Redis Sentinel集群双机房容灾实施步骤

概要目标
防止双机房情况下任一个机房完全无法提供服务时如何让Redis继续提供服务。
架构设计
A、B两机房，其中A机房有一Master一Slave和两个Sentinel，B机房只有2个Sentinel，如下图。

初始规划
A机房
192.168.71.213 S+哨兵
192.168.71.214 M+哨兵
B机房
192.168.70.214 S
192.168.70.215 S

目录创建
--redis软件目录
mkdir -p /home/redis
--pidfile文件存放目录
mkdir -p /home/redis/redisrun/
解压redis截止到 /home/redis

集群配置
【Master】
选择71.214作为Master
[root@node-71 redis]# vi /home/redis/redis.conf
#后台启动
daemonize yes
pidfile "/home/redis/redisrun/redis_6379.pid"
port 6379
timeout 0
tcp-keepalive 0
loglevel notice
logfile "/home/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 "/home/redis/redisdb"
#如果做故障切换，不论主从节点都要填写密码且要保持一致
masterauth "123456"
slave-serve-stale-data yes
slave-read-only yes
repl-disable-tcp-nodelay no
slave-priority 98
#当前redis密码
requirepass "123456"
appendonly yes
# appendfsync always
appendfsync everysec
# appendfsync no
no-appendfsync-on-rewrite no
auto-aof-rewrite-percentage 100
auto-aof-rewrite-min-size 64mb
lua-time-limit 5000
slowlog-log-slower-than 10000
slowlog-max-len 128
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
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
# Generated by CONFIG REWRITE

【Slave】
选择其余3个几点作为Slave
[root@node-71 redis]# vi /home/redis/redis.conf
daemonize yes
pidfile "/home/redis/redisrun/redis_6379.pid"
port 6379
timeout 0
tcp-keepalive 0
loglevel notice
logfile "/home/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 "/home/redis/redisdb"
#主节点密码
masterauth "123456"
slave-serve-stale-data yes
slave-read-only yes
repl-disable-tcp-nodelay no
slave-priority 98
requirepass "123456"
appendonly yes
# appendfsync always
appendfsync everysec
# appendfsync no
no-appendfsync-on-rewrite no
auto-aof-rewrite-percentage 100
auto-aof-rewrite-min-size 64mb
lua-time-limit 5000
slowlog-log-slower-than 10000
slowlog-max-len 128
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
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
# Generated by CONFIG REWRITE
#配置主节点信息
slaveof 192.168.71.214 6379

--检查修正
daemonize yes
pidfile "/home/redis/redisrun//redis_6379.pid"
logfile "/home/redis/redis.log"

【sentinel.conf】
选择A机房2节点作为sentinel
vi /home/redis/sentinel.conf
port 26379
#1表示在sentinel集群中只要有两个节点检测到redis主节点出故障就进行切换，单sentinel节点无效(自己测试发现的)
#如果3s内mymaster无响应，则认为mymaster宕机了
#如果10秒后,mysater仍没活过来，则启动failover
sentinel monitor mymaster 192.168.71.214 6379 1
sentinel down-after-milliseconds mymaster 3000
sentinel failover-timeout mymaster 10000
daemonize yes
#指定工作目录
dir "/home/redis/sentinel-work"
protected-mode no
logfile "/home/redis/sentinellog/sentinel.log"
#redis主节点密码
sentinel auth-pass mymaster 123456
# Generated by CONFIG REWRITE

--检查修正
sentinel monitor mymaster 192.168.71.214 6379 1
dir "/home/redis/sentinel-work"
logfile "/home/redis/sentinellog/sentinel.log"

启动检查
【启动集群与日志监控】
每个几点都执行
cd /home/redis/src/
./redis-server /home/redis/redis.conf

tail -f /home/redis/redis.log

只在sentinel节点执行
cd /home/redis/src/
./redis-sentinel /home/redis/sentinel.conf

tail -f /home/redis/sentinellog/sentinel.log

【Master检查】
cd /home/redis/src/
[root@localhost src]# ./redis-cli -h 192.168.70.214 -p 6379 -a 123456
192.168.70.214:6379> info Replication
# Replication
role:master
connected_slaves:3
slave0:ip=192.168.71.213,port=6379,state=online,offset=1107595,lag=1
slave1:ip=192.168.70.214,port=6379,state=online,offset=1107742,lag=0
slave2:ip=192.168.70.215,port=6379,state=online,offset=1107889,lag=0
master_repl_offset:1107889
repl_backlog_active:1
repl_backlog_size:1048576
repl_backlog_first_byte_offset:59314
repl_backlog_histlen:1048576
192.168.70.214:6379> set test zgy
OK
192.168.70.214:6379> get test
"zgy"
192.168.70.214:6379>

【Slave检查,只读】
192.168.71.214:6379> get test
"zgy"
192.168.71.214:6379> set test zgy2
(error) READONLY You can't write against a read only slave.
192.168.71.214:6379> info Replication
# Replication
role:slave
master_host:192.168.70.214
master_port:6379
master_link_status:up
master_last_io_seconds_ago:1
master_sync_in_progress:0
slave_repl_offset:42385
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
192.168.71.214:6379>

断网断电测试
断网
通过开启防火墙来模拟
service iptables status
--service iptables start
--70网段2节点的防火墙配置
[root@localhost redis]# cat /etc/sysconfig/iptables
# Firewall configuration written by system-config-firewall
# Manual customization of this file is not recommended.
*filter
:INPUT ACCEPT [0:0]
:FORWARD ACCEPT [0:0]
:OUTPUT ACCEPT [0:0]
#屏蔽A机房2个节点
-I INPUT -s 192.168.71.213 -j DROP
-I INPUT -s 192.168.71.214 -j DROP
-A INPUT -m state --state ESTABLISHED,RELATED -j ACCEPT
-A INPUT -p icmp -j ACCEPT
-A INPUT -i lo -j ACCEPT
-A INPUT -m state --state NEW -m tcp -p tcp --dport 22 -j ACCEPT
-A INPUT -j REJECT --reject-with icmp-host-prohibited
-A FORWARD -j REJECT --reject-with icmp-host-prohibited
COMMIT

断网
B机房断网前
--前
192.168.71.214:6379> info Replication
# Replication
role:master
connected_slaves:3
slave0:ip=192.168.71.213,port=6379,state=online,offset=12825868,lag=1
slave1:ip=192.168.70.214,port=6379,state=online,offset=12825868,lag=1
slave2:ip=192.168.70.215,port=6379,state=online,offset=12826015,lag=0
master_repl_offset:12826162
repl_backlog_active:1
repl_backlog_size:1048576
repl_backlog_first_byte_offset:11777587
repl_backlog_histlen:1048576
192.168.71.214:6379>

--后
--明显找不到70网段的那2个节点啦
192.168.71.214:6379> info Replication
# Replication
role:master
connected_slaves:1
slave0:ip=192.168.71.213,port=6379,state=online,offset=12909588,lag=1
master_repl_offset:12909588
repl_backlog_active:1
repl_backlog_size:1048576
repl_backlog_first_byte_offset:11861013
repl_backlog_histlen:1048576
192.168.71.214:6379>

而Master还能继续对外提供服务

A机房断网前、后
前
192.168.71.214:6379> info Replication
# Replication
role:master
connected_slaves:3
slave0:ip=192.168.71.213,port=6379,state=online,offset=12942691,lag=1
slave1:ip=192.168.70.214,port=6379,state=online,offset=12942691,lag=1
slave2:ip=192.168.70.215,port=6379,state=online,offset=12942838,lag=0
master_repl_offset:12942838
repl_backlog_active:1
repl_backlog_size:1048576
repl_backlog_first_byte_offset:11894263
repl_backlog_histlen:1048576

后，出现2个Master？？
192.168.71.214:6379> info Replication
# Replication
role:master
connected_slaves:0
master_repl_offset:12957363
repl_backlog_active:1
repl_backlog_size:1048576
repl_backlog_first_byte_offset:11908788
repl_backlog_histlen:1048576
192.168.71.214:6379>

192.168.71.213:6379> info replication
# Replication
role:master
connected_slaves:0
master_repl_offset:12943881
repl_backlog_active:0
repl_backlog_size:1048576
repl_backlog_first_byte_offset:0
repl_backlog_histlen:0
192.168.71.213:6379>

断电
通过kill redis进程来模拟
ps -ef|grep redis
断电前
192.168.71.213:6379> info replication
# Replication
role:master
connected_slaves:3
slave0:ip=192.168.70.215,port=6379,state=online,offset=13091227,lag=0
slave1:ip=192.168.70.214,port=6379,state=online,offset=13091227,lag=0
slave2:ip=192.168.71.214,port=6379,state=online,offset=13091080,lag=1
master_repl_offset:13091227
repl_backlog_active:1
repl_backlog_size:1048576
repl_backlog_first_byte_offset:13087442
repl_backlog_histlen:3786

192.168.71.214:6379> info Replication
# Replication
role:master
connected_slaves:1
slave0:ip=192.168.71.213,port=6379,state=online,offset=13096642,lag=1
master_repl_offset:13096642
repl_backlog_active:1
repl_backlog_size:1048576
repl_backlog_first_byte_offset:13092272
repl_backlog_histlen:4371
192.168.71.214:6379>

断电后
192.168.70.214:6379> info Replication
# Replication
role:slave
master_host:192.168.71.214
master_port:6379
master_link_status:down
master_last_io_seconds_ago:-1
master_sync_in_progress:0
slave_repl_offset:13159324
master_link_down_since_seconds:18
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

92.168.70.215:6379> info Replication
# Replication
role:slave
master_host:192.168.71.214
master_port:6379
master_link_status:down
master_last_io_seconds_ago:-1
master_sync_in_progress:0
slave_repl_offset:13159324
master_link_down_since_seconds:28
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

70网段都变成Slave无法正常提供服务了。。。

此时，需要修改其中一个节点的配置来向外提供服务
先Kill掉redis进程，再修改某一节点的redis参数，指向其中一个节点，如70.215，并检查另外一台，删除这一项，最后重启2个节点，对外正常提供服务
vi /home/redis/redis.conf
slaveof 192.168.70.214 6379

[root@localhost src]# ./redis-cli -h 192.168.70.214 -p 6379 -a 123456
192.168.70.214:6379> info Replication
# Replication
role:master
connected_slaves:1
slave0:ip=192.168.70.215,port=6379,state=online,offset=15,lag=1
master_repl_offset:15
repl_backlog_active:1
repl_backlog_size:1048576
repl_backlog_first_byte_offset:2
repl_backlog_histlen:14
192.168.70.214:6379>

【还原初始】
修改71.214 之外的参数
vi /home/redis/redis.conf
slaveof 192.168.71.214 6379

vi /home/redis/sentinel.conf
sentinel monitor mymaster 192.168.71.214 6379 1
并删除最后几行

数据校验
Master执行更新数据会同步Slave
注意事项
见每步后面