本文最后更新于 257 天前，其中的信息可能已经过时，如有错误请发送邮件到wuxianglongblog@163.com
Redis的高可用架构单哨兵(sentinel)模式部署实战
一.redis 哨兵(sentinel)模式工作原理概述
redis 哨兵(sentinel)模式的优点:
(1)监控master库，当master库宕机后会自动进行选主切换;
(2)切换过程是应用透明的，因此应用程序无法感知;
(3)自动处理故障节点;
生产环境中，我们可以将多个Redis sentinel实例部署在K8S集群上，这样就可以借助K8S的一些工作机制来实现Redis的故障自愈功能哟~
二.部署redis 哨兵(sentinel)模式
1.准备3个Redis实例(可以直接基于上一节搭建的主从复制环境即可)
[root@redis201.oldboyedu.com ~]# ss -ntl
State Recv-Q Send-Q Local Address:Port Peer Address:Port
LISTEN 0 128 127.0.0.1:16370 *:*
LISTEN 0 128 172.200.1.201:16370 *:*
LISTEN 0 128 127.0.0.1:16371 *:*
LISTEN 0 128 172.200.1.201:16371 *:*
LISTEN 0 128 127.0.0.1:16372 *:*
LISTEN 0 128 172.200.1.201:16372 *:*
LISTEN 0 128 *:22 *:*
LISTEN 0 128 [::]:22 [::]:*
[root@redis201.oldboyedu.com ~]#
[root@redis201.oldboyedu.com ~]# redis-cli -p 16370 -a oldboyedu INFO REPLICATION
# Replication
role:master
connected_slaves:2
slave0:ip=172.200.1.201,port=16371,state=online,offset=90616,lag=0
slave1:ip=172.200.1.201,port=16372,state=online,offset=90616,lag=0
master_repl_offset:90616
repl_backlog_active:1
repl_backlog_size:1048576
repl_backlog_first_byte_offset:2
repl_backlog_histlen:90615
[root@redis201.oldboyedu.com ~]#
2.创建sentinel的配置文件，数据，日志存储目录
[root@redis201.oldboyedu.com ~]# mkdir -pv /oldboyedu/softwares/redis16379
mkdir: 已创建目录 "/oldboyedu/softwares/redis16379"
[root@redis201.oldboyedu.com ~]#
[root@redis201.oldboyedu.com ~]# mkdir -pv /oldboyedu/data/redis16379
mkdir: 已创建目录 "/oldboyedu/data/redis16379"
[root@redis201.oldboyedu.com ~]#
[root@redis201.oldboyedu.com ~]# mkdir -pv /oldboyedu/logs/redis16379
mkdir: 已创建目录 "/oldboyedu/logs/redis16379"
[root@redis201.oldboyedu.com ~]#
3.创建sentinel的配置文件
[root@redis201.oldboyedu.com ~]# cat > /oldboyedu/softwares/redis16379/sentinel.conf <<EOF
> bind 172.200.1.201 127.0.0.1
> port 16379
> dir /oldboyedu/data/redis16379
> sentinel monitor oldboyedu_master 172.200.1.201 16370 1
> sentinel down-after-milliseconds oldboyedu_master 5000
> sentinel auth-pass oldboyedu_master oldboyedu
> EOF
[root@redis201.oldboyedu.com ~]#
[root@redis201.oldboyedu.com ~]# cat /oldboyedu/softwares/redis16379/sentinel.conf
bind 172.200.1.201 127.0.0.1
port 16379
dir /oldboyedu/data/redis16379
sentinel monitor oldboyedu_master 172.200.1.201 16370 1
sentinel down-after-milliseconds oldboyedu_master 5000
sentinel auth-pass oldboyedu_master oldboyedu
[root@redis201.oldboyedu.com ~]#
配置文件参数说明如下:
sentinel monitor oldboyedu_master 172.200.1.201 16370 1:
指定Redis初始集群的master库节点并命名为"oldboyedu_master"，注意这个"1"的含义，他表示sentinel需要最少的投票数。这在于多sentinel的场景下很有用，由于本案例只是用了1个sentinel，因此我就设置为1。
在生产环境中，我建议大家配置多sentinel的模式，建议最少设置为3(建议大家设置奇数)个sentinel物理节点，而后将此处的"1"改为"2"，表示3个sentinel中最少有2个节点认为master库宕机时，才会真正意义上认为master库宕机。
如果想要了解多sentinel如何配置可参考下一篇笔记中有详细案例哟~
sentinel down-after-milliseconds oldboyedu_master 5000:
监控sentinel集群时，如果超过了5000毫秒(即5秒)仍然没有响应，则sentinel会判定master库宕机了。
sentinel auth-pass oldboyedu_master oldboyedu:
由于sentinel需要访问Redis集群，因此我们要设置访问整个集群的密码，我这里指定的密码为"oldboyedu"，这意味着所有的节点都使用相同的密码。
4.启动sentinel进程并查看日志
(1)启动sentinel进程
[root@redis201.oldboyedu.com ~]# redis-sentinel /oldboyedu/softwares/redis16379/sentinel.conf &> /oldboyedu/logs/redis16379/sentinel.log &
[1] 14144
[root@redis201.oldboyedu.com ~]#
(1)查看日志信息
[root@redis201.oldboyedu.com ~]# tail -100f /oldboyedu/logs/redis16379/sentinel.log
14144:X 04 Mar 07:09:26.063 * Increased maximum number of open files to 10032 (it was originally set to 1024).
_._
_.-``__ ''-._
_.-`` `. `_. ''-._ Redis 3.2.13 (00000000/0) 64 bit
.-`` .-```. ```\/ _.,_ ''-._
( ' , .-` | `, ) Running in sentinel mode
|`-._`-...-` __...-.``-._|'` _.-'| Port: 16379
| `-._ `._ / _.-' | PID: 14144
`-._ `-._ `-./ _.-' _.-'
|`-._`-._ `-.__.-' _.-'_.-'|
| `-._`-._ _.-'_.-' | http://redis.io
`-._ `-._`-.__.-'_.-' _.-'
|`-._`-._ `-.__.-' _.-'_.-'|
| `-._`-._ _.-'_.-' |
`-._ `-._`-.__.-'_.-' _.-'
`-._ `-.__.-' _.-'
`-._ _.-'
`-.__.-'
14144:X 04 Mar 07:09:26.063 # WARNING: The TCP backlog setting of 511 cannot be enforced because /proc/sys/net/core/somaxconn is set to the lower value of 128.
14144:X 04 Mar 07:09:26.064 # Sentinel ID is 13c5048514a67a5a66b0792cb3b64a3a85e10b4b
14144:X 04 Mar 07:09:26.064 # +monitor master oldboyedu_master 172.200.1.201 16370 quorum 1
14144:X 04 Mar 07:09:26.065 * +slave slave 172.200.1.201:16371 172.200.1.201 16371 @ oldboyedu_master 172.200.1.201 16370
14144:X 04 Mar 07:09:26.066 * +slave slave 172.200.1.201:16372 172.200.1.201 16372 @ oldboyedu_master 172.200.1.201 16370
(3)如下所示，我们可以登录到"sentinel"后在查看日子信息:
[root@redis201.oldboyedu.com ~]# ss -ntl
State Recv-Q Send-Q Local Address:Port Peer Address:Port
LISTEN 0 128 127.0.0.1:16370 *:*
LISTEN 0 128 172.200.1.201:16370 *:*
LISTEN 0 128 127.0.0.1:16371 *:*
LISTEN 0 128 172.200.1.201:16371 *:*
LISTEN 0 128 127.0.0.1:16372 *:*
LISTEN 0 128 172.200.1.201:16372 *:*
LISTEN 0 128 *:22 *:*
LISTEN 0 128 127.0.0.1:16379 *:*
LISTEN 0 128 172.200.1.201:16379 *:*
LISTEN 0 128 [::]:22 [::]:*
[root@redis201.oldboyedu.com ~]#
[root@redis201.oldboyedu.com ~]# redis-cli -p 16379
127.0.0.1:16379> PING
PONG
127.0.0.1:16379>
127.0.0.1:16379> SHUTDOWN
not connected>
not connected> QUIT
[1]+ 完成 redis-sentinel /oldboyedu/softwares/redis16379/sentinel.conf &>/oldboyedu/logs/redis16379/sentinel.log
[root@redis201.oldboyedu.com ~]#
[root@redis201.oldboyedu.com ~]# ss -ntl
State Recv-Q Send-Q Local Address:Port Peer Address:Port
LISTEN 0 128 127.0.0.1:16370 *:*
LISTEN 0 128 172.200.1.201:16370 *:*
LISTEN 0 128 127.0.0.1:16371 *:*
LISTEN 0 128 172.200.1.201:16371 *:*
LISTEN 0 128 127.0.0.1:16372 *:*
LISTEN 0 128 172.200.1.201:16372 *:*
LISTEN 0 128 *:22 *:*
LISTEN 0 128 [::]:22 [::]:*
[root@redis201.oldboyedu.com ~]#
5.手动停止主库运行，模拟主库宕机
(1)切换前主从状态正常:
[root@redis201.oldboyedu.com ~]# ss -ntl
State Recv-Q Send-Q Local Address:Port Peer Address:Port
LISTEN 0 128 127.0.0.1:16370 *:*
LISTEN 0 128 172.200.1.201:16370 *:*
LISTEN 0 128 127.0.0.1:16371 *:*
LISTEN 0 128 172.200.1.201:16371 *:*
LISTEN 0 128 127.0.0.1:16372 *:*
LISTEN 0 128 172.200.1.201:16372 *:*
LISTEN 0 128 *:22 *:*
LISTEN 0 128 127.0.0.1:16379 *:*
LISTEN 0 128 172.200.1.201:16379 *:*
LISTEN 0 128 [::]:22 [::]:*
[root@redis201.oldboyedu.com ~]#
[root@redis201.oldboyedu.com ~]# redis-cli -p 16370 -a oldboyedu INFO REPLICATION
# Replication
role:master
connected_slaves:2
slave0:ip=172.200.1.201,port=16371,state=online,offset=203023,lag=1
slave1:ip=172.200.1.201,port=16372,state=online,offset=202881,lag=1
master_repl_offset:203023
repl_backlog_active:1
repl_backlog_size:1048576
repl_backlog_first_byte_offset:2
repl_backlog_histlen:203022
[root@redis201.oldboyedu.com ~]#
[root@redis201.oldboyedu.com ~]# redis-cli -p 16371 -a oldboyedu INFO REPLICATION
# Replication
role:slave
master_host:172.200.1.201
master_port:16370
master_link_status:up
master_last_io_seconds_ago:1
master_sync_in_progress:0
slave_repl_offset:379
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
[root@redis201.oldboyedu.com ~]#
[root@redis201.oldboyedu.com ~]# redis-cli -p 16372 -a oldboyedu INFO REPLICATION
# Replication
role:slave
master_host:172.200.1.201
master_port:16370
master_link_status:up
master_last_io_seconds_ago:4
master_sync_in_progress:0
slave_repl_offset:379
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
[root@redis201.oldboyedu.com ~]#
(2)手动将主库宕机，观察从库被sentinel实现了自动切换:
[root@redis201.oldboyedu.com ~]# redis-cli -p 16370 -a oldboyedu SHUTDOWN
[root@redis201.oldboyedu.com ~]#
[root@redis201.oldboyedu.com ~]# redis-cli -p 16370 -a oldboyedu INFO REPLICATION # 由于我们将主库服务停了，因此无法再次连接主库了
Could not connect to Redis at 127.0.0.1:16370: Connection refused
[root@redis201.oldboyedu.com ~]#
[root@redis201.oldboyedu.com ~]# redis-cli -p 16371 -a oldboyedu INFO REPLICATION
# Replication
role:slave
master_host:172.200.1.201
master_port:16372
master_link_status:up
master_last_io_seconds_ago:2
master_sync_in_progress:0
slave_repl_offset:748
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
[root@redis201.oldboyedu.com ~]#
[root@redis201.oldboyedu.com ~]# redis-cli -p 16372 -a oldboyedu INFO REPLICATION
# Replication
role:master
connected_slaves:1
slave0:ip=172.200.1.201,port=16371,state=online,offset=1032,lag=0
master_repl_offset:1046
repl_backlog_active:1
repl_backlog_size:1048576
repl_backlog_first_byte_offset:2
repl_backlog_histlen:1045
[root@redis201.oldboyedu.com ~]#
(3)查看sentinel的日志信息，不难发现有记录切换的过程，如下所示:
[root@redis201.oldboyedu.com ~]# tail -100f /oldboyedu/logs/redis16379/sentinel.log
...
14823:X 04 Mar 07:35:30.162 # +new-epoch 1
14823:X 04 Mar 07:35:30.162 # +try-failover master oldboyedu_master 172.200.1.201 16370
14823:X 04 Mar 07:35:30.164 # +vote-for-leader 13c5048514a67a5a66b0792cb3b64a3a85e10b4b 1
14823:X 04 Mar 07:35:30.164 # +elected-leader master oldboyedu_master 172.200.1.201 16370
14823:X 04 Mar 07:35:30.164 # +failover-state-select-slave master oldboyedu_master 172.200.1.201 16370
14823:X 04 Mar 07:35:30.236 # +selected-slave slave 172.200.1.201:16372 172.200.1.201 16372 @ oldboyedu_master 172.200.1.201 16370
14823:X 04 Mar 07:35:30.236 * +failover-state-send-slaveof-noone slave 172.200.1.201:16372 172.200.1.201 16372 @ oldboyedu_master 172.200.1.201 16370
14823:X 04 Mar 07:35:30.327 * +failover-state-wait-promotion slave 172.200.1.201:16372 172.200.1.201 16372 @ oldboyedu_master 172.200.1.201 16370
14823:X 04 Mar 07:35:31.170 # +promoted-slave slave 172.200.1.201:16372 172.200.1.201 16372 @ oldboyedu_master 172.200.1.201 16370
14823:X 04 Mar 07:35:31.170 # +failover-state-reconf-slaves master oldboyedu_master 172.200.1.201 16370
14823:X 04 Mar 07:35:31.240 * +slave-reconf-sent slave 172.200.1.201:16371 172.200.1.201 16371 @ oldboyedu_master 172.200.1.201 16370
14823:X 04 Mar 07:35:32.190 * +slave-reconf-inprog slave 172.200.1.201:16371 172.200.1.201 16371 @ oldboyedu_master 172.200.1.201 16370
14823:X 04 Mar 07:35:32.190 * +slave-reconf-done slave 172.200.1.201:16371 172.200.1.201 16371 @ oldboyedu_master 172.200.1.201 16370
14823:X 04 Mar 07:35:32.256 # +failover-end master oldboyedu_master 172.200.1.201 16370
14823:X 04 Mar 07:35:32.256 # +switch-master oldboyedu_master 172.200.1.201 16370 172.200.1.201 16372
14823:X 04 Mar 07:35:32.256 * +slave slave 172.200.1.201:16371 172.200.1.201 16371 @ oldboyedu_master 172.200.1.201 16372
14823:X 04 Mar 07:35:32.256 * +slave slave 172.200.1.201:16370 172.200.1.201 16370 @ oldboyedu_master 172.200.1.201 16372
14823:X 04 Mar 07:35:37.293 # +sdown slave 172.200.1.201:16370 172.200.1.201 16370 @ oldboyedu_master 172.200.1.201 16372
6.手动修复宕机的主库，sentinel会自动发现
(1)修复之前，16370端口是连接不上的:
[root@redis201.oldboyedu.com ~]# redis-cli -p 16370 -a oldboyedu INFO REPLICATION
Could not connect to Redis at 127.0.0.1:16370: Connection refused
[root@redis201.oldboyedu.com ~]#
[root@redis201.oldboyedu.com ~]# redis-cli -p 16371 -a oldboyedu INFO REPLICATION
# Replication
role:slave
master_host:172.200.1.201
master_port:16372
master_link_status:up
master_last_io_seconds_ago:2
master_sync_in_progress:0
slave_repl_offset:28558
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
[root@redis201.oldboyedu.com ~]#
[root@redis201.oldboyedu.com ~]# redis-cli -p 16372 -a oldboyedu INFO REPLICATION
# Replication
role:master
connected_slaves:1
slave0:ip=172.200.1.201,port=16371,state=online,offset=28856,lag=1
master_repl_offset:28856
repl_backlog_active:1
repl_backlog_size:1048576
repl_backlog_first_byte_offset:2
repl_backlog_histlen:28855
[root@redis201.oldboyedu.com ~]#
(2)修复之后，之前宕机的主库会自动加入集群并成为slaves节点:
[root@redis201.oldboyedu.com ~]# redis-server /oldboyedu/softwares/redis16370/redis.conf
[root@redis201.oldboyedu.com ~]#
[root@redis201.oldboyedu.com ~]# redis-cli -p 16370 -a oldboyedu INFO REPLICATION
# Replication
role:slave
master_host:172.200.1.201
master_port:16372
master_link_status:up
master_last_io_seconds_ago:1
master_sync_in_progress:0
slave_repl_offset:38731
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
[root@redis201.oldboyedu.com ~]#
[root@redis201.oldboyedu.com ~]# redis-cli -p 16371 -a oldboyedu INFO REPLICATION
# Replication
role:slave
master_host:172.200.1.201
master_port:16372
master_link_status:up
master_last_io_seconds_ago:0
master_sync_in_progress:0
slave_repl_offset:39015
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
[root@redis201.oldboyedu.com ~]#
[root@redis201.oldboyedu.com ~]# redis-cli -p 16372 -a oldboyedu INFO REPLICATION
# Replication
role:master
connected_slaves:2
slave0:ip=172.200.1.201,port=16371,state=online,offset=39157,lag=1
slave1:ip=172.200.1.201,port=16370,state=online,offset=39157,lag=1
master_repl_offset:39157
repl_backlog_active:1
repl_backlog_size:1048576
repl_backlog_first_byte_offset:2
repl_backlog_histlen:39156
[root@redis201.oldboyedu.com ~]#
(3)注意观察sentinel日志哟，会有新日志记录:
[root@redis201.oldboyedu.com ~]# tail -100f /oldboyedu/logs/redis16379/sentinel.log
...
14823:X 04 Mar 07:42:37.709 # -sdown slave 172.200.1.201:16370 172.200.1.201 16370 @ oldboyedu_master 172.200.1.201 16372
14823:X 04 Mar 07:42:47.723 * +convert-to-slave slave 172.200.1.201:16370 172.200.1.201 16370 @ oldboyedu_master 172.200.1.201 16372
温馨提示:
除了提到上面的变化外，还有一点变化就是配置文件也修改了，每个"slave"节点都新增了以下两行:(本案例只有"redis16370"和"redis16371"这两个实例的配置文件是slave节点，因此这两个节点中是存在的!)
[root@redis201.oldboyedu.com ~]# tail -2 /oldboyedu/softwares/redis16370/redis.conf
# Generated by CONFIG REWRITE
slaveof 172.200.1.201 16372
[root@redis201.oldboyedu.com ~]#
[root@redis201.oldboyedu.com ~]# tail -2 /oldboyedu/softwares/redis16371/redis.conf
# Generated by CONFIG REWRITE
slaveof 172.200.1.201 16372
[root@redis201.oldboyedu.com ~]#
三.Sentinel管理命令
1.使用"PING"指令监测Redis或者sentinel实例是否正常工作，如果返回"PONG"说明是可以正常相应的!
[root@redis201.oldboyedu.com ~]# redis-cli -p 16379
127.0.0.1:16379> PING
PONG
127.0.0.1:16379>
127.0.0.1:16379> QUIT
[root@redis201.oldboyedu.com ~]#
2.使用"SENTINEL MASTERS"指令列出所有被监视的主服务器
[root@redis201.oldboyedu.com ~]# redis-cli -p 16379
127.0.0.1:16379> SENTINEL MASTERS
1) 1) "name"
2) "oldboyedu_master"
3) "ip"
4) "172.200.1.201"
5) "port"
6) "16372"
7) "runid"
8) "197c7c9da1f90db94becbad0c4a06f7ebfdc58f0"
9) "flags"
10) "master"
11) "link-pending-commands"
12) "0"
13) "link-refcount"
14) "1"
15) "last-ping-sent"
16) "0"
17) "last-ok-ping-reply"
18) "841"
19) "last-ping-reply"
20) "841"
21) "down-after-milliseconds"
22) "5000"
23) "info-refresh"
24) "3770"
25) "role-reported"
26) "master"
27) "role-reported-time"
28) "1460019"
29) "config-epoch"
30) "1"
31) "num-slaves"
32) "2"
33) "num-other-sentinels"
34) "0"
35) "quorum"
36) "1"
37) "failover-timeout"
38) "180000"
39) "parallel-syncs"
40) "1"
127.0.0.1:16379> QUIT
[root@redis201.oldboyedu.com ~]#
3.使用"SENTINEL slaves "指令列出所有被监视的从服务器
[root@redis201.oldboyedu.com ~]# redis-cli -p 16379
127.0.0.1:16379> SENTINEL SLAVES oldboyedu_master
1) 1) "name"
2) "172.200.1.201:16370"
3) "ip"
4) "172.200.1.201"
5) "port"
6) "16370"
7) "runid"
8) "806728fde23bba8950da1b4dc6efde83ac9a5d10"
9) "flags"
10) "slave"
11) "link-pending-commands"
12) "0"
13) "link-refcount"
14) "1"
15) "last-ping-sent"
16) "0"
17) "last-ok-ping-reply"
18) "424"
19) "last-ping-reply"
20) "424"
21) "down-after-milliseconds"
22) "5000"
23) "info-refresh"
24) "5726"
25) "role-reported"
26) "slave"
27) "role-reported-time"
28) "1100125"
29) "master-link-down-time"
30) "0"
31) "master-link-status"
32) "ok"
33) "master-host"
34) "172.200.1.201"
35) "master-port"
36) "16372"
37) "slave-priority"
38) "100"
39) "slave-repl-offset"
40) "108121"
2) 1) "name"
2) "172.200.1.201:16371"
3) "ip"
4) "172.200.1.201"
5) "port"
6) "16371"
7) "runid"
8) "cae263ec4a59df2a099a69be088748fe64b2f0fa"
9) "flags"
10) "slave"
11) "link-pending-commands"
12) "0"
13) "link-refcount"
14) "1"
15) "last-ping-sent"
16) "0"
17) "last-ok-ping-reply"
18) "334"
19) "last-ping-reply"
20) "334"
21) "down-after-milliseconds"
22) "5000"
23) "info-refresh"
24) "9472"
25) "role-reported"
26) "slave"
27) "role-reported-time"
28) "1545671"
29) "master-link-down-time"
30) "0"
31) "master-link-status"
32) "ok"
33) "master-host"
34) "172.200.1.201"
35) "master-port"
36) "16372"
37) "slave-priority"
38) "100"
39) "slave-repl-offset"
40) "107979"
127.0.0.1:16379>
127.0.0.1:16379> QUIT
[root@redis201.oldboyedu.com ~]#
4.使用"SENTINEL get-master-addr-by-name "指令返回给定名字的主服务器的IP地址和端口号
[root@redis201.oldboyedu.com ~]# redis-cli -p 16379
127.0.0.1:16379> SENTINEL get-master-addr-by-name oldboyedu_master
1) "172.200.1.201"
2) "16372"
127.0.0.1:16379>
127.0.0.1:16379> QUIT
[root@redis201.oldboyedu.com ~]#
5.使用"SENTINEL reset "指令重置所有名字和给定模式"pattern"相匹配的主服务器
[root@redis201.oldboyedu.com ~]# redis-cli -p 16379
127.0.0.1:16379> SENTINEL RESET my*
(integer) 1
127.0.0.1:16379>
127.0.0.1:16379> QUIT
[root@redis201.oldboyedu.com ~]#
[root@redis201.oldboyedu.com ~]# tail -100f /oldboyedu/logs/redis16379/sentinel.log
...
14823:X 04 Mar 08:14:21.191 # +reset-master master oldboyedu_master 172.200.1.201 16372
14823:X 04 Mar 08:14:21.672 * +slave slave 172.200.1.201:16371 172.200.1.201 16371 @ oldboyedu_master 172.200.1.201 16372
14823:X 04 Mar 08:14:21.675 * +slave slave 172.200.1.201:16370 172.200.1.201 16370 @ oldboyedu_master 172.200.1.201 16372
...
6.当主服务器失效时，使用"SENTINEL failover "指令在不询问其他Sentinel意见的情况下， 强制开始一次自动故障迁移。
(1)手动"failover"操作前:
[root@redis201.oldboyedu.com ~]# redis-cli -p 16370 -a oldboyedu INFO REPLICATION
# Replication
role:slave
master_host:172.200.1.201
master_port:16372
master_link_status:up
master_last_io_seconds_ago:1
master_sync_in_progress:0
slave_repl_offset:120585
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
[root@redis201.oldboyedu.com ~]#
[root@redis201.oldboyedu.com ~]# redis-cli -p 16371 -a oldboyedu INFO REPLICATION
# Replication
role:slave
master_host:172.200.1.201
master_port:16372
master_link_status:up
master_last_io_seconds_ago:2
master_sync_in_progress:0
slave_repl_offset:120869
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
[root@redis201.oldboyedu.com ~]#
[root@redis201.oldboyedu.com ~]# redis-cli -p 16372 -a oldboyedu INFO REPLICATION
# Replication
role:master
connected_slaves:2
slave0:ip=172.200.1.201,port=16371,state=online,offset=121153,lag=0
slave1:ip=172.200.1.201,port=16370,state=online,offset=121153,lag=0
master_repl_offset:121153
repl_backlog_active:1
repl_backlog_size:1048576
repl_backlog_first_byte_offset:2
repl_backlog_histlen:121152
[root@redis201.oldboyedu.com ~]#
(2)手动触发故障转移
[root@redis201.oldboyedu.com ~]# redis-cli -p 16379
127.0.0.1:16379> SENTINEL failover oldboyedu_master
OK
127.0.0.1:16379>
127.0.0.1:16379> QUIT
[root@redis201.oldboyedu.com ~]#
[root@redis201.oldboyedu.com ~]# tail -100f /oldboyedu/logs/redis16379/sentinel.log # 注意观察日志的变化
...
14823:X 04 Mar 08:06:16.688 # Executing user requested FAILOVER of 'oldboyedu_master'
14823:X 04 Mar 08:06:16.688 # +new-epoch 2
14823:X 04 Mar 08:06:16.688 # +try-failover master oldboyedu_master 172.200.1.201 16372
14823:X 04 Mar 08:06:16.697 # +vote-for-leader 13c5048514a67a5a66b0792cb3b64a3a85e10b4b 2
14823:X 04 Mar 08:06:16.697 # +elected-leader master oldboyedu_master 172.200.1.201 16372
14823:X 04 Mar 08:06:16.697 # +failover-state-select-slave master oldboyedu_master 172.200.1.201 16372
14823:X 04 Mar 08:06:16.753 # +selected-slave slave 172.200.1.201:16370 172.200.1.201 16370 @ oldboyedu_master 172.200.1.201 16372
14823:X 04 Mar 08:06:16.753 * +failover-state-send-slaveof-noone slave 172.200.1.201:16370 172.200.1.201 16370 @ oldboyedu_master 172.200.1.201 16372
14823:X 04 Mar 08:06:16.820 * +failover-state-wait-promotion slave 172.200.1.201:16370 172.200.1.201 16370 @ oldboyedu_master 172.200.1.201 16372
14823:X 04 Mar 08:06:17.740 # +promoted-slave slave 172.200.1.201:16370 172.200.1.201 16370 @ oldboyedu_master 172.200.1.201 16372
14823:X 04 Mar 08:06:17.740 # +failover-state-reconf-slaves master oldboyedu_master 172.200.1.201 16372
14823:X 04 Mar 08:06:17.800 * +slave-reconf-sent slave 172.200.1.201:16371 172.200.1.201 16371 @ oldboyedu_master 172.200.1.201 16372
14823:X 04 Mar 08:06:18.798 * +slave-reconf-inprog slave 172.200.1.201:16371 172.200.1.201 16371 @ oldboyedu_master 172.200.1.201 16372
14823:X 04 Mar 08:06:18.798 * +slave-reconf-done slave 172.200.1.201:16371 172.200.1.201 16371 @ oldboyedu_master 172.200.1.201 16372
14823:X 04 Mar 08:06:18.882 # +failover-end master oldboyedu_master 172.200.1.201 16372
14823:X 04 Mar 08:06:18.882 # +switch-master oldboyedu_master 172.200.1.201 16372 172.200.1.201 16370
14823:X 04 Mar 08:06:18.882 * +slave slave 172.200.1.201:16371 172.200.1.201 16371 @ oldboyedu_master 172.200.1.201 16370
14823:X 04 Mar 08:06:18.882 * +slave slave 172.200.1.201:16372 172.200.1.201 16372 @ oldboyedu_master 172.200.1.201 16370
14823:X 04 Mar 08:06:29.000 * +convert-to-slave slave 172.200.1.201:16372 172.200.1.201 16372 @ oldboyedu_master 172.200.1.201 16370
...
(3)手动"failover"操作后:
[root@redis201.oldboyedu.com ~]# redis-cli -p 16370 -a oldboyedu INFO REPLICATION
# Replication
role:master
connected_slaves:2
slave0:ip=172.200.1.201,port=16371,state=online,offset=135098,lag=1
slave1:ip=172.200.1.201,port=16372,state=online,offset=135098,lag=0
master_repl_offset:135098
repl_backlog_active:1
repl_backlog_size:1048576
repl_backlog_first_byte_offset:129559
repl_backlog_histlen:5540
[root@redis201.oldboyedu.com ~]#
[root@redis201.oldboyedu.com ~]# redis-cli -p 16371 -a oldboyedu INFO REPLICATION
# Replication
role:slave
master_host:172.200.1.201
master_port:16370
master_link_status:up
master_last_io_seconds_ago:1
master_sync_in_progress:0
slave_repl_offset:135396
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
[root@redis201.oldboyedu.com ~]#
[root@redis201.oldboyedu.com ~]# redis-cli -p 16372 -a oldboyedu INFO REPLICATION
# Replication
role:slave
master_host:172.200.1.201
master_port:16370
master_link_status:up
master_last_io_seconds_ago:1
master_sync_in_progress:0
slave_repl_offset:135538
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:2
repl_backlog_histlen:130295
[root@redis201.oldboyedu.com ~]# 四.作业
(1)创建sentinel的配置文件
install -d /oldboyedu/{data,softwares}/sentinel
cat > /oldboyedu/softwares/sentinel/sentinel.conf <<EOF
bind 10.0.0.106
port 18106
dir /oldboyedu/data/sentinel
sentinel monitor oldboyedu_master 10.0.0.108 8108 1
sentinel down-after-milliseconds oldboyedu_master 5000
sentinel auth-pass oldboyedu_master oldboyedu
EOF
(2)前台启动sentinel
redis-sentinel /oldboyedu/softwares/sentinel/sentinel.conf
作业：
(1)完成课上的所有练习，包括但不限于字符串，列表，字典，集合，有序集合，主从复制，单哨兵模式;
(2)实现单哨兵的VIP飘逸功能(扩展题);