基于Sentinel实现redis主从自动切换
Sentinel(哨兵)是用于监控redis集群中Master状态的工具,它可以实现对redis的监控、通知、自动故障转移。
Sentinel作用:
- Master状态检测
- 当被监控的某个 Redis Master异常无法连接时 Sentinel 可以向系统管理员发送通知, 也可以通过 API 向其他程序发送通知,并且进行Master-Slave切换,将其中一个Slave作为Master,将之前的Master作为Slave
- Master-Slave切换后,master_redis.conf、slave_redis.conf和sentinel.conf的内容都会发生改变,即master_redis.conf中会多一行slaveof的配置,sentinel.conf的监控目标会随之调换。当应用程序连接Redis 服务器时, Redis Sentinel会告之新的主服务器地址和端口。
Redis 主从部署:
192.168.11.12:6379 master
192.168.11.12:6381 slave
192.168.11.13:6379 slave
192.168.11.13:6381 slave
192.168.11.14:6379 slave
192.168.11.14:6381 slave
Sentinel:
192.168.11.12
192.168.11.13
192.168.11.14
具体部署(简写)
mkdir -p /opt/redis6379/{bin,etc,var,log}
mkdir -p /opt/redis6381/{bin,etc,var,log}
tar zxmf redis-2.8.19.tar.gz
cd redis-2.8.19
make
make PREFIX=/opt/redis6379 install
make PREFIX=/opt/redis6381 install
master config:
cat redis6379.conf
daemonize yes
pidfile /opt/redis6379/redis6379.pid
port 6379
tcp-backlog 10240
maxclients 10000
bind 0.0.0.0
timeout 0
tcp-keepalive 0
loglevel notice
logfile /opt/redis6379/log/redis6379.log
databases 16
stop-writes-on-bgsave-error yes
rdbcompression yes
rdbchecksum yes
dbfilename dump6379.rdb
dir /export/redis6379/
slave-priority 100
maxmemory 2500mb
maxmemory-policy allkeys-lru
appendonly no
appendfilename "appendonly01.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-entries 512
list-max-ziplist-value 64
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
slave config:
cat redis6379.conf
daemonize yes
pidfile /opt/redis6379/redis6379.pid
port 6379
tcp-backlog 10240
maxclients 10000
bind 0.0.0.0
timeout 0
tcp-keepalive 0
loglevel notice
logfile /opt/redis6379/log/redis6379.log
databases 16
stop-writes-on-bgsave-error yes
rdbcompression yes
rdbchecksum yes
dbfilename dump6379.rdb
dir /export/redis6379/
slave-priority 100
slave-serve-stale-data yes
slave-read-only yes
slave-priority 100
slaveof 192.168.11.12 6379
maxmemory 2500mb
maxmemory-policy allkeys-lru
appendonly no
appendfilename "appendonly01.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 ""
save 900 1
save 300 100
save 60 10000
hash-max-ziplist-entries 512
hash-max-ziplist-value 64
list-max-ziplist-entries 512
list-max-ziplist-value 64
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
启动:
/opt/redis6379/bin/redis-server /opt/redis6379/etc/redis6379.conf
/opt/redis6381/bin/redis-server /opt/redis6381/etc/redis6381.conf
# ps aux | grep redis
root 6379 0.2 0.7 137368 7520 ? Ssl 08:19 0:00 /opt/redis6379/bin/redis-server 0.0.0.0:6379
root 6457 0.1 0.7 137368 7504 ? Ssl 08:19 0:00 /opt/redis6381/bin/redis-server 0.0.0.0:6381
查看主从状态:
# /opt/redis6379/bin/redis-cli -h 127.0.0.1 -p 6379 INFO Replication
# Replication
role:master
connected_slaves:5
slave0:ip=192.168.11.12,port=6381,state=online,offset=449,lag=1
slave2:ip=192.168.11.13,port=6379,state=online,offset=449,lag=1
slave4:ip=192.168.11.13,port=6381,state=online,offset=449,lag=1
slave2:ip=192.168.11.14,port=6379,state=online,offset=449,lag=1
slave4:ip=192.168.11.14,port=6381,state=online,offset=449,lag=1
master_repl_offset:449
repl_backlog_active:1
repl_backlog_size:1048576
repl_backlog_first_byte_offset:2
repl_backlog_histlen:448
查看同步:
192.168.11.12:6379
127.0.0.1:6379> set aa 123
OK
192.168.11.13:6379
127.0.0.1:6379> get aa
"123"
192.168.11.14:6381
127.0.0.1:6381> get aa
"123"
数据以同步
配置sentinel:
mkdir -p /opt/redis6379/sentinel
sentinel.conf:
port 26379
dir "/opt/redis6379/sentinel"
sentinel monitor master 192.168.11.12 6379 2
sentinel down-after-milliseconds master 60000
sentinel config-epoch master 1
sentinel leader-epoch master 1
sentinel known-slave master 192.168.11.12 6381
sentinel known-slave master 192.168.11.14 6381
# Generated by CONFIG REWRITE
sentinel known-slave master 192.168.11.112 6379
sentinel known-slave master 192.168.11.112 6381
sentinel known-slave master 192.168.11.14 6379
sentinel known-sentinel master 192.168.11.112 26379 459b949dd141a301f93d764d92bb04af9450870b
sentinel known-sentinel master 192.168.11.14 26379 9638d9c07c83c8cf6b3cac0b419867d9a4eeb17c
sentinel current-epoch 1
说明:
port 监听端口
dir Sentinel服务运行时使用的临时文件夹
sentinel monitor master IP 端口 x 监控的 redis master ip 端口 判断为失效至少需要2个 Sentinel进程的同意,只要同意Sentinel的数量不达标,自动failover就不会执行
一下配置启动之后无法看到
sentinel failover-timeout master 180000 : 如果在该时间(ms)内未能完成failover操作,则认为该failover失败
sentinel parallel-syncs master 1:指定了在执行故障转移时,最多可以有多少个从Redis实例在同步新的主实例,在从Redis实例较多的情况下这个数字越小,同步的时间越长,完成故障转移所需的时间就越长
启动 redis-sentinel :
/opt/redis6379/bin/redis-sentinel /opt/redis6379/sentinel/sentinel.conf 2>&1 /opt/redis6379/sentinel.log &
模拟故障:
192.168.11.12:6379
/opt/redis6379/bin/redis-cli -h 127.0.0.1 -p 6379 shutdown
查看:
192.168.11.12:6381
/opt/redis6379/bin/redis-cli -h 127.0.0.1 -p 6381 info Replication
# Replication
role:slave
master_host:192.168.11.14
master_port:6381
master_link_status:up
master_last_io_seconds_ago:0
master_sync_in_progress:0
slave_repl_offset:4265
slave_priority:100
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
master 已经主观转移