数据存储我们在应用设计过程中非常重要的一部分，无论是关系型数据库，还是Redis、MongoDB等非关系型数据库，都有很多的高可用方案，还有一些针对不同业务设计的中间件，使其性能更有特色，更能保证数据存储的稳定和安全。
目前主流的Redis的数据存储架构有Redis单点，Redis主从，基于Sentinel的Redis主备、基于keepalive的redis主备，以及Redis集群Cluster，还有豌豆荚开源的Codis等是目前业内比较流行的解决方案，不同的存储架构，是若干个技术工程师，根据自己公司的业务，踩过N次坑，总结出的结果，没有最好的解决方案，只有最适合的解决方案。
在上家公司出，采用了redis的主从架构，版本是2.8，上线一年多，尚未出现过问题，但是潜在的隐患是大家都知道的，主从架构中，当主宕机或者意外情况时，主与从不能平滑切换，将会影响业务，所以这里就是一个定时炸弹，而Sentinel(哨兵)方案是主从架构的升级版，能够让主从之间自动切换。提高系统稳定性和高可用性。以下是最近自己测试哨兵架构的相关记录。
Sentinel官方文档：https://redis.io/topics/sentinel
操作系统：CentOS 6.7
redis版本：3.2.9
主机规划：
| 序号 | Host |
Redis Role | Redis Port | Sentinel Port |
| 1 | 10.1.4.82 | slave | 6379 | 26379 26380 |
| 2 | 10.1.4.99 | master | 6379 | 26379 26380 |
整体采用，主从结构，每台主机再部署2个哨兵，整体是2个Redis 4个Sentinel集群的架构。
步骤：
1、配置Redis主从
由于配置redis主从比较简单，所以在这里不会去详细的介绍，想了解如何配置的朋友，请参考以下文章。
http://dreamlinux.blog.51cto.com/9079323/1832029
http://dreamlinux.blog.51cto.com/9079323/1832069
2、检验redis主从是否正常
登录redis的Master主节点检查redis主从
|
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
|
127.0.0.1:6379> info
127.0.0.1:6379> info
# Server
redis_version:3.2.9
redis_git_sha1:00000000
redis_git_dirty:0
redis_build_id:66cf707c3c51d9f
redis_mode:standalone
os:Linux 2.6.32-696.6.3.el6.x86_64 x86_64
arch_bits:64
multiplexing_api:epoll
gcc_version:4.4.7
process_id:26318
run_id:7954f18bcb4084964efb58460027399d583fba57
tcp_port:6379
uptime_in_seconds:14402
uptime_in_days:0
hz:10
lru_clock:9455690
executable:
/usr/local/bin/redis-server
config_file:
/etc/redis/6379
.conf
# Clients
connected_clients:7
client_longest_output_list:0
client_biggest_input_buf:0
blocked_clients:0
# Memory
used_memory:2017176
used_memory_human:1.92M
used_memory_rss:9969664
used_memory_rss_human:9.51M
used_memory_peak:2098064
used_memory_peak_human:2.00M
total_system_memory:2103533568
total_system_memory_human:1.96G
used_memory_lua:37888
used_memory_lua_human:37.00K
maxmemory:0
maxmemory_human:0B
maxmemory_policy:noeviction
mem_fragmentation_ratio:4.94
mem_allocator:jemalloc-4.0.3
# Persistence
loading:0
rdb_changes_since_last_save:0
rdb_bgsave_in_progress:0
rdb_last_save_time:1502627793
rdb_last_bgsave_status:ok
rdb_last_bgsave_time_sec:0
rdb_current_bgsave_time_sec:-1
aof_enabled:0
aof_rewrite_in_progress:0
aof_rewrite_scheduled:0
aof_last_rewrite_time_sec:-1
aof_current_rewrite_time_sec:-1
aof_last_bgrewrite_status:ok
aof_last_write_status:ok
# Stats
total_connections_received:20
total_commands_processed:89212
instantaneous_ops_per_sec:6
total_net_input_bytes:6172607
total_net_output_bytes:33781397
instantaneous_input_kbps:0.26
instantaneous_output_kbps:0.67
rejected_connections:0
sync_full:1
sync_partial_ok:0
sync_partial_err:0
expired_keys:0
evicted_keys:0
keyspace_hits:0
keyspace_misses:0
pubsub_channels:1
pubsub_patterns:0
latest_fork_usec:643
migrate_cached_sockets:0
# Replication
role:master
#当前host的角色
connected_slaves:1
slave0:ip=10.1.4.82,port=6379,state=online,offset=23160,lag=1
#看到从节点的一些信息
master_repl_offset:23160
repl_backlog_active:1
repl_backlog_size:1048576
repl_backlog_first_byte_offset:2
repl_backlog_histlen:23159
# CPU
used_cpu_sys:10.40
used_cpu_user:6.24
used_cpu_sys_children:0.00
used_cpu_user_children:0.00
# Cluster
cluster_enabled:0
# Keyspace
db0:keys=5,expires=0,avg_ttl=0
|
设置相关的key检查
master
|
1
2
3
4
5
6
7
8
9
|
127.0.0.1:6379> dbsize
(integer) 5
127.0.0.1:6379> keys *
1)
"winnerlook"
2)
"redis"
3)
"bund"
4)
"baidu"
5)
"bund33"
127.0.0.1:6379>
|
slave
|
1
2
3
4
5
6
7
8
9
|
127.0.0.1:6379> dbsize
(integer) 5
127.0.0.1:6379> keys *
1)
"winnerlook"
2)
"redis"
3)
"bund"
4)
"baidu"
5)
"bund33"
127.0.0.1:6379>
|
由此可见配置的redis主从是可以正常工作的，接下来就是准备配置Sentinel，并启动相关的服务。
10.1.4.99 sentinel 1配置文件
|
1
2
3
4
5
6
7
8
9
10
|
pdaemonize
yes
#后台运行
protected-mode no
#保护模式关闭
# Generated by CONFIG REWRITE
dir
"/etc/redis"
port 26379
#哨兵端口
sentinel monitor master1 10.1.4.99 6379 2
# 监听IP 端口 默认最少投票数量
sentinel down-after-milliseconds master1 5000
sentinel failover-timeout master1 5000
#选举超时时间
logfile
"/data/redis/sentinel/sentinel_26379.log"
#日志文件
sentinel auth-pass master1 Redis2017
#redis密码
|
10.1.4.99 sentinel 2配置文件
|
1
2
3
4
5
6
7
8
9
10
|
pdaemonize
yes
#后台运行
protected-mode no
#保护模式关闭
# Generated by CONFIG REWRITE
dir
"/etc/redis"
port 26380
#哨兵端口
sentinel monitor master1 10.1.4.99 6379 2
# 监听IP 端口 默认最少投票数量
sentinel down-after-milliseconds master1 5000
sentinel failover-timeout master1 5000
#选举超时时间
logfile
"/data/redis/sentinel/sentinel_26379.log"
#日志文件
sentinel auth-pass master1 Redis2017
#redis密码
|
10.1.4.82 sentinel 1配置
|
1
2
3
4
5
6
7
8
9
10
|
pdaemonize
yes
#后台运行
protected-mode no
#保护模式关闭
# Generated by CONFIG REWRITE
dir
"/etc/redis"
port 26379
#哨兵端口
sentinel monitor master1 10.1.4.99 6379 2
# 监听IP 端口 默认最少投票数量
sentinel down-after-milliseconds master1 5000
sentinel failover-timeout master1 5000
#选举超时时间
logfile
"/data/redis/sentinel/sentinel_26379.log"
#日志文件
sentinel auth-pass master1 Redis2017
#redis密码
|
10.1.4.82 sentinel 2配置
|
1
2
3
4
5
6
7
8
9
10
|
pdaemonize
yes
#后台运行
protected-mode no
#保护模式关闭
# Generated by CONFIG REWRITE
dir
"/etc/redis"
port 26380
#哨兵端口
sentinel monitor master1 10.1.4.99 6379 2
# 监听IP 端口 默认最少投票数量
sentinel down-after-milliseconds master1 5000
sentinel failover-timeout master1 5000
#选举超时时间
logfile
"/data/redis/sentinel/sentinel_26379.log"
#日志文件
sentinel auth-pass master1 Redis2017
#redis密码
|
启动哨兵服务的两种方式
|
1
|
1、redis-sentinel
/path/to/sentinel
.conf
|
|
1
|
2、redis-server
/path/to/sentinel
.conf --sentinel
|
启动10.1.4.99的哨兵服务
|
1
2
3
|
redis-sentinel
/etc/redis/sentinel_26380
.conf
redis-sentinel
/etc/redis/sentinel_26379
.conf
|
检查26379端口相关日志
|
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
|
[root@bundcredit sentinel]
# tail -f sentinel_26379.log
`-._ `-._`-.__.-
'_.-'
_.-'
`-._ `-.__.-
' _.-'
`-._ _.-'
`-.__.-'
27098:X 13 Aug 21:51:25.242
# WARNING: The TCP backlog setting of 511 cannot be enforced because /proc/sys/net/core/somaxconn is set to the lower value of 128.
27098:X 13 Aug 21:51:25.242
# Sentinel ID is 361c7d3014a041a1a74163355882d88f5fab7581
27098:X 13 Aug 21:51:25.242
# +monitor master master1 10.1.4.99 6379 quorum 2
27098:X 13 Aug 21:51:30.271
# +sdown sentinel 76d5a26748410e72ca703e255f44a5ed2b3f83e4 10.1.4.99 26380 @ master1 10.1.4.99 6379
27098:X 13 Aug 21:51:30.410
# -sdown sentinel 76d5a26748410e72ca703e255f44a5ed2b3f83e4 10.1.4.99 26380 @ master1 10.1.4.99 6379
27098:X 13 Aug 21:52:30.617
# +sdown sentinel 5c3033c0ea6d1e0d861e9d9e99e94edb49c5daac 10.1.4.82 26379 @ master1 10.1.4.99 6379
27098:X 13 Aug 21:53:23.855
# -sdown sentinel 5c3033c0ea6d1e0d861e9d9e99e94edb49c5daac 10.1.4.82 26379 @ master1 10.1.4.99 6379
27098:X 13 Aug 21:53:34.463
# +sdown master master1 10.1.4.99 6379
27098:X 13 Aug 21:53:34.532
# +odown master master1 10.1.4.99 6379 #quorum 2/2
27098:X 13 Aug 21:53:34.532
# +new-epoch 1
27098:X 13 Aug 21:53:34.532
# +try-failover master master1 10.1.4.99 6379
27098:X 13 Aug 21:53:34.541
# +vote-for-leader 361c7d3014a041a1a74163355882d88f5fab7581 1
27098:X 13 Aug 21:53:34.542
# b73258f1019f65c5fee3b88b9c2eae7f85d985f3 voted for b73258f1019f65c5fee3b88b9c2eae7f85d985f3 1
27098:X 13 Aug 21:53:34.559
# 5c3033c0ea6d1e0d861e9d9e99e94edb49c5daac voted for b73258f1019f65c5fee3b88b9c2eae7f85d985f3 1
27098:X 13 Aug 21:53:34.560
# 76d5a26748410e72ca703e255f44a5ed2b3f83e4 voted for b73258f1019f65c5fee3b88b9c2eae7f85d985f3 1
27098:X 13 Aug 21:53:35.156
# +config-update-from sentinel b73258f1019f65c5fee3b88b9c2eae7f85d985f3 10.1.4.82 26380 @ master1 10.1.4.99 6379
27098:X 13 Aug 21:53:35.156
# +switch-master master1 10.1.4.99 6379 10.1.4.82 6379
27098:X 13 Aug 21:53:35.156 * +slave slave 10.1.4.99:6379 10.1.4.99 6379 @ master1 10.1.4.82 6379
27098:X 13 Aug 21:53:40.173
# +sdown slave 10.1.4.99:6379 10.1.4.99 6379 @ master1 10.1.4.82 6379
27098:X 13 Aug 21:54:07.583
# -sdown slave 10.1.4.99:6379 10.1.4.99 6379 @ master1 10.1.4.82 6379
|
检查26380端口相关日志
|
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
|
[root@bundcredit sentinel]
# tail -f sentinel_26379.log
`-._ `-._`-.__.-
'_.-'
_.-'
`-._ `-.__.-
' _.-'
`-._ _.-'
`-.__.-'
27098:X 13 Aug 21:51:25.242
# WARNING: The TCP backlog setting of 511 cannot be enforced because /proc/sys/net/core/somaxconn is set to the lower value of 128.
27098:X 13 Aug 21:51:25.242
# Sentinel ID is 361c7d3014a041a1a74163355882d88f5fab7581
27098:X 13 Aug 21:51:25.242
# +monitor master master1 10.1.4.99 6379 quorum 2
27098:X 13 Aug 21:51:30.271
# +sdown sentinel 76d5a26748410e72ca703e255f44a5ed2b3f83e4 10.1.4.99 26380 @ master1 10.1.4.99 6379
27098:X 13 Aug 21:51:30.410
# -sdown sentinel 76d5a26748410e72ca703e255f44a5ed2b3f83e4 10.1.4.99 26380 @ master1 10.1.4.99 6379
27098:X 13 Aug 21:52:30.617
# +sdown sentinel 5c3033c0ea6d1e0d861e9d9e99e94edb49c5daac 10.1.4.82 26379 @ master1 10.1.4.99 6379
27098:X 13 Aug 21:53:23.855
# -sdown sentinel 5c3033c0ea6d1e0d861e9d9e99e94edb49c5daac 10.1.4.82 26379 @ master1 10.1.4.99 6379
27098:X 13 Aug 21:53:34.463
# +sdown master master1 10.1.4.99 6379
27098:X 13 Aug 21:53:34.532
# +odown master master1 10.1.4.99 6379 #quorum 2/2
27098:X 13 Aug 21:53:34.532
# +new-epoch 1
27098:X 13 Aug 21:53:34.532
# +try-failover master master1 10.1.4.99 6379
27098:X 13 Aug 21:53:34.541
# +vote-for-leader 361c7d3014a041a1a74163355882d88f5fab7581 1
27098:X 13 Aug 21:53:34.542
# b73258f1019f65c5fee3b88b9c2eae7f85d985f3 voted for b73258f1019f65c5fee3b88b9c2eae7f85d985f3 1
27098:X 13 Aug 21:53:34.559
# 5c3033c0ea6d1e0d861e9d9e99e94edb49c5daac voted for b73258f1019f65c5fee3b88b9c2eae7f85d985f3 1
27098:X 13 Aug 21:53:34.560
# 76d5a26748410e72ca703e255f44a5ed2b3f83e4 voted for b73258f1019f65c5fee3b88b9c2eae7f85d985f3 1
27098:X 13 Aug 21:53:35.156
# +config-update-from sentinel b73258f1019f65c5fee3b88b9c2eae7f85d985f3 10.1.4.82 26380 @ master1 10.1.4.99 6379
27098:X 13 Aug 21:53:35.156
# +switch-master master1 10.1.4.99 6379 10.1.4.82 6379
27098:X 13 Aug 21:53:35.156 * +slave slave 10.1.4.99:6379 10.1.4.99 6379 @ master1 10.1.4.82 6379
27098:X 13 Aug 21:53:40.173
# +sdown slave 10.1.4.99:6379 10.1.4.99 6379 @ master1 10.1.4.82 6379
27098:X 13 Aug 21:54:07.583
# -sdown slave 10.1.4.99:6379 10.1.4.99 6379 @ master1 10.1.4.82 6379
^C
[root@bundcredit sentinel]
# tail -f sentinel_26369.log
tail
: cannot
open
`sentinel_26369.log'
for
reading: No such
file
or directory
tail
: no files remaining
[root@bundcredit sentinel]
# tail -f sentinel_26380.log
27102:X 13 Aug 21:53:34.496
# +sdown master master1 10.1.4.99 6379
27102:X 13 Aug 21:53:34.558
# +new-epoch 1
27102:X 13 Aug 21:53:34.560
# +vote-for-leader b73258f1019f65c5fee3b88b9c2eae7f85d985f3 1
27102:X 13 Aug 21:53:34.580
# +odown master master1 10.1.4.99 6379 #quorum 4/2
27102:X 13 Aug 21:53:34.580
# Next failover delay: I will not start a failover before Sun Aug 13 21:53:45 2017
27102:X 13 Aug 21:53:35.156
# +config-update-from sentinel b73258f1019f65c5fee3b88b9c2eae7f85d985f3 10.1.4.82 26380 @ master1 10.1.4.99 6379
27102:X 13 Aug 21:53:35.156
# +switch-master master1 10.1.4.99 6379 10.1.4.82 6379
27102:X 13 Aug 21:53:35.156 * +slave slave 10.1.4.99:6379 10.1.4.99 6379 @ master1 10.1.4.82 6379
27102:X 13 Aug 21:53:40.201
# +sdown slave 10.1.4.99:6379 10.1.4.99 6379 @ master1 10.1.4.82 6379
27102:X 13 Aug 21:54:07.583
# -sdown slave 10.1.4.99:6379 10.1.4.99 6379 @ master1 10.1.4.82 6379
|
启动10.1.4.82服务器相关服务
|
1
2
3
|
redis-sentinel
/etc/redis/sentinel_26380
.conf
redis-sentinel
/etc/redis/sentinel_26379
.conf
|
检查26379端口相关日志
|
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
|
15947:X 13 Aug 21:45:33.882
# WARNING: The TCP backlog setting of 511 cannot be enforced because /proc/sys/net/core/somaxconn is set to the lower value of 128.
15947:X 13 Aug 21:45:33.893
# Sentinel ID is 5c3033c0ea6d1e0d861e9d9e99e94edb49c5daac
15947:X 13 Aug 21:45:33.893
# +monitor master master1 10.1.4.99 6379 quorum 2
15947:X 13 Aug 21:45:33.894 * +slave slave 10.1.4.82:6379 10.1.4.82 6379 @ master1 10.1.4.99 6379
15951:X 13 Aug 21:45:39.170 * Increased maximum number of
open
files to 10032 (it was originally
set
to 1024).
_._
_.-``__
''
-._
_.-`` `. `_.
''
-._ Redis 3.2.9 (00000000
/0
) 64 bit
.-`` .-```. ```\/ _.,_
''
-._
( ' , .-` | `, ) Running
in
sentinel mode
|`-._`-...-` __...-.``-._|
'` _.-'
| Port: 26380
| `-._ `._ / _.-' | PID: 15951
`-._ `-._ `-./ _.-
' _.-'
|`-._`-._ `-.__.-
' _.-'
_.-'|
| `-._`-._ _.-
'_.-'
| http:
//redis
.io
`-._ `-._`-.__.-
'_.-'
_.-'
|`-._`-._ `-.__.-
' _.-'
_.-'|
| `-._`-._ _.-
'_.-'
|
`-._ `-._`-.__.-
'_.-'
_.-'
`-._ `-.__.-
' _.-'
`-._ _.-'
`-.__.-'
15951:X 13 Aug 21:45:39.171
# WARNING: The TCP backlog setting of 511 cannot be enforced because /proc/sys/net/core/somaxconn is set to the lower value of 128.
15951:X 13 Aug 21:45:39.174
# Sentinel ID is b73258f1019f65c5fee3b88b9c2eae7f85d985f3
15951:X 13 Aug 21:45:39.174
# +monitor master master1 10.1.4.99 6379 quorum 2
15951:X 13 Aug 21:45:39.175 * +slave slave 10.1.4.82:6379 10.1.4.82 6379 @ master1 10.1.4.99 6379
15951:X 13 Aug 21:45:40.025 * +sentinel sentinel 5c3033c0ea6d1e0d861e9d9e99e94edb49c5daac 10.1.4.82 26379 @ master1 10.1.4.99 6379
15947:X 13 Aug 21:45:41.196 * +sentinel sentinel b73258f1019f65c5fee3b88b9c2eae7f85d985f3 10.1.4.82 26380 @ master1 10.1.4.99 6379
15951:X 13 Aug 21:45:57.729 * +sentinel sentinel 361c7d3014a041a1a74163355882d88f5fab7581 10.1.4.99 26379 @ master1 10.1.4.99 6379
15947:X 13 Aug 21:45:57.729 * +sentinel sentinel 361c7d3014a041a1a74163355882d88f5fab7581 10.1.4.99 26379 @ master1 10.1.4.99 6379
15947:X 13 Aug 21:46:05.114 * +sentinel sentinel 76d5a26748410e72ca703e255f44a5ed2b3f83e4 10.1.4.99 26380 @ master1 10.1.4.99 6379
15951:X 13 Aug 21:46:05.115 * +sentinel sentinel 76d5a26748410e72ca703e255f44a5ed2b3f83e4 10.1.4.99 26380 @ master1 10.1.4.99 6379
15951:X 13 Aug 21:47:16.648
# +sdown master master1 10.1.4.99 6379
15951:X 13 Aug 21:47:20.907
# +sdown sentinel 361c7d3014a041a1a74163355882d88f5fab7581 10.1.4.99 26379 @ master1 10.1.4.99 6379
15951:X 13 Aug 21:47:20.908
# +sdown sentinel 76d5a26748410e72ca703e255f44a5ed2b3f83e4 10.1.4.99 26380 @ master1 10.1.4.99 6379
15947:X 13 Aug 21:47:20.908
# +sdown sentinel 361c7d3014a041a1a74163355882d88f5fab7581 10.1.4.99 26379 @ master1 10.1.4.99 6379
15951:X 13 Aug 21:47:20.979
# +tilt #tilt mode entered
15947:X 13 Aug 21:47:20.979
# +tilt #tilt mode entered
15947:X 13 Aug 21:47:50.982
# -tilt #tilt mode exited
15947:X 13 Aug 21:47:50.983
# -sdown sentinel 361c7d3014a041a1a74163355882d88f5fab7581 10.1.4.99 26379 @ master1 10.1.4.99 6379
15951:X 13 Aug 21:47:51.030
# -tilt #tilt mode exited
15951:X 13 Aug 21:47:51.030
# -sdown master master1 10.1.4.99 6379
15951:X 13 Aug 21:47:51.030
# -sdown sentinel 361c7d3014a041a1a74163355882d88f5fab7581 10.1.4.99 26379 @ master1 10.1.4.99 6379
15951:X 13 Aug 21:47:51.030
# -sdown sentinel 76d5a26748410e72ca703e255f44a5ed2b3f83e4 10.1.4.99 26380 @ master1 10.1.4.99 6379
15947:X 13 Aug 21:48:42.378
# +sdown sentinel b73258f1019f65c5fee3b88b9c2eae7f85d985f3 10.1.4.82 26380 @ master1 10.1.4.99 6379
15947:X 13 Aug 21:49:06.458
# -sdown sentinel b73258f1019f65c5fee3b88b9c2eae7f85d985f3 10.1.4.82 26380 @ master1 10.1.4.99 6379
15947:X 13 Aug 21:49:42.140
# +sdown sentinel 361c7d3014a041a1a74163355882d88f5fab7581 10.1.4.99 26379 @ master1 10.1.4.99 6379
15947:X 13 Aug 21:50:59.960
# +sdown sentinel 76d5a26748410e72ca703e255f44a5ed2b3f83e4 10.1.4.99 26380 @ master1 10.1.4.99 6379
15947:X 13 Aug 21:51:20.923
# -sdown sentinel 361c7d3014a041a1a74163355882d88f5fab7581 10.1.4.99 26379 @ master1 10.1.4.99 6379
15947:X 13 Aug 21:51:26.062
# -sdown sentinel 76d5a26748410e72ca703e255f44a5ed2b3f83e4 10.1.4.99 26380 @ master1 10.1.4.99 6379
15947:X 13 Aug 21:52:20.592
# User requested shutdown...
15947:X 13 Aug 21:52:20.592 * Removing the pid
file
.
15947:X 13 Aug 21:52:20.592
# Sentinel is now ready to exit, bye bye...
17056:X 13 Aug 21:53:18.720 * Increased maximum number of
open
files to 10032 (it was originally
set
to 1024).
_._
_.-``__
''
-._
_.-`` `. `_.
''
-._ Redis 3.2.9 (00000000
/0
) 64 bit
.-`` .-```. ```\/ _.,_
''
-._
( ' , .-` | `, ) Running
in
sentinel mode
|`-._`-...-` __...-.``-._|
'` _.-'
| Port: 26379
| `-._ `._ / _.-' | PID: 17056
`-._ `-._ `-./ _.-
' _.-'
|`-._`-._ `-.__.-
' _.-'
_.-'|
| `-._`-._ _.-
'_.-'
| http:
//redis
.io
`-._ `-._`-.__.-
'_.-'
_.-'
|`-._`-._ `-.__.-
' _.-'
_.-'|
| `-._`-._ _.-
'_.-'
|
`-._ `-._`-.__.-
'_.-'
_.-'
`-._ `-.__.-
' _.-'
`-._ _.-'
`-.__.-'
17056:X 13 Aug 21:53:18.721
# WARNING: The TCP backlog setting of 511 cannot be enforced because /proc/sys/net/core/somaxconn is set to the lower value of 128.
17056:X 13 Aug 21:53:18.721
# Sentinel ID is 5c3033c0ea6d1e0d861e9d9e99e94edb49c5daac
17056:X 13 Aug 21:53:18.721
# +monitor master master1 10.1.4.99 6379 quorum 2
17056:X 13 Aug 21:53:29.549
# +sdown master master1 10.1.4.99 6379
17056:X 13 Aug 21:53:29.643
# +new-epoch 1
17056:X 13 Aug 21:53:29.645
# +vote-for-leader b73258f1019f65c5fee3b88b9c2eae7f85d985f3 1
17056:X 13 Aug 21:53:29.645
# +odown master master1 10.1.4.99 6379 #quorum 2/2
17056:X 13 Aug 21:53:29.645
# Next failover delay: I will not start a failover before Sun Aug 13 21:53:39 2017
17056:X 13 Aug 21:53:30.241
# +config-update-from sentinel b73258f1019f65c5fee3b88b9c2eae7f85d985f3 10.1.4.82 26380 @ master1 10.1.4.99 6379
17056:X 13 Aug 21:53:30.241
# +switch-master master1 10.1.4.99 6379 10.1.4.82 6379
17056:X 13 Aug 21:53:30.241 * +slave slave 10.1.4.99:6379 10.1.4.99 6379 @ master1 10.1.4.82 6379
17056:X 13 Aug 21:53:35.285
# +sdown slave 10.1.4.99:6379 10.1.4.99 6379 @ master1 10.1.4.82 6379
17056:X 13 Aug 21:54:02.488
# -sdown slave 10.1.4.99:6379 10.1.4.99 6379 @ master1 10.1.4.82 6379
17056:X 13 Aug 21:54:12.441 * +convert-to-slave slave 10.1.4.99:6379 10.1.4.99 6379 @ master1 10.1.4.82 6379
|
检查26380端口相关日志
|
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
|
[root@localhost redis]
# tail -200f /data/redis/sentinel/sentinel_26380.log
16390:X 13 Aug 21:49:05.965 * Increased maximum number of
open
files to 10032 (it was originally
set
to 1024).
_._
_.-``__
''
-._
_.-`` `. `_.
''
-._ Redis 3.2.9 (00000000
/0
) 64 bit
.-`` .-```. ```\/ _.,_
''
-._
( ' , .-` | `, ) Running
in
sentinel mode
|`-._`-...-` __...-.``-._|
'` _.-'
| Port: 26380
| `-._ `._ / _.-' | PID: 16390
`-._ `-._ `-./ _.-
' _.-'
|`-._`-._ `-.__.-
' _.-'
_.-'|
| `-._`-._ _.-
'_.-'
| http:
//redis
.io
`-._ `-._`-.__.-
'_.-'
_.-'
|`-._`-._ `-.__.-
' _.-'
_.-'|
| `-._`-._ _.-
'_.-'
|
`-._ `-._`-.__.-
'_.-'
_.-'
`-._ `-.__.-
' _.-'
`-._ _.-'
`-.__.-'
16390:X 13 Aug 21:49:05.966
# WARNING: The TCP backlog setting of 511 cannot be enforced because /proc/sys/net/core/somaxconn is set to the lower value of 128.
16390:X 13 Aug 21:49:05.966
# Sentinel ID is b73258f1019f65c5fee3b88b9c2eae7f85d985f3
16390:X 13 Aug 21:49:05.966
# +monitor master master1 10.1.4.99 6379 quorum 2
16390:X 13 Aug 21:49:42.173
# +sdown sentinel 361c7d3014a041a1a74163355882d88f5fab7581 10.1.4.99 26379 @ master1 10.1.4.99 6379
16390:X 13 Aug 21:50:59.918
# +sdown sentinel 76d5a26748410e72ca703e255f44a5ed2b3f83e4 10.1.4.99 26380 @ master1 10.1.4.99 6379
16390:X 13 Aug 21:51:21.049
# -sdown sentinel 361c7d3014a041a1a74163355882d88f5fab7581 10.1.4.99 26379 @ master1 10.1.4.99 6379
16390:X 13 Aug 21:51:25.188
# -sdown sentinel 76d5a26748410e72ca703e255f44a5ed2b3f83e4 10.1.4.99 26380 @ master1 10.1.4.99 6379
16390:X 13 Aug 21:52:25.659
# +sdown sentinel 5c3033c0ea6d1e0d861e9d9e99e94edb49c5daac 10.1.4.82 26379 @ master1 10.1.4.99 6379
16390:X 13 Aug 21:53:19.651
# -sdown sentinel 5c3033c0ea6d1e0d861e9d9e99e94edb49c5daac 10.1.4.82 26379 @ master1 10.1.4.99 6379
16390:X 13 Aug 21:53:29.563
# +sdown master master1 10.1.4.99 6379
16390:X 13 Aug 21:53:29.615
# +odown master master1 10.1.4.99 6379 #quorum 3/2
16390:X 13 Aug 21:53:29.615
# +new-epoch 1
16390:X 13 Aug 21:53:29.615
# +try-failover master master1 10.1.4.99 6379
16390:X 13 Aug 21:53:29.626
# +vote-for-leader b73258f1019f65c5fee3b88b9c2eae7f85d985f3 1
16390:X 13 Aug 21:53:29.627
# 361c7d3014a041a1a74163355882d88f5fab7581 voted for 361c7d3014a041a1a74163355882d88f5fab7581 1
16390:X 13 Aug 21:53:29.645
# 5c3033c0ea6d1e0d861e9d9e99e94edb49c5daac voted for b73258f1019f65c5fee3b88b9c2eae7f85d985f3 1
16390:X 13 Aug 21:53:29.646
# 76d5a26748410e72ca703e255f44a5ed2b3f83e4 voted for b73258f1019f65c5fee3b88b9c2eae7f85d985f3 1
16390:X 13 Aug 21:53:29.698
# +elected-leader master master1 10.1.4.99 6379
16390:X 13 Aug 21:53:29.698
# +failover-state-select-slave master master1 10.1.4.99 6379
16390:X 13 Aug 21:53:29.764
# +selected-slave slave 10.1.4.82:6379 10.1.4.82 6379 @ master1 10.1.4.99 6379
16390:X 13 Aug 21:53:29.764 * +failover-state-send-slaveof-noone slave 10.1.4.82:6379 10.1.4.82 6379 @ master1 10.1.4.99 6379
16390:X 13 Aug 21:53:29.826 * +failover-state-wait-promotion slave 10.1.4.82:6379 10.1.4.82 6379 @ master1 10.1.4.99 6379
16390:X 13 Aug 21:53:30.179
# +promoted-slave slave 10.1.4.82:6379 10.1.4.82 6379 @ master1 10.1.4.99 6379
16390:X 13 Aug 21:53:30.179
# +failover-state-reconf-slaves master master1 10.1.4.99 6379
16390:X 13 Aug 21:53:30.237
# +failover-end master master1 10.1.4.99 6379
16390:X 13 Aug 21:53:30.237
# +switch-master master1 10.1.4.99 6379 10.1.4.82 6379
16390:X 13 Aug 21:53:30.238 * +slave slave 10.1.4.99:6379 10.1.4.99 6379 @ master1 10.1.4.82 6379
16390:X 13 Aug 21:53:35.295
# +sdown slave 10.1.4.99:6379 10.1.4.99 6379 @ master1 10.1.4.82 6379
16390:X 13 Aug 21:54:02.671
# -sdown slave 10.1.4.99:6379 10.1.4.99 6379 @ master1 10.1.4.82 6379
|
检查redis主从和哨兵的状态
10.1.4.99 主从状态
|
1
2
3
4
5
6
7
8
9
10
|
127.0.0.1:6379> info [section]
# Replication
role:master
connected_slaves:1
slave0:ip=10.1.4.82,port=6379,state=online,offset=2678,lag=0
master_repl_offset:2678
repl_backlog_active:1
repl_backlog_size:1048576
repl_backlog_first_byte_offset:2
repl_backlog_histlen:2677
|
哨兵状态
|
1
2
3
4
5
6
7
8
9
|
[root@bundcredit redis]
# redis-cli -p 26379 -a "Redis2017"
127.0.0.1:26379> info
# Sentinel
sentinel_masters:1
sentinel_tilt:0
sentinel_running_scripts:0
sentinel_scripts_queue_length:0
sentinel_simulate_failure_flags:0
master0:name=master1,status=ok,address=10.1.4.99:6379,slaves=1,sentinels=4
|
10.1.4.82 主从状态
|
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
|
127.0.0.1:6379> info [section]
# Replication
# Replication
role:slave
master_host:10.1.4.99
master_port:6379
master_link_status:up
master_last_io_seconds_ago:1
master_sync_in_progress:0
slave_repl_offset:49526
slave_priority:100
slave_read_only:0
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
|
哨兵状态
|
1
2
3
4
5
6
7
8
9
10
|
[root@bundcredit redis]
# redis-cli -p 26379 -a "Redis2017"
127.0.0.1:26379> info
# Sentinel
sentinel_masters:1
sentinel_tilt:0
sentinel_running_scripts:0
sentinel_scripts_queue_length:0
sentinel_simulate_failure_flags:0
master0:name=master1,status=ok,address=10.1.4.99:6379,slaves=1,sentinels=4
127.0.0.1:26379>
|
异常测试
1、停止10.1.4.99的redis服务
|
1
2
3
|
[root@bundcredit redis]
# redis-cli -h 10.1.4.99 -p 6379 -a "Redis2017"
10.1.4.99:6379> SHUTDOWN
not connected>
|
2、哨兵日志
|
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
|
27102:X 13 Aug 21:53:40.201
# +sdown slave 10.1.4.99:6379 10.1.4.99 6379 @ master1 10.1.4.82 6379
27102:X 13 Aug 21:54:07.583
# -sdown slave 10.1.4.99:6379 10.1.4.99 6379 @ master1 10.1.4.82 6379
27102:X 13 Aug 22:17:23.410
# +sdown master master1 10.1.4.82 6379
27102:X 13 Aug 22:17:23.445
# +new-epoch 2
27102:X 13 Aug 22:17:23.447
# +vote-for-leader 5c3033c0ea6d1e0d861e9d9e99e94edb49c5daac 2
27102:X 13 Aug 22:17:23.462
# +odown master master1 10.1.4.82 6379 #quorum 3/2
27102:X 13 Aug 22:17:23.462
# Next failover delay: I will not start a failover before Sun Aug 13 22:17:34 2017
27102:X 13 Aug 22:17:24.499
# -odown master master1 10.1.4.82 6379
27102:X 13 Aug 22:17:24.533
# +config-update-from sentinel 5c3033c0ea6d1e0d861e9d9e99e94edb49c5daac 10.1.4.82 26379 @ master1 10.1.4.82 6379
27102:X 13 Aug 22:17:24.533
# +switch-master master1 10.1.4.82 6379 10.1.4.99 6379
27102:X 13 Aug 22:17:24.534 * +slave slave 10.1.4.82:6379 10.1.4.82 6379 @ master1 10.1.4.99 6379
27102:X 13 Aug 22:17:29.580
# +sdown slave 10.1.4.82:6379 10.1.4.82 6379 @ master1 10.1.4.99 6379
27102:X 13 Aug 22:17:33.951
# -sdown slave 10.1.4.82:6379 10.1.4.82 6379 @ master1 10.1.4.99 6379
27102:X 13 Aug 22:17:43.855 * +convert-to-slave slave 10.1.4.82:6379 10.1.4.82 6379 @ master1 10.1.4.99 6379
27102:X 13 Aug 22:24:14.405
# +sdown master master1 10.1.4.99 6379
27102:X 13 Aug 22:24:14.445
# +new-epoch 3
27102:X 13 Aug 22:24:14.446
# +vote-for-leader b73258f1019f65c5fee3b88b9c2eae7f85d985f3 3
27102:X 13 Aug 22:24:14.477
# +odown master master1 10.1.4.99 6379 #quorum 4/2
27102:X 13 Aug 22:24:14.477
# Next failover delay: I will not start a failover before Sun Aug 13 22:24:24 2017
27102:X 13 Aug 22:24:15.600
# +config-update-from sentinel b73258f1019f65c5fee3b88b9c2eae7f85d985f3 10.1.4.82 26380 @ master1 10.1.4.99 6379
27102:X 13 Aug 22:24:15.601
# +switch-master master1 10.1.4.99 6379 10.1.4.82 6379
27102:X 13 Aug 22:24:15.601 * +slave slave 10.1.4.99:6379 10.1.4.99 6379 @ master1 10.1.4.82 6379
27102:X 13 Aug 22:24:20.627
# +sdown slave 10.1.4.99:6379 10.1.4.99 6379 @ master1 10.1.4.82 6379
27102:X 13 Aug 22:25:49.007
# -sdown slave 10.1.4.99:6379 10.1.4.99 6379 @ master1 10.1.4.82 6379
27102:X 13 Aug 22:25:58.984 * +convert-to-slave slave 10.1.4.99:6379 10.1.4.99 6379 @ master1 10.1.4.82 6379
|
3、检查服务
|
1
2
3
4
5
6
7
8
9
10
|
10.1.4.82
# Replication
role:master
connected_slaves:1
slave0:ip=10.1.4.99,port=6379,state=online,offset=17942,lag=0
master_repl_offset:17942
repl_backlog_active:1
repl_backlog_size:1048576
repl_backlog_first_byte_offset:2
repl_backlog_histlen:17941
|
启动10.1.4.99
检查服务
|
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
|
# Replication
role:slave
master_host:10.1.4.82
master_port:6379
master_link_status:up
master_last_io_seconds_ago:0
master_sync_in_progress:0
slave_repl_offset:2546
slave_priority:100
slave_read_only:0
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
|
再次停止10.1.4.82 redis服务服务正常
注意事项：
在配置redis.conf时需要注意以下事项
|
1
2
|
1、protected-mode no
#保护模式关闭 在3.2版本后添加的安全模式 最好关闭
2、slave-
read
-only no
#默认的从库不支持写入，这里需要开启，当经过哨兵写入时才支持写入数据
|
在配置sentinel.conf时需要注意
|
1
2
3
4
5
6
7
8
|
1、pdaemonize
yes
#后台运行
2、protected-mode no
#保护模式关闭
3、sentinel monitor master1 10.1.4.99 6379 2
# 监听IP 端口 默认最少投票数量
4、sentinel down-after-milliseconds master1 5000
5、sentinel failover-timeout master1 5000
#选举超时时间
6、logfile
"/data/redis/sentinel/sentinel_26379.log"
#日志文件
7、sentinel auth-pass master1 Redis2017
#redis密码 如果主从设置密码后记得要设置密码，否则不能登录
|
本文转自 tianya1993 51CTO博客，原文链接：http://blog.51cto.com/dreamlinux/1955969，如需转载请自行联系原作者
