first_byte_offset:1 repl_backlog_histlen:87953 # 現(xiàn)在集群里面的情況就是 一主一從
4.3 恢復(fù)16370 實(shí)例會成為slaves節(jié)點(diǎn)
# 修復(fù)之前，16370端口是連接不上的 redis-cli -h 192.168.1.100 -p 16370 -a dongdong info replication [root@master ~]# redis-cli -h 192.168.1.100 -p 16370 -a dongdong info replication Warning: Using a password with '-a' or '-u' option on the command line interface may not be safe. Could not connect to Redis at 192.168.1.100:16370: Connection refused # 修復(fù)之后，之前宕機(jī)的主庫會自動加入集群并成為slaves節(jié)點(diǎn) # 啟動Redis redis-server /usr/local/redis/conf/16370/redis_16370.conf # 可能需要過了幾秒才會顯示哦 因?yàn)樾畔⑦€沒有同步過來哦 [root@redis_master ~]# redis-cli -h 192.168.1.100 -p 16370 -a dongdong info replication Warning: Using a password with '-a' or '-u' option on the command line interface may not be safe. # Replication role:slave master_host:192.168.1.200 master_port:16371 master_link_status:up master_last_io_seconds_ago:0 master_sync_in_progress:0 slave_read_repl_offset:165354 slave_repl_offset:165354 slave_priority:100 slave_read_only:1 replica_announced:1 connected_slaves:0 master_failover_state:no-failover master_replid:c6e88908dbe348999f5406625612e76ee4d6ffd1 master_replid2:0000000000000000000000000000000000000000 master_repl_offset:165354 second_repl_offset:-1 repl_backlog_active:1 repl_backlog_size:1048576 repl_backlog_first_byte_offset:164602 repl_backlog_histlen:753
5. 手動停掉一個sentinel實(shí)例
5.1 將redis_master主機(jī)中的sentinel實(shí)例給停止掉
# 查看redis進(jìn)程 [root@redis_master ~]# ps -ef | grep redis root 30646 30499 0 11:23 pts/1 00:00:06 redis-sentinel 192.168.1.100:26370 [sentinel] root 30663 1 0 11:42 ? 00:00:01 redis-server 192.168.1.100:16370 root 30672 30499 0 11:50 pts/1 00:00:00 grep --color=auto redis # 殺掉 進(jìn)程id為30646服務(wù) kill -9 30646 # 再次查看redis進(jìn)程 sentinel 已經(jīng)被殺掉了 [root@redis_master ~]# ps -ef | grep redis root 30663 1 0 11:42 ? 00:00:01 redis-server 192.168.1.100:16370 root 30676 30499 0 11:51 pts/1 00:00:00 grep --color=auto redis
5.2 驗(yàn)證主從切換
# 查看各節(jié)點(diǎn)狀況 此時slave1這臺是主庫 # 查詢從庫狀態(tài) redis_master主機(jī) redis-cli -h 192.168.1.100 -p 16370 -a dongdong info replication [root@redis_master ~]# redis-cli -h 192.168.1.100 -p 16370 -a dongdong info replication Warning: Using a password with '-a' or '-u' option on the command line interface may not be safe. # Replication role:slave master_host:192.168.1.200 master_port:16371 master_link_status:up master_last_io_seconds_ago:1 master_sync_in_progress:0 slave_read_repl_offset:251306 slave_repl_offset:251306 slave_priority:100 slave_read_only:1 replica_announced:1 connected_slaves:0 master_failover_state:no-failover master_replid:c6e88908dbe348999f5406625612e76ee4d6ffd1 master_replid2:0000000000000000000000000000000000000000 master_repl_offset:251306 second_repl_offset:-1 repl_backlog_active:1 repl_backlog_size:1048576 repl_backlog_first_byte_offset:164602 repl_backlog_histlen:86705 # 查詢從庫狀態(tài) redis_slave2主機(jī) redis-cli -h 192.168.1.250 -p 16372 -a dongdong info replication [root@redis_master ~]# redis-cli -h 192.168.1.250 -p 16372 -a dongdong info replication Warning: Using a password with '-a' or '-u' option on the command line interface may not be safe. # Replication role:slave master_host:192.168.1.200 master_port:16371 master_link_status:up master_last_io_seconds_ago:0 master_sync_in_progress:0 slave_read_repl_offset:254136 slave_repl_offset:254136 slave_priority:100 slave_read_only:1 replica_announced:1 connected_slaves:0 master_failover_state:no-failover master_replid:c6e88908dbe348999f5406625612e76ee4d6ffd1 master_replid2:b23a9a971ad73ed23fb81186529c671e54acbac4 master_repl_offset:254136 second_repl_offset:14231 repl_backlog_active:1 repl_backlog_size:1048576 repl_backlog_first_byte_offset:1 repl_backlog_histlen:254136 # 查詢主庫狀態(tài) redis-cli -h 192.168.1.200 -p 16371 -a dongdong info replication [root@redis_master ~]# redis-cli -h 192.168.1.200 -p 16371 -a dongdong info replication Warning: Using a password with '-a' or '-u' option on the command line interface may not be safe. # Replication role:master connected_slaves:2 slave0:ip=192.168.1.250,port=16372,state=online,offset=261722,lag=0 slave1:ip=192.168.1.100,port=16370,state=online,offset=261736,lag=0 master_failover_state:no-failover master_replid:c6e88908dbe348999f5406625612e76ee4d6ffd1 master_replid2:b23a9a971ad73ed23fb81186529c671e54acbac4 master_repl_offset:261882 second_repl_offset:14231 repl_backlog_active:1 repl_backlog_size:1048576 repl_backlog_first_byte_offset:1 repl_backlog_histlen:261882 # 此時干掉主庫 redis-cli -h 192.168.1.200 -p 16371 -a dongdong shutdown [root@redis_slave1 ~]# redis-cli -h 192.168.1.200 -p 16371 -a dongdong shutdown Warning: Using a password with '-a' or '-u' option on the command line interface may not be safe. # 再次查詢 [root@redis_slave1 ~]# redis-cli -h 192.168.1.200 -p 16371 -a dongdong info replication Warning: Using a password with '-a' or '-u' option on the command line interface may not be safe. Could not connect to Redis at 192.168.1.200:16371: Connection refused # 此時查詢集群個節(jié)點(diǎn)狀態(tài) 看看哪個節(jié)點(diǎn)成功master 此時集群中只有兩個節(jié)點(diǎn) 分別是redis_master主機(jī) 和 redis_slave2主機(jī) # 先查看redis_master redis-cli -h 192.168.1.100 -p 16370 -a dongdong info replication [root@redis_master ~]# redis-cli -h 192.168.1.100 -p 16370 -a dongdong info replication Warning: Using a password with '-a' or '-u' option on the command line interface may not be safe. # Replication role:slave master_host:192.168.1.250 master_port:16372 master_link_status:up master_last_io_seconds_ago:1 master_sync_in_progress:0 slave_read_repl_offset:56377 slave_repl_offset:56377 slave_priority:100 slave_read_only:1 replica_announced:1 connected_slaves:0 master_failover_state:no-failover master_replid:a375e44a67edd4fb81880032ef927964da6c376f master_replid2:3288009c32135c1d1df6e8d56ceff7d45e572875 master_repl_offset:56377 second_repl_offset:55465 repl_backlog_active:1 repl_backlog_size:1048576 repl_backlog_first_byte_offset:30454 repl_backlog_histlen:25924 # 是正常的
5.3 恢復(fù)Redis 實(shí)例
# 查看各節(jié)點(diǎn) redis 服務(wù)是否停用 如果停用 則重新啟動 ps -ef | grep redis # redis_master redis-server /usr/local/redis/conf/16370/redis_16370.conf # redis_slave1 redis-server /usr/local/redis/conf/16371/redis_16371.conf # redis_slave2 redis-server /usr/local/redis/conf/16372/redis_16372.conf
6. 手動停掉兩個sentinel 實(shí)例
6.1 將redis_slave1 主機(jī)中的sentinel 實(shí)例停止
[root@redis_slave1 ~]# ps -ef | grep redis root 39196 1 0 13:43 ? 00:00:02 redis-server 192.168.1.200:16371 root 39207 38840 0 13:45 pts/0 00:00:03 redis-sentinel 192.168.1.200:26370 [sentinel] root 39220 38840 0 13:57 pts/0 00:00:00 grep --color=auto redis kill -9 39207 root@redis_slave1 ~]# ps -ef | grep redis root 39196 1 0 13:43 ? 00:00:02 redis-server 192.168.1.200:16371 root 39224 38840 0 13:59 pts/0 00:00:00 grep --color=auto redis
6.2 驗(yàn)證主從切換
# 查詢redis_master主機(jī)集群狀態(tài) [root@redis_master ~]# redis-cli -h 192.168.1.100 -p 16370 -a dongdong info replication Warning: Using a password with '-a' or '-u' option on the command line interface may not be safe. # Replication role:slave master_host:192.168.1.200 master_port:16371 master_link_status:up master_last_io_seconds_ago:1 master_sync_in_progress:0 slave_read_repl_offset:140766 slave_repl_offset:140766 slave_priority:100 slave_read_only:1 replica_announced:1 connected_slaves:0 master_failover_state:no-failover master_replid:a375e44a67edd4fb81880032ef927964da6c376f master_replid2:3288009c32135c1d1df6e8d56ceff7d45e572875 master_repl_offset:140766 second_repl_offset:55465 repl_backlog_active:1 repl_backlog_size:1048576 repl_backlog_first_byte_offset:30454 repl_backlog_histlen:110313 # 查詢redis_slvae1主機(jī)集群狀態(tài) [root@redis_slave1 ~]# redis-cli -h 192.168.1.200 -p 16371 -a dongdong info replication Warning: Using a password with '-a' or '-u' option on the command line interface may not be safe. # Replication role:master connected_slaves:2 slave0:ip=192.168.1.100,port=16370,state=online,offset=143290,lag=0 slave1:ip=192.168.1.250,port=16372,state=online,offset=143290,lag=1 master_failover_state:no-failover master_replid:a375e44a67edd4fb81880032ef927964da6c376f master_replid2:3288009c32135c1d1df6e8d56ceff7d45e572875 master_repl_offset:143436 second_repl_offset:55465 repl_backlog_active:1 repl_backlog_size:1048576 repl_backlog_first_byte_offset:15 repl_backlog_histlen:143422 # 查詢redis_slave2主機(jī)集群狀態(tài) [root@redis_slave2 conf]# redis-cli -h 192.168.1.250 -p 16372 -a dongdong info replication Warning: Using a password with '-a' or '-u' option on the command line interface may not be safe. # Replication role:slave master_host:192.168.1.200 master_port:16371 master_link_status:down master_last_io_seconds_ago:-1 master_sync_in_progress:0 slave_read_repl_offset:1 slave_repl_offset:1 master_link_down_since_seconds:-1 slave_priority:100 slave_read_only:1 replica_announced:1 connected_slaves:0 master_failover_state:no-failover master_replid:eb44c8dbaaf01ca3b091a4005ecb563cc467dc76 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 # 停掉redis_slvae1主機(jī)的redis 服務(wù) 看看是否可以進(jìn)行主從切換 [root@redis_slave1 ~]# ps -ef | grep redis root 39196 1 0 13:43 ? 00:00:03 redis-server 192.168.1.200:16371 root 39242 38840 0 14:03 pts/0 00:00:00 grep --color=auto redis kill -9 39196 # 再次查看redis_master主機(jī)集群狀態(tài) [root@redis_master ~]# redis-cli -h 192.168.1.100 -p 16370 -a dongdong info replication Warning: Using a password with '-a' or '-u' option on the command line interface may not be safe. # Replication role:slave master_host:192.168.1.200 master_port:16371 master_link_status:down # master 已經(jīng)是 down 狀態(tài) 說明沒有進(jìn)行主從切換 master_last_io_seconds_ago:-1 master_sync_in_progress:0 slave_read_repl_offset:154925 slave_repl_offset:154925 master_link_down_since_seconds:49 slave_priority:100 slave_read_only:1 replica_announced:1 connected_slaves:0 master_failover_state:no-failover master_replid:a375e44a67edd4fb81880032ef927964da6c376f master_replid2:3288009c32135c1d1df6e8d56ceff7d45e572875 master_repl_offset:154925 second_repl_offset:55465 repl_backlog_active:1 repl_backlog_size:1048576 repl_backlog_first_byte_offset:30454 repl_backlog_histlen:124472 # 再次查詢redis_slvae2主機(jī)集群狀態(tài) [root@redis_slave2 conf]# redis-cli -h 192.168.1.250 -p 16372 -a dongdong info replication Warning: Using a password with '-a' or '-u' option on the command line interface may not be safe. # Replication role:slave master_host:192.168.1.200 master_port:16371 master_link_status:down # master 已經(jīng)是 down 狀態(tài) 說明沒有進(jìn)行主從切換 master_last_io_seconds_ago:-1 master_sync_in_progress:0 slave_read_repl_offset:154925 slave_repl_offset:154925 master_link_down_since_seconds:133 slave_priority:100 slave_read_only:1 replica_announced:1 connected_slaves:0 master_failover_state:no-failover master_replid:a375e44a67edd4fb81880032ef927964da6c376f master_replid2:0000000000000000000000000000000000000000 master_repl_offset:154925 second_repl_offset:-1 repl_backlog_active:1 repl_backlog_size:1048576 repl_backlog_first_byte_offset:149549 repl_backlog_histlen:5377 # 是因?yàn)槲覀冊倥渲梦募兄付?sentinel monitor redis_master 192.168.1.100 16370 2 這個2 表示3個sentinel中最少有2個節(jié)點(diǎn)認(rèn)為master庫宕機(jī)時，才會真正意義上認(rèn)為master庫宕機(jī)。 # 把redis_slvae1主機(jī)的redis 服務(wù)恢復(fù) redis-server /usr/local/redis/conf/16371/redis_16371.conf # 查看狀態(tài) [root@redis_master ~]# redis-cli -h 192.168.1.100 -p 16370 -a dongdong info replication Warning: Using a password with '-a' or '-u' option on the command line interface may not be safe. # Replication role:slave master_host:192.168.1.200 master_port:16371 master_link_status:up # 恢復(fù)redis_slvae1主機(jī)的redis后 master服務(wù)up了 master_last_io_seconds_ago:1 master_sync_in_progress:0 slave_read_repl_offset:3474 slave_repl_offset:3474 slave_priority:100 slave_read_only:1 replica_announced:1 connected_slaves:0 master_failover_state:no-failover master_replid:527f1bfc1c67513d0813d0c0fddb07964640d907 master_replid2:0000000000000000000000000000000000000000 master_repl_offset:3474 second_repl_offset:-1 repl_backlog_active:1 repl_backlog_size:1048576 repl_backlog_first_byte_offset:476 repl_backlog_histlen:2999
到此這篇關(guān)于Redis 哨兵搭建的實(shí)現(xiàn)示例的文章就介紹到這了,更多相關(guān)Redis 哨兵搭建內(nèi)容請搜索腳本之家以前的文章或繼續(xù)瀏覽下面的相關(guān)文章希望大家以后多多支持腳本之家！
相關(guān)文章
Redis是一個開源的使用ANSI C語言編寫、支持網(wǎng)絡(luò)、可基于內(nèi)存亦可持久化的日志型、Key-Value數(shù)據(jù)庫，并提供多種語言的API。從2010年3月15日起，Redis的開發(fā)工作由VMware主持。這篇文章主要介紹的是利用yum安裝Redis的方法，有需要的朋友們可以參考借鑒，下面來一起看看吧2016-11-11
一致性就是數(shù)據(jù)保持一致，在分布式系統(tǒng)中，可以理解為多個節(jié)點(diǎn)中數(shù)據(jù)的值是一致的。本文給大家分享Redis與MySQL雙寫一致性該如何保證，感興趣的朋友一起看看吧2021-06-06
Redis實(shí)現(xiàn)分布式鎖(setnx、getset、incr)以及如何處理超時情況
本文主要介紹了Redis實(shí)現(xiàn)分布式鎖(setnx、getset、incr)以及如何處理超時情況，文中通過示例代碼介紹的非常詳細(xì)，具有一定的參考價值，感興趣的小伙伴們可以參考一下2021-11-11
如何基于Session實(shí)現(xiàn)短信登錄功能
對比起Cookie,Session是存儲在服務(wù)器端的會話,相對安全,并且不像Cookie那樣有存儲長度限制,下面這篇文章主要給大家介紹了關(guān)于如何基于Session實(shí)現(xiàn)短信登錄功能的相關(guān)資料,需要的朋友可以參考下2022-10-10
Python利用redis限制用戶重復(fù)刷新帶來的數(shù)據(jù)問題
在網(wǎng)站開發(fā)中,我們經(jīng)常會遇到需要控制用戶重復(fù)刷新頁面的情況,本文就來介紹了Python利用redis限制用戶重復(fù)刷新帶來的數(shù)據(jù)問題,感興趣的可以了解一下2024-03-03
這篇文章主要給大家介紹了關(guān)于Redis中LRU淘汰策略的相關(guān)資料，文中通過示例代碼介紹的非常詳細(xì)，對大家學(xué)習(xí)或者使用Redis具有一定的參考學(xué)習(xí)價值，需要的朋友們下面來一起學(xué)習(xí)學(xué)習(xí)吧2019-06-06
超強(qiáng)、超詳細(xì)Redis數(shù)據(jù)庫入門教程
這篇文章主要介紹了超強(qiáng)、超詳細(xì)Redis入門教程,本文詳細(xì)介紹了Redis數(shù)據(jù)庫各個方面的知識,需要的朋友可以參考下2014-10-10
詳解Redis高效恢復(fù)策略內(nèi)存快照與AOF
這篇文章主要為大家介紹了Redis高效恢復(fù)策略內(nèi)存快照與AOF及對比詳解,有需要的朋友可以借鑒參考下,希望能夠有所幫助,祝大家多多進(jìn)步,早日升職加薪2023-12-12
這篇文章主要介紹了Redis禁用命令、危險命令及規(guī)避方法,本文介紹了個非常致命的兩個命令以及用配置文件禁用這些命令的方法,需要的朋友可以參考下2015-06-06
今天小編就為大家分享一篇redis禁止幾個危險命令的方法，具有很好的參考價值，希望對大家有所幫助。一起跟隨小編過來看看吧2018-05-05