Redis3.x Sentinel集群配置

redis高可用的三种常见的集群方式：redis sentinel 、redis cluster（多主机+分布式）、redis sharding。本文主机介绍redis sentinel的部署过程。

centos7部署redis的主机ip地址：

master：10.11.11.109

slave：10.11.11.110，10.11.11.111

3.1、Redis3.x Sentinel集群配置

Redis服务启动和配置文件端口的查询

# ansible redmon -i /root/ans/ansible_inventory.txt -m systemd -a "name=redis enabled=yes state=started"

# ansible redmon -i /root/ans/ansible_inventory.txt -m shell -a "rpm -ql redis"

# ansible redmon -i /root/ans/ansible_inventory.txt -m shell -a "ps -ef | grep redis"

# ansible redmon -i /root/ans/ansible_inventory.txt -m shell -a "ss -tunlp | grep redis"

3.1.1、准备Redis配置文件：

# ansible redmon -i /root/ans/ansible_inventory.txt -m fetch -a "src=/etc/redis.conf dest=/root/ans/conf.d flat=yes" --limit=10.11.11.109

# ansible redmon -i /root/ans/ansible_inventory.txt -m fetch -a "src=/etc/redis-sentinel.conf dest=/root/ans/conf.d flat=yes" --limit=10.11.11.109

3.1.2、三台机器主从节点的准备：

1）、master节点：10.11.11.109

修改后，包含默认的情况如下：

# sed -e "s/#.*//g" redis.conf | awk '{if (length !=0) print $0}'

bind 10.11.11.109

protected-mode no

port 6379

tcp-backlog 511

timeout 0

tcp-keepalive 300

daemonize yes

supervised no

pidfile /var/run/redis/redis.pid

loglevel notice

logfile /var/log/redis/redis.log

databases 16

save 900 1

save 300 10

save 60 10000

stop-writes-on-bgsave-error yes

rdbcompression yes

rdbchecksum yes

dbfilename dump.rdb

dir /var/lib/redis

slave-serve-stale-data yes

slave-read-only yes

repl-diskless-sync no

repl-diskless-sync-delay 5

repl-disable-tcp-nodelay no

slave-priority 100

appendonly yes

appendfilename "appendonly.aof"

appendfsync everysec

no-appendfsync-on-rewrite no

auto-aof-rewrite-percentage 100

auto-aof-rewrite-min-size 64mb

aof-load-truncated yes

lua-time-limit 5000

slowlog-log-slower-than 10000

slowlog-max-len 128

latency-monitor-threshold 0

notify-keyspace-events ""

hash-max-ziplist-entries 512

hash-max-ziplist-value 64

list-max-ziplist-size -2

list-compress-depth 0

set-max-intset-entries 512

zset-max-ziplist-entries 128

zset-max-ziplist-value 64

hll-sparse-max-bytes 3000

activerehashing yes

client-output-buffer-limit normal 0 0 0

client-output-buffer-limit slave 256mb 64mb 60

client-output-buffer-limit pubsub 32mb 8mb 60

hz 10

aof-rewrite-incremental-fsync yes

下面是可以或需要修改的选项：

仅仅修改这四项，为的是方便管理，也可以保持默认配置

daemonize    yes    使Redis以守护进程模式运行

pidfile    /var/run/redis_端口号.pid    设置Redis的PID文件位置

port    端口号6379    设置Redis监听的端口号

dir     自己定义目录    设置持久文件存放位置

logfile /var/log/redis/redis.log

dir /var/lib/redis

 

详细情况：

bind 10.11.11.109

protected-mode no

默认情况下，Redis node和sentinel的protected-mode都是yes，在搭建集群时，若想从远程连接redis集群，需要将redis.conf和sentinel.conf的protected-mode修改为no，若只修改redis node，从远程连接sentinel后，依然是无法正常使用的，且sentinel的配置文件中没有protected-mode配置项，需要手工添加。

protected-mode在默认开启的情况下要是配置里没有指定bind和密码。开启该参数后，redis只会本地进行访问，拒绝外部访问

 

##启用增量（Master禁用）

appendonly yes

appendfsync everysec

2）、slave节点：10.11.11.110-111

修改后，包含默认的情况如下：

# sed -e "s/#.*//g" redis.conf.201-202 | awk '{if (length !=0) print $0}'

bind 10.11.11.110-111

protected-mode no

port 6379

tcp-backlog 511

timeout 0

tcp-keepalive 300

daemonize yes

supervised no

pidfile /var/run/redis/redis.pid

loglevel notice

logfile /var/log/redis/redis.log

databases 16

save 900 1

save 300 10

save 60 10000

stop-writes-on-bgsave-error yes

rdbcompression yes

rdbchecksum yes

dbfilename dump.rdb

dir /var/lib/redis

slaveof 10.11.7.205 6379

slave-serve-stale-data yes

slave-read-only yes

repl-diskless-sync no

repl-diskless-sync-delay 5

repl-disable-tcp-nodelay no

slave-priority 100

appendonly yes

appendfilename "appendonly.aof"

appendfsync everysec

no-appendfsync-on-rewrite no

auto-aof-rewrite-percentage 100

auto-aof-rewrite-min-size 64mb

aof-load-truncated yes

lua-time-limit 5000

slowlog-log-slower-than 10000

slowlog-max-len 128

latency-monitor-threshold 0

notify-keyspace-events ""

hash-max-ziplist-entries 512

hash-max-ziplist-value 64

list-max-ziplist-size -2

list-compress-depth 0

set-max-intset-entries 512

zset-max-ziplist-entries 128

zset-max-ziplist-value 64

hll-sparse-max-bytes 3000

activerehashing yes

client-output-buffer-limit normal 0 0 0

client-output-buffer-limit slave 256mb 64mb 60

client-output-buffer-limit pubsub 32mb 8mb 60

hz 10

aof-rewrite-incremental-fsync yes

详细说明：

下面是可以或需要修改的选项：

仅仅修改这四项，为的是方便管理，也可以保持默认配置

daemonize    yes    使Redis以守护进程模式运行

pidfile    /var/run/redis_端口号.pid    设置Redis的PID文件位置

port    端口号6379    设置Redis监听的端口号

dir     自己定义目录    设置持久文件存放位置

protected-mode no

默认情况下，Redis node和sentinel的protected-mode都是yes，在搭建集群时，若想从远程连接redis集群，需要将redis.conf和sentinel.conf的protected-mode修改为no，若只修改redis node，从远程连接sentinel后，依然是无法正常使用的，且sentinel的配置文件中没有protected-mode配置项，需要手工添加。

protected-mode在默认开启的情况下要是配置里没有指定bind和密码。开启该参数后，redis只会本地进行访问，拒绝外部访问

logfile "/var/log/redis/redis.log"

 

bind 10.11.11.110-111

##启用增量（Master禁用）

appendonly yes

appendfsync everysec

 

slave-priority 80                ---->110

slave-priority 50                --->111

添加为从节点：

# slaveof

slaveof 192.168.0.247 6379

copy到指定的主从节点：

# ls

redis.conf  redis.conf.201  redis.conf.202  redis-sentinel.conf 

# ansible redmon -i /root/ans/ansible_inventory.txt -m copy -a "src=/root/ans/conf.d/redis.conf dest=/etc/redis.conf backup=yes" --limit=10.11.11.109

# ansible redmon -i /root/ans/ansible_inventory.txt -m copy -a "src=/root/ans/conf.d/redis.conf.201 dest=/etc/redis.conf backup=yes" --limit=10.11.11.110

# ansible redmon -i /root/ans/ansible_inventory.txt -m copy -a "src=/root/ans/conf.d/redis.conf.202 dest=/etc/redis.conf backup=yes" --limit=10.11.11.111

3）、启动Redis主从集群

先启动主节点（master）：

# ansible redmon -i /root/ans/ansible_inventory.txt -m systemd -a "name=redis enabled=yes state=restarted" --limit=10.11.11.109

再启动从节点：

# ansible redmon -i /root/ans/ansible_inventory.txt -m systemd -a "name=redis enabled=yes state=restarted" --limit=10.11.11.110,10.11.11.111

4）、防火墙放行

#  ansible redmon -i /root/ans/ansible_inventory.txt -m firewalld -a "zone=public state=enabled permanent=yes port=6379/tcp"

#  ansible redmon -i /root/ans/ansible_inventory.txt -m firewalld -a "zone=public state=enabled permanent=true port=6379/udp"

注意使其生效喔

#  ansible redmon -i /root/ans/ansible_inventory.txt -m shell -a "firewall-cmd --reload"

5）、查看redis主从状态

# ansible redmon -i /root/ans/ansible_inventory.txt -m shell -a "redis-cli -h 10.11.11.109 info Replication" --limit=10.11.11.109

10.11.11.109 | SUCCESS | rc=0 >>

# Replication

role:master

connected_slaves:2

slave0:ip=10.11.11.111,port=6379,state=online,offset=57,lag=1

slave1:ip=10.11.11.110,port=6379,state=online,offset=57,lag=1

master_repl_offset:57

repl_backlog_active:1

repl_backlog_size:1048576

repl_backlog_first_byte_offset:2

repl_backlog_histlen:56

 

# ansible redmon -i /root/ans/ansible_inventory.txt -m shell -a "redis-cli -h 10.11.11.111 info Replication" --limit=10.11.11.111

10.11.11.111 | SUCCESS | rc=0 >>

# Replication

role:slave

master_host:10.11.11.109

master_port:6379

master_link_status:up

master_last_io_seconds_ago:2

master_sync_in_progress:0

slave_repl_offset:85

slave_priority:50

slave_read_only:1

connected_slaves:0

master_repl_offset:0

repl_backlog_active:0

repl_backlog_size:1048576

repl_backlog_first_byte_offset:0

repl_backlog_histlen:0

 

# ansible redmon -i /root/ans/ansible_inventory.txt -m shell -a "redis-cli -h 10.11.11.110 info Replication" --limit=10.11.11.110

10.11.11.110 | SUCCESS | rc=0 >>

# Replication

role:slave

master_host:10.11.11.109

master_port:6379

master_link_status:up

master_last_io_seconds_ago:6

master_sync_in_progress:0

slave_repl_offset:99

slave_priority:80

slave_read_only:1

connected_slaves:0

master_repl_offset:0

repl_backlog_active:0

repl_backlog_size:1048576

repl_backlog_first_byte_offset:0

repl_backlog_histlen:0

3.1.3、配置sentinel集群

1）、主备主机的sentinel.conf配置如下：

master节点：

# sed -e "s/#.*//g" redis-sentinel.conf | awk '{if (length !=0) print $0}'

port 26379

dir /data/var/tmp

sentinel monitor mymaster 10.11.11.109 6379 2

sentinel down-after-milliseconds mymaster 30000

sentinel parallel-syncs mymaster 1

sentinel failover-timeout mymaster 180000

logfile /data/var/log/redis/sentinel.log

bind 10.11.11.109

protected-mode no

slave节点：

# sed -e "s/#.*//g" redis-sentinel.conf.110 | awk '{if (length !=0) print $0}'

port 26379

dir /data/var/tmp

sentinel monitor mymaster 10.11.11.109 6379 2

sentinel down-after-milliseconds mymaster 30000

sentinel parallel-syncs mymaster 1

sentinel failover-timeout mymaster 180000

logfile /data/var/log/redis/sentinel.log

bind 10.11.11.110

protected-mode no

# sed -e "s/#.*//g" redis-sentinel.conf.backup111 | awk '{if (length !=0) print $0}'

port 26379

dir /data/var/tmp

sentinel monitor mymaster 10.11.11.109 6379 2

sentinel down-after-milliseconds mymaster 30000

sentinel parallel-syncs mymaster 1

sentinel failover-timeout mymaster 180000

logfile /data/var/log/redis/sentinel.log

bind 10.11.11.111

protected-mode no

copy到指定的主从节点：

# ls

 redis-sentinel.conf.110  redis-sentinel.conf.backup111  redis-sentinel.conf.master

 

# ansible redmon -i /root/ans/ansible_inventory.txt -m copy -a "src=/root/ans/conf.d/redis-sentinel.conf.master dest=/etc/redis-sentinel.conf backup=yes" --limit=10.11.11.109

# ansible redmon -i /root/ans/ansible_inventory.txt -m copy -a "src=/root/ans/conf.d/redis-sentinel.conf.110 dest=/etc/redis-sentinel.conf backup=yes" --limit=10.11.11.110

# ansible redmon -i /root/ans/ansible_inventory.txt -m copy -a "src=/root/ans/conf.d/redis-sentinel.conf.backup111 dest=/etc/redis-sentinel.conf backup=yes" --limit=10.11.11.111

2）、防火墙配置

# ansible redmon -i /root/ans/ansible_inventory.txt -m firewalld -a "zone=public state=enabled permanent=yes port=26379/tcp"

# ansible redmon -i /root/ans/ansible_inventory.txt -m firewalld -a "zone=public state=enabled permanent=yes port=26379/udp"

# ansible redmon -i /root/ans/ansible_inventory.txt -m shell -a "firewall-cmd --reload"

3）、启动sentinel集群

注意点：

1)：首次启动时，必须先启动Master

2)：Sentinel 只在 server 端做主从切换，app端要自己开发(例如Jedis库的SentinelJedis，能够监控Sentinel的状态)

3)：若Master已经被判定为下线，Sentinel已经选择了新的Master，也已经将old Master改成Slave，但是还没有将其改成new Master。若此时重启old Master，则Redis集群将处于无Master状态，此时只能手动修改配置文件，然后重新启动集群

# ansible redmon -i /root/ans/ansible_inventory.txt -m systemd -a "name=redis-sentinel state=started enabled=yes" --limit=10.11.11.109

# ansible redmon -i /root/ans/ansible_inventory.txt -m systemd -a "name=redis-sentinel state=started enabled=yes" --limit=10.11.11.110

# ansible redmon -i /root/ans/ansible_inventory.txt -m systemd -a "name=redis-sentinel state=started enabled=yes" --limit=10.11.11.111

查询状态：

主从节点查看的结果是一样的：

# ansible redmon -i /root/ans/ansible_inventory.txt -m shell -a "redis-cli -h 10.11.11.109 -p 26379 info Sentinel"

# ansible redmon -i /root/ans/ansible_inventory.txt -m shell -a "redis-cli -h 10.11.11.110 -p 26379 info Sentinel"

# ansible redmon -i /root/ans/ansible_inventory.txt -m shell -a "redis-cli -h 10.11.11.110 -p 26379 info Sentinel"

10.11.11.111 | SUCCESS | rc=0 >>

# 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=10.11.11.109:6379,slaves=2,sentinels=3

 

10.11.11.109 | SUCCESS | rc=0 >>

# 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=10.11.11.109:6379,slaves=2,sentinels=3

 

10.11.11.110 | SUCCESS | rc=0 >>

# 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=10.11.11.109:6379,slaves=2,sentinels=3

4）、宕机演示

添加测试数据：

[root@redis247 ~]# redis-cli -h 192.168.0.8 -p 6379

192.168.0.9:6379> set lll 245

OK

192.168.0.8:6379> save

OK

 

[root@redis247 ~]# redis-cli -h 192.168.0.247 -p 26379 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=redismaster,status=ok,address=192.168.0.247:6379,slaves=2,sentinels=3

关掉redis-server服务：

[root@redis247 ~]# redis-cli -h 192.168.0.247 -p 6379 shutdown

[root@redis247 ~]# ss -tunlp | grep redis

tcp    LISTEN     0      128        192.168.0.247:26379                 *:*      users:(("redis-sentinel",1241,4))

状态切换的结果：

[root@redis247 ~]# redis-cli -h 192.168.0.247 -p 26379 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=redismaster,status=ok,address=192.168.0.9:6379,slaves=2,sentinels=3

[root@redis247 ~]# redis-cli -h 192.168.0.9 -p 26381 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=redismaster,status=ok,address=192.168.0.9:6379,slaves=2,sentinels=3

再关闭sentinel：（相当于主机宕机）

[root@redis247 ~]# ss -tunlp | grep redis

tcp    LISTEN     0      128        192.168.0.247:26379                 *:*      users:(("redis-sentinel",1241,4))

[root@redis247 ~]# ps -aux | grep redis

Warning: bad syntax, perhaps a bogus '-'? See /usr/share/doc/procps-3.2.8/FAQ

root      1241  0.5  0.4 133532  2412 ?        Ssl  14:26   0:07 redis-sentinel 192.168.0.247:26379 [sentinel]

root      1272  0.0  0.1 103252   840 pts/0    S+   14:49   0:00 grep redis

[root@redis247 ~]# kill -9 1241

[root@redis247 ~]# ps -aux | grep redis

Warning: bad syntax, perhaps a bogus '-'? See /usr/share/doc/procps-3.2.8/FAQ

root      1274  0.0  0.1 103252   840 pts/0    S+   14:49   0:00 grep redis

[root@redis247 ~]# ss -tunlp | grep redis

状态同上

 

重新启动：

[root@redis247 ~]# /etc/init.d/redis restart

[root@redis247 ~]# redis-sentinel /usr/local/redis/sentinel.conf

[root@redis247 ~]# redis-cli -h 192.168.0.8 -p 26380 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=redismaster,status=ok,address=192.168.0.9:6379,slaves=2,sentinels=3

#注意：原来的主宕机重新启动后，充当从的角色

# 如果不是主宕机，而是从宕机，那么不会发生切换行为，只会把宕机的那台从集群中剔除。

# 已宕机的机器，如果再次加入集群，只要它成为了当前主的从机，则Sentinel会自动发现，并将其加入集群成员。

 

再次查看测试数据：

[root@redis247 ~]# redis-cli -h 192.168.0.9 -p 6379

192.168.0.9:6379> get lll

"245"