如今業(yè)務(wù)系統(tǒng)對于緩存redis的依賴似乎是必不可少的，我們可以在各種各樣的系統(tǒng)中看到Redis的身影。考慮到系統(tǒng)運(yùn)行的穩(wěn)定性，Redis的應(yīng)用和MySQL數(shù)據(jù)庫一樣需要做到高可用部署。
一、Redis 的多種高可用方案
常見的Redis的高可用方案有以下幾種：
- Redis Replication（主從復(fù)制）：Redis的主從復(fù)制可以實(shí)現(xiàn)數(shù)據(jù)的備份和讀寫分離。通過配置主節(jié)點(diǎn)和從節(jié)點(diǎn)，主節(jié)點(diǎn)將數(shù)據(jù)異步復(fù)制到從節(jié)點(diǎn)上。當(dāng)主節(jié)點(diǎn)發(fā)生故障時(shí)，一個(gè)從節(jié)點(diǎn)可以被提升為新的主節(jié)點(diǎn)，實(shí)現(xiàn)故障轉(zhuǎn)移。主從復(fù)制適用于對讀操作較多、對可用性要求較高的場景。
- Redis Sentinel（哨兵模式）：哨兵模式是Redis官方推薦的實(shí)現(xiàn)高可用的方案之一。通過運(yùn)行一個(gè)或多個(gè)Sentinel進(jìn)程，監(jiān)控Redis主節(jié)點(diǎn)的狀態(tài)。當(dāng)主節(jié)點(diǎn)故障時(shí)，Sentinel會(huì)自動(dòng)進(jìn)行故障轉(zhuǎn)移，將其中一個(gè)從節(jié)點(diǎn)提升為新的主節(jié)點(diǎn)。哨兵還可以監(jiān)控從節(jié)點(diǎn)并進(jìn)行故障恢復(fù)。哨兵模式適用于對高可用性要求不是特別高的場景。
- Redis Cluster（集群模式）：Redis Cluster是Redis官方提供的高可用和分布式解決方案。通過將多個(gè)Redis實(shí)例組成一個(gè)集群，Redis Cluster提供了自動(dòng)的數(shù)據(jù)分片和高可用性。數(shù)據(jù)被分配到不同的節(jié)點(diǎn)上，并使用Gossip協(xié)議進(jìn)行節(jié)點(diǎn)之間的通信。當(dāng)有節(jié)點(diǎn)發(fā)生故障時(shí)，Redis Cluster可以自動(dòng)將數(shù)據(jù)遷移到其他正常的節(jié)點(diǎn)上。Redis Cluster適用于對可用性和擴(kuò)展性要求較高的場景。集群模式只能存儲在db0。
- 第三方中間件/解決方案：除了Redis官方提供的高可用方案，還有一些第三方中間件或解決方案可以用于實(shí)現(xiàn)Redis的高可用，如Codis、Twemproxy等。這些中間件提供了更多的功能和擴(kuò)展性，如代理、負(fù)載均衡、故障恢復(fù)等。
在選擇高可用方案時(shí)，需要考慮系統(tǒng)的可用性需求、數(shù)據(jù)一致性要求、網(wǎng)絡(luò)拓?fù)涞纫蛩亍Ｍ瑫r(shí)，也要注意進(jìn)行適當(dāng)?shù)臏y試和監(jiān)控，確保Redis集群的穩(wěn)定性和高可用性。
二、使用Docker Compose安裝Redis并配置哨兵模式（Redis Sentinel）
1. 環(huán)境準(zhǔn)備
集群的架構(gòu)一般服務(wù)器為奇數(shù)臺，所以，如果是采用集群模式，那么至少準(zhǔn)備3臺linux服務(wù)器，受生產(chǎn)環(huán)境所限，我們只有兩臺Linux服務(wù)器，但是我們可以使用Docker搭建多個(gè)Redis服務(wù)(Redis主服務(wù)1、Redis從服務(wù)2、Redis從服務(wù)3)：
- 192.168.0.210 (Redis主服務(wù)1、Redis從服務(wù)2)
- 192.168.0.195 (Redis從服務(wù)3)
2. 準(zhǔn)備Redis文件存放目錄
- 準(zhǔn)備Redis存儲目錄，在兩臺主從服務(wù)器上分別執(zhí)行一下命令
在192.168.0.210服務(wù)器上執(zhí)行Redis主服務(wù)1所需目錄及權(quán)限命令
mkdir -p /opt/contAIner/redis/master/data /opt/container/redis/master/conf /opt/container/redis/master/logs /opt/container/redis/sentinel/data /opt/container/redis/sentinel/conf /opt/container/redis/sentinel/logs
chmod -R 777 /opt/container/redis/master/data /opt/container/redis/master/conf /opt/container/redis/master/logs /opt/container/redis/sentinel/data /opt/container/redis/sentinel/conf /opt/container/redis/sentinel/logs
在192.168.0.210服務(wù)器上執(zhí)行Redis從服務(wù)2所需目錄及權(quán)限命令
mkdir -p /opt/container/redis/slave1/data /opt/container/redis/slave1/conf /opt/container/redis/slave1/logs /opt/container/redis/sentinel1/data /opt/container/redis/sentinel1/conf /opt/container/redis/sentinel1/logs
chmod -R 777 /opt/container/redis/slave1/data /opt/container/redis/slave1/conf /opt/container/redis/slave1/logs /opt/container/redis/sentinel1/data /opt/container/redis/sentinel1/conf /opt/container/redis/sentinel1/logs
在192.168.0.195服務(wù)器上執(zhí)行Redis從服務(wù)3所需目錄及權(quán)限命令
mkdir -p /opt/container/redis/slave2/data /opt/container/redis/slave2/conf /opt/container/redis/slave2/logs /opt/container/redis/sentinel2/data /opt/container/redis/sentinel2/conf /opt/container/redis/sentinel2/logs
chmod -R 777 /opt/container/redis/slave2/data /opt/container/redis/slave2/conf /opt/container/redis/slave2/logs /opt/container/redis/sentinel2/data /opt/container/redis/sentinel2/conf /opt/container/redis/sentinel2/logs
/opt/container/redis/ ** /data 用于存放Redis數(shù)據(jù)文件
/opt/container/redis/ ** /conf 用于存放Redis配置文件
/opt/container/redis/ ** /logs 用于存放Redis日志文件
/opt/container/sentinel/ ** /data 用于存放Redis哨兵數(shù)據(jù)文件
/opt/container/sentinel/ ** /conf 用于存放Redis哨兵配置文件
/opt/container/sentinel/ ** /logs 用于存放Redis哨兵日志文件
3. Redis服務(wù)器docker-compose.yml文件
- 192.168.0.210服務(wù)器上部署兩個(gè)Redis服務(wù)(Redis主服務(wù)1、Redis從服務(wù)2)的docker-compose-redis.yml文件
version: '3'
services:
##redis主配置
redisMaster:
image: redis:latest
restart: always
container_name: redis-master
command: redis-server /usr/local/etc/redis/redis.conf
##將26381映射到26381上
ports:
- "26381:26381"
volumes:
##數(shù)據(jù)目錄，要確保先創(chuàng)建好
- /opt/container/redis/master/data:/data
- /opt/container/redis/master/conf/redis.conf:/usr/local/etc/redis/redis.conf
- /opt/container/redis/master/logs:/logs
- "/etc/localtime:/etc/localtime"
- "/usr/share/zoneinfo/Asia/Shanghai:/etc/timezone"
##redis從配置
redisSlave1:
image: redis:latest
restart: always
container_name: redis-slave-1
command: redis-server /usr/local/etc/redis/redis.conf
##將26382映射到26382上
ports:
- "26382:26382"
volumes:
##數(shù)據(jù)目錄，要確保先創(chuàng)建好
- /opt/container/redis/slave1/data:/data
- /opt/container/redis/slave1/conf/redis.conf:/usr/local/etc/redis/redis.conf
- /opt/container/redis/slave1/logs:/logs
- "/etc/localtime:/etc/localtime"
- "/usr/share/zoneinfo/Asia/Shanghai:/etc/timezone"
sentinel:
image: redis:latest
restart: always
container_name: redis-sentinel
ports:
- 36381:36381
command: redis-sentinel /opt/redis/sentinel/sentinel.conf
volumes:
- /opt/container/redis/sentinel/data:/data
- /opt/container/redis/sentinel/conf/sentinel.conf:/opt/redis/sentinel/sentinel.conf
- /opt/container/redis/sentinel/logs:/logs
- "/etc/localtime:/etc/localtime"
- "/usr/share/zoneinfo/Asia/Shanghai:/etc/timezone"
sentinel1:
image: redis:latest
restart: always
container_name: redis-sentinel-1
ports:
- 36382:36382
command: redis-sentinel /opt/redis/sentinel/sentinel1.conf
volumes:
- /opt/container/redis/sentinel1/data:/data
- /opt/container/redis/sentinel1/conf/sentinel1.conf:/opt/redis/sentinel/sentinel1.conf
- /opt/container/redis/sentinel1/logs:/logs
- "/etc/localtime:/etc/localtime"
- "/usr/share/zoneinfo/Asia/Shanghai:/etc/timezone"
- 192.168.0.195服務(wù)器上部署Redis服務(wù)(Redis從服務(wù)3)的docker-compose-redis-slave.yml文件
version: '3'
services:
##redis從2配置
redisSlave2:
image: redis:latest
restart: always
container_name: redis-slave-2
command: redis-server /usr/local/etc/redis/redis.conf
##將26383映射到26383上
ports:
- "26383:26383"
volumes:
##數(shù)據(jù)目錄，要確保先創(chuàng)建好
- /opt/container/redis/slave2/data:/data
- /opt/container/redis/slave2/conf/redis.conf:/usr/local/etc/redis/redis.conf
- /opt/container/redis/slave2/logs:/logs
- "/etc/localtime:/etc/localtime"
- "/usr/share/zoneinfo/Asia/Shanghai:/etc/timezone"
sentinel2:
image: redis:latest
restart: always
container_name: redis-sentinel-2
ports:
- 36383:36383
command: redis-sentinel /opt/redis/sentinel/sentinel2.conf
volumes:
- /opt/container/redis/sentinel2/data:/data
- /opt/container/redis/sentinel2/conf/sentinel2.conf:/opt/redis/sentinel/sentinel2.conf
- /opt/container/redis/sentinel2/logs:/logs
- "/etc/localtime:/etc/localtime"
- "/usr/share/zoneinfo/Asia/Shanghai:/etc/timezone"
4. 準(zhǔn)備Redis服務(wù)的配置文件
- 192.168.0.210主Redis配置文件，配置端口26381，Redis密碼，并將配置文件redis.conf上傳至/opt/container/redis/master/conf目錄
Appendonly yes
port 26381
appendfilename appendonly.aof
appendfsync everysec
auto-aof-rewrite-min-size 10M
auto-aof-rewrite-percentage 100
requirepass "設(shè)置密碼"
masterauth "設(shè)置密碼"
replica-read-only no
- 192.168.0.210從Redis配置文件，配置端口26382，密碼，并將配置文件redis.conf上傳至/opt/container/redis/slave1/conf目錄
appendonly yes
port 26382
appendfilename appendonly.aof
appendfsync everysec
auto-aof-rewrite-min-size 10M
auto-aof-rewrite-percentage 100
requirepass "設(shè)置密碼"
replicaof 192.168.0.210 26381
masterauth "設(shè)置密碼"
replica-read-only no
- 192.168.0.195從Redis配置文件，配置端口26383，密碼，并將配置文件redis.conf上傳至/opt/container/redis/slave2/conf目錄
appendonly yes
port 26383
appendfilename appendonly.aof
appendfsync everysec
auto-aof-rewrite-min-size 10M
auto-aof-rewrite-percentage 100
requirepass "設(shè)置密碼"
replicaof 192.168.0.210 26381
masterauth "設(shè)置密碼"
replica-read-only no
4. 準(zhǔn)備Redis哨兵的配置文件，每個(gè)Redis服務(wù)對應(yīng)一個(gè)哨兵配置
- 192.168.0.210主Redis哨兵配置文件，配置哨兵sentinel實(shí)例運(yùn)行的端口36381，Redis密碼，并將配置文件sentinel.conf上傳至/opt/container/redis/sentinel/conf目錄
# 哨兵sentinel實(shí)例運(yùn)行的端口
port 36381
daemonize no
pidfile /var/run/redis-sentinel.pid
dir /tmp
sentinel monitor mymaster 192.168.0.210 26381 2
sentinel auth-pass mymaster "Redis密碼"
# 指定多少毫秒之后 主節(jié)點(diǎn)沒有應(yīng)答哨兵sentinel 此時(shí) 哨兵主觀上認(rèn)為主節(jié)點(diǎn)下線 默認(rèn)30秒
sentinel down-after-milliseconds mymaster 30000
# 指定了在發(fā)生failover主備切換時(shí)最多可以有多少個(gè)slave同時(shí)對新的master進(jìn)行同步,這個(gè)數(shù)字越小，完成failover所需的時(shí)間就越長
sentinel parallel-syncs mymaster 1
# 故障轉(zhuǎn)移的超時(shí)時(shí)間
sentinel failover-timeout mymaster 180000
sentinel deny-scripts-reconfig yes
- 192.168.0.210從Redis哨兵配置文件，配置哨兵sentinel1實(shí)例運(yùn)行的端口36382，Redis密碼，并將配置文件sentinel1.conf上傳至/opt/container/redis/sentinel1/conf目錄
# 哨兵sentinel實(shí)例運(yùn)行的端口
port 36382
daemonize no
pidfile /var/run/redis-sentinel1.pid
dir /tmp
sentinel monitor mymaster 192.168.0.210 26381 2
sentinel auth-pass mymaster "Redis密碼"
# 指定多少毫秒之后 主節(jié)點(diǎn)沒有應(yīng)答哨兵sentinel 此時(shí) 哨兵主觀上認(rèn)為主節(jié)點(diǎn)下線 默認(rèn)30秒
sentinel down-after-milliseconds mymaster 30000
# 指定了在發(fā)生failover主備切換時(shí)最多可以有多少個(gè)slave同時(shí)對新的master進(jìn)行同步,這個(gè)數(shù)字越小，完成failover所需的時(shí)間就越長
sentinel parallel-syncs mymaster 1
# 故障轉(zhuǎn)移的超時(shí)時(shí)間
sentinel failover-timeout mymaster 180000
sentinel deny-scripts-reconfig yes
- 192.168.0.195從Redis哨兵配置文件，配置哨兵sentinel2實(shí)例運(yùn)行的端口36383，Redis密碼，并將配置文件sentinel2.conf上傳至/opt/container/redis/sentinel2/conf目錄
appendonly yes
port 26383
appendfilename appendonly.aof
appendfsync everysec
auto-aof-rewrite-min-size 10M
auto-aof-rewrite-percentage 100
requirepass "設(shè)置密碼"
replicaof 192.168.0.210 26381
masterauth "設(shè)置密碼"
replica-read-only no
5. 在兩臺服務(wù)器上分別執(zhí)行docker-compose安裝啟動(dòng)命令
將docker-compose-redis.yml上傳至/opt/software目錄，這個(gè)目錄可以自己選擇，然后到目錄下執(zhí)行安裝啟動(dòng)命令
docker-compose -f docker-compose-redis.yml up -d
[root@localhost software]# docker-compose -f docker-compose-redis.yml up -d
[+] Running 10/10
? sentinel1 Pulled 15.7s
? redisSlave1 Pulled 15.7s
? sentinel Pulled 3.4s
? a2abf6c4d29d Already exists 0.0s
? c7a4e4382001 Pull complete 0.4s
? 4044b9ba67c9 Pull complete 1.0s
? c8388a79482f Pull complete 2.3s
? 413c8bb60be2 Pull complete 2.3s
? 1abfd3011519 Pull complete 2.4s
? redisMaster Pulled 15.7s
WARN[0015] Found orphan containers ([mysql Nginx]) for this project. If you removed or renamed this service in your compose file, you can run this command with the --remove-orphans flag to clean it up.
[+] Running 4/4
? Container redis-sentinel-1 Started 0.5s
? Container redis-slave-1 Started 0.6s
? Container redis-master Started 0.6s
? Container redis-sentinel Started 0.5s
通過docker ps命令可以看到redis和redis哨兵已經(jīng)安裝并啟動(dòng)成功
[root@localhost software]# docker ps
CONTAINER ID IMAGE COMMAND CREATED STATUS PORTS NAMES
f609322cabaa redis:latest "docker-entrypoint.s…" 59 seconds ago Up 58 seconds 6379/tcp, 0.0.0.0:26381->26381/tcp, :::26381->26381/tcp redis-master
18b75828b5b7 redis:latest "docker-entrypoint.s…" 59 seconds ago Up 58 seconds 6379/tcp, 0.0.0.0:36381->36381/tcp, :::36381->36381/tcp redis-sentinel
f0f9a037c7ae redis:latest "docker-entrypoint.s…" 59 seconds ago Up 58 seconds 6379/tcp, 0.0.0.0:26382->26382/tcp, :::26382->26382/tcp redis-slave-1
e51d3b0bc696 redis:latest "docker-entrypoint.s…" 59 seconds ago Up 58 seconds 6379/tcp, 0.0.0.0:36382->36382/tcp, :::36382->36382/tcp redis-sentinel-1
6. 通過命令查看redis哨兵狀態(tài)
- 進(jìn)入docker容器
docker exec -it f609322cabaa bash
- 進(jìn)入Redis目錄
cd /usr/local/bin
- 運(yùn)行info Replication命令查看主節(jié)點(diǎn)信息，我們可以看到connected_slaves有2個(gè)，其中slave0的ip為172.18.0.1，這是因?yàn)槠浍@取的是docker的ip地址
./redis-cli -h 127.0.0.1 -p 26381 -a "密碼" info Replication
# Replication
role:master
connected_slaves:2
slave0:ip=172.18.0.1,port=26382,state=online,offset=700123,lag=0
slave1:ip=192.168.0.195,port=26383,state=online,offset=700123,lag=0
master_failover_state:no-failover
master_replid:9ee56f68d25b71158544f6cfafc677822c401ec3
master_replid2:0000000000000000000000000000000000000000
master_repl_offset:700123
second_repl_offset:-1
repl_backlog_active:1
repl_backlog_size:1048576
repl_backlog_first_byte_offset:1
repl_backlog_histlen:700123
- 同樣的操作方式進(jìn)入到Sentinel容器，通過INFO Sentinel命令查看哨兵信息，我們可以看到有兩個(gè)redis從服務(wù)，三個(gè)哨兵
root@fba6d91e10f6:/usr/local/bin# ./redis-cli -h 127.0.0.1 -p 36381 INFO Sentinel
# Sentinel
sentinel_masters:1
sentinel_tilt:0
sentinel_running_scripts:0
sentinel_scripts_queue_length:0
sentinel_simulate_failure_flags:0
master0:name=mymaster,status=ok,address=192.168.0.210:26381,slaves=2,sentinels=3
7. 測試redis哨兵主備切換
- 手動(dòng)關(guān)閉主Redis服務(wù)
docker stop 5541698b65a1
- 進(jìn)入到Sentinel容器，通過INFO Sentinel命令查看哨兵信息，可以看到主Redis服務(wù)地址已經(jīng)切換到192.168.0.195:26383
root@fba6d91e10f6:/usr/local/bin# ./redis-cli -h 127.0.0.1 -p 36381 INFO Sentinel
# Sentinel
sentinel_masters:1
sentinel_tilt:0
sentinel_running_scripts:0
sentinel_scripts_queue_length:0
sentinel_simulate_failure_flags:0
master0:name=mymaster,status=ok,address=192.168.0.195:26383,slaves=3,sentinels=3
- 進(jìn)入到192.168.0.195的Redis容器，info Replication命令查看節(jié)點(diǎn)信息，我們可以看到此Redis節(jié)點(diǎn)已變?yōu)橹鞣?wù)，有一個(gè)從服務(wù)slave0:ip=192.168.0.210,port=26382
./redis-cli -h 127.0.0.1 -p 26383 -a "密碼" 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=192.168.0.210,port=26382,state=online,offset=894509,lag=0
master_failover_state:no-failover
master_replid:e9ebebdff0a5f7b7622c6c5fbfed7b1e44d84ae2
master_replid2:9ee56f68d25b71158544f6cfafc677822c401ec3
master_repl_offset:894509
second_repl_offset:884279
repl_backlog_active:1
repl_backlog_size:1048576
repl_backlog_first_byte_offset:846
repl_backlog_histlen:893664
三、配置并測試Redis集群模式（Redis Cluster）
Redis哨兵模式其實(shí)質(zhì)還是主從復(fù)制，只不過加了哨兵進(jìn)行自動(dòng)主備切換，Redis集群模式（Redis Cluster）才是真正意義上的集群部署，它可以將數(shù)據(jù)進(jìn)行分布式分片存儲，但是其只能存儲到db0。
1. 環(huán)境準(zhǔn)備
我們?nèi)匀皇褂蒙厦娴腖inux服務(wù)器環(huán)境進(jìn)行安裝redis集群：在192.168.0.210上安裝redisCluster1、redisCluster2、redisCluster3三臺Reids服務(wù)；在在192.168.0.195上安裝redisCluster4、redisCluster5、redisCluster6三臺Reids服務(wù)。
2. 準(zhǔn)備Redis文件存放目錄
- 準(zhǔn)備Redis存儲目錄，在兩臺主從服務(wù)器上分別執(zhí)行一下命令
在192.168.0.210服務(wù)器上執(zhí)行Redis集群所需目錄及權(quán)限命令
mkdir -p /opt/container/redis/cluster1/data /opt/container/redis/cluster1/conf /opt/container/redis/cluster1/logs /opt/container/redis/cluster2/data /opt/container/redis/cluster2/conf /opt/container/redis/cluster2/logs /opt/container/redis/cluster3/data /opt/container/redis/cluster3/conf /opt/container/redis/cluster3/logs
chmod -R 777 /opt/container/redis/cluster1/data /opt/container/redis/cluster1/conf /opt/container/redis/cluster1/logs /opt/container/redis/cluster2/data /opt/container/redis/cluster2/conf /opt/container/redis/cluster2/logs /opt/container/redis/cluster3/data /opt/container/redis/cluster3/conf /opt/container/redis/cluster3/logs
在192.168.0.195服務(wù)器上執(zhí)行Redis集群所需目錄及權(quán)限命令
mkdir -p /opt/container/redis/cluster4/data /opt/container/redis/cluster4/conf /opt/container/redis/cluster4/logs /opt/container/redis/cluster5/data /opt/container/redis/cluster5/conf /opt/container/redis/cluster5/logs /opt/container/redis/cluster6/data /opt/container/redis/cluster6/conf /opt/container/redis/cluster6/logs
chmod -R 777 /opt/container/redis/cluster4/data /opt/container/redis/cluster4/conf /opt/container/redis/cluster4/logs /opt/container/redis/cluster5/data /opt/container/redis/cluster5/conf /opt/container/redis/cluster5/logs /opt/container/redis/cluster6/data /opt/container/redis/cluster6/conf /opt/container/redis/cluster6/logs
- 192.168.0.210 /192.168.0.195Redis配置文件、端口（12381、12382、12383、12384、12385、12386）、密碼，并將配置文件redis.conf上傳至對應(yīng)目錄，只需要端口不同
port 12381
cluster-enabled yes #啟動(dòng)集群模式
cluster-config-file nodes-1.conf
cluster-node-timeout 5000
cluster-announce-ip 192.168.0.210
cluster-announce-port 12381
cluster-announce-bus-port 22381
bind 0.0.0.0
protected-mode no
appendonly yes
#如果要設(shè)置密碼需要增加如下配置：
#(設(shè)置redis訪問密碼)
requirepass 密碼
#(設(shè)置集群節(jié)點(diǎn)間訪問密碼，跟上面一致)
masterauth 密碼
3. Redis服務(wù)器docker-compose.yml文件
- 192.168.0.210服務(wù)器上部署兩個(gè)Redis服務(wù)的docker-compose-redis-cluster.yml文件
version: '3'
services:
##redis節(jié)點(diǎn)1配置
redisCluster1:
image: redis:latest
restart: always
.NETwork_mode: "host"
container_name: redis-cluster1
command: redis-server /usr/local/etc/redis/redis.conf
volumes:
##數(shù)據(jù)目錄，要確保先創(chuàng)建好
- /opt/container/redis/cluster1/data:/data
- /opt/container/redis/cluster1/conf/redis.conf:/usr/local/etc/redis/redis.conf
- /opt/container/redis/cluster1/logs:/logs
- "/etc/localtime:/etc/localtime"
- "/usr/share/zoneinfo/Asia/Shanghai:/etc/timezone"
##redis節(jié)點(diǎn)2配置
redisCluster2:
image: redis:latest
restart: always
network_mode: "host"
container_name: redis-cluster2
command: redis-server /usr/local/etc/redis/redis.conf
volumes:
##數(shù)據(jù)目錄，要確保先創(chuàng)建好
- /opt/container/redis/cluster2/data:/data
- /opt/container/redis/cluster2/conf/redis.conf:/usr/local/etc/redis/redis.conf
- /opt/container/redis/cluster2/logs:/logs
- "/etc/localtime:/etc/localtime"
- "/usr/share/zoneinfo/Asia/Shanghai:/etc/timezone"
##redis節(jié)點(diǎn)3配置
redisCluster3:
image: redis:latest
restart: always
network_mode: "host"
container_name: redis-cluster3
command: redis-server /usr/local/etc/redis/redis.conf
volumes:
##數(shù)據(jù)目錄，要確保先創(chuàng)建好
- /opt/container/redis/cluster3/data:/data
- /opt/container/redis/cluster3/conf/redis.conf:/usr/local/etc/redis/redis.conf
- /opt/container/redis/cluster3/logs:/logs
- "/etc/localtime:/etc/localtime"
- "/usr/share/zoneinfo/Asia/Shanghai:/etc/timezone"
- 192.168.0.195服務(wù)器上部署三個(gè)Redis服務(wù)的docker-compose-redis-cluster.yml文件
version: '3'
services:
##redis節(jié)點(diǎn)4配置
redisCluster4:
image: redis:latest
restart: always
network_mode: "host"
container_name: redis-cluster4
command: redis-server /usr/local/etc/redis/redis.conf
volumes:
##數(shù)據(jù)目錄，要確保先創(chuàng)建好
- /opt/container/redis/cluster4/data:/data
- /opt/container/redis/cluster4/conf/redis.conf:/usr/local/etc/redis/redis.conf
- /opt/container/redis/cluster4/logs:/logs
- "/etc/localtime:/etc/localtime"
- "/usr/share/zoneinfo/Asia/Shanghai:/etc/timezone"
##redis節(jié)點(diǎn)5配置
redisCluster5:
image: redis:latest
restart: always
network_mode: "host"
container_name: redis-cluster5
command: redis-server /usr/local/etc/redis/redis.conf
volumes:
##數(shù)據(jù)目錄，要確保先創(chuàng)建好
- /opt/container/redis/cluster5/data:/data
- /opt/container/redis/cluster5/conf/redis.conf:/usr/local/etc/redis/redis.conf
- /opt/container/redis/cluster5/logs:/logs
- "/etc/localtime:/etc/localtime"
- "/usr/share/zoneinfo/Asia/Shanghai:/etc/timezone"
##redis節(jié)點(diǎn)6配置
redisCluster6:
image: redis:latest
restart: always
network_mode: "host"
container_name: redis-cluster6
command: redis-server /usr/local/etc/redis/redis.conf
volumes:
##數(shù)據(jù)目錄，要確保先創(chuàng)建好
- /opt/container/redis/cluster6/data:/data
- /opt/container/redis/cluster6/conf/redis.conf:/usr/local/etc/redis/redis.conf
- /opt/container/redis/cluster6/logs:/logs
- "/etc/localtime:/etc/localtime"
- "/usr/share/zoneinfo/Asia/Shanghai:/etc/timezone"
4. 在兩臺服務(wù)器上分別執(zhí)行docker-compose安裝啟動(dòng)命令
將
docker-compose-redis-cluster.yml上傳至/opt/software目錄，這個(gè)目錄可以自己選擇，然后到目錄下執(zhí)行安裝啟動(dòng)命令
docker-compose -f docker-compose-redis-cluster.yml up -d
[+] Running 3/3
? Container redis-cluster3 Started 0.5s
? Container redis-cluster2 Started 0.4s
? Container redis-cluster1 Started 0.5s
通過docker ps命令可以看到redis和redis哨兵已經(jīng)安裝并啟動(dòng)成功
[root@localhost software]# docker ps
CONTAINER ID IMAGE COMMAND CREATED STATUS PORTS NAMES
67a48962160d redis:latest "docker-entrypoint.s…" 10 seconds ago Up 9 seconds 6379/tcp, 0.0.0.0:52381->52381/tcp, :::52381->52381/tcp redis-cluster1
b10f669691b3 redis:latest "docker-entrypoint.s…" 10 seconds ago Up 9 seconds 6379/tcp, 0.0.0.0:52383->52383/tcp, :::52383->52383/tcp redis-cluster3
d3899c9c01f6 redis:latest "docker-entrypoint.s…" 10 seconds ago Up 9 seconds 6379/tcp, 0.0.0.0:52382->52382/tcp, :::52382->52382/tcp redis-cluster2
5. 在兩臺服務(wù)器上分別執(zhí)行docker-compose安裝啟動(dòng)命令
- 登錄到docker容器中
docker exec -it 67a48962160d bash
- 用redis-cli創(chuàng)建整個(gè)redis集群
cd /usr/local/bin
root@localhost:/usr/local/bin# ./redis-cli --cluster create 192.168.0.210:12381 192.168.0.210:12382 192.168.0.210:12383 192.168.0.195:12384 192.168.0.195:12385 192.168.0.195:12386 --cluster-replicas 1 -a "密碼"
Warning: Using a password with '-a' or '-u' option on the command line interface may not be safe.
>>> Performing hash slots allocation on 6 nodes...
Master[0] -> Slots 0 - 5460
Master[1] -> Slots 5461 - 10922
Master[2] -> Slots 10923 - 16383
Adding replica 192.168.0.195:12386 to 192.168.0.210:12381
Adding replica 192.168.0.210:12383 to 192.168.0.195:12384
Adding replica 192.168.0.195:12385 to 192.168.0.210:12382
M: 061d7b1bf2f93df7cbf261e47a7981800d636e63 192.168.0.210:12381
slots:[0-5460] (5461 slots) master
M: 6cfbf9677ab802483ddc7cbb715fe770c8de884a 192.168.0.210:12382
slots:[10923-16383] (5461 slots) master
S: 5afc2e7d2da8f9d7f7ad6e99d5ad04ffbf5bdfe5 192.168.0.210:12383
replicates d54562615048044b43e368db71789829d76fa263
M: d54562615048044b43e368db71789829d76fa263 192.168.0.195:12384
slots:[5461-10922] (5462 slots) master
S: 9137f05da40ce173a975fa4a5e86e65b9d3fe4e3 192.168.0.195:12385
replicates 6cfbf9677ab802483ddc7cbb715fe770c8de884a
S: de04b0b6d207bd8653f2cb738cb47443c427810e 192.168.0.195:12386
replicates 061d7b1bf2f93df7cbf261e47a7981800d636e63
Can I set the above configuration? (type 'yes' to accept): yes
>>> Nodes configuration updated
>>> Assign a different config epoch to each node
>>> Sending CLUSTER MEET messages to join the cluster
Waiting for the cluster to join
>>> Performing Cluster Check (using node 192.168.0.210:12381)
M: 061d7b1bf2f93df7cbf261e47a7981800d636e63 192.168.0.210:12381
slots:[0-5460] (5461 slots) master
1 additional replica(s)
S: 5afc2e7d2da8f9d7f7ad6e99d5ad04ffbf5bdfe5 192.168.0.210:12383
slots: (0 slots) slave
replicates d54562615048044b43e368db71789829d76fa263
S: 9137f05da40ce173a975fa4a5e86e65b9d3fe4e3 192.168.0.195:12385
slots: (0 slots) slave
replicates 6cfbf9677ab802483ddc7cbb715fe770c8de884a
M: d54562615048044b43e368db71789829d76fa263 192.168.0.195:12384
slots:[5461-10922] (5462 slots) master
1 additional replica(s)
S: de04b0b6d207bd8653f2cb738cb47443c427810e 192.168.0.195:12386
slots: (0 slots) slave
replicates 061d7b1bf2f93df7cbf261e47a7981800d636e63
M: 6cfbf9677ab802483ddc7cbb715fe770c8de884a 192.168.0.210:12382
slots:[10923-16383] (5461 slots) master
1 additional replica(s)
[OK] All nodes agree about slots configuration.
>>> Check for open slots...
>>> Check slots coverage...
[OK] All 16384 slots covered.
- 登錄redis，并查看集群信息
# 查看集群信息
cluster info
# 查看節(jié)點(diǎn)列表
cluster nodes
[root@localhost software]# docker exec -it 407e3847371a bash
root@localhost:/data# cd /usr/local/bin
root@localhost:/usr/local/bin# redis-cli -c -h 127.0.0.1 -p 12383 -a "密碼"
127.0.0.1:12383> cluster info
cluster_state:ok
cluster_slots_assigned:16384
cluster_slots_ok:16384
cluster_slots_pfail:0
cluster_slots_fail:0
cluster_known_nodes:6
cluster_size:3
cluster_current_epoch:6
cluster_my_epoch:4
cluster_stats_messages_ping_sent:173
cluster_stats_messages_pong_sent:163
cluster_stats_messages_meet_sent:1
cluster_stats_messages_sent:337
cluster_stats_messages_ping_received:163
cluster_stats_messages_pong_received:174
cluster_stats_messages_received:337
127.0.0.1:12383> cluster nodes
061d7b1bf2f93df7cbf261e47a7981800d636e63 192.168.0.210:12381@22381 master - 0 1696952366541 1 connected 0-5460
de04b0b6d207bd8653f2cb738cb47443c427810e 192.168.0.195:12386@22386 slave 061d7b1bf2f93df7cbf261e47a7981800d636e63 0 1696952366541 1 connected
5afc2e7d2da8f9d7f7ad6e99d5ad04ffbf5bdfe5 192.168.0.210:12383@22383 slave d54562615048044b43e368db71789829d76fa263 0 1696952366039 4 connected
9137f05da40ce173a975fa4a5e86e65b9d3fe4e3 192.168.0.195:12385@22385 slave 6cfbf9677ab802483ddc7cbb715fe770c8de884a 0 1696952367043 2 connected
6cfbf9677ab802483ddc7cbb715fe770c8de884a 192.168.0.210:12382@22382 master - 0 1696952365537 2 connected 10923-16383
d54562615048044b43e368db71789829d76fa263 192.168.0.195:12383@22384 myself,master - 0 1696952366000 4 connected 5461-10922
搭建Redis集群模式一定要注意，docker-compose配置一定要將network_mode設(shè)置為"host"模式，且不需要端口映射。如果創(chuàng)建集群時(shí)，一直顯示等待連接，那么需要配置防火墻，放開Redis集群端口。
