redis實(shí)現(xiàn)sentinel哨兵架構(gòu)的方法
更新時(shí)間：2022年11月12日 08:57:01 作者：gy001
哨兵是一個(gè)分布式系統(tǒng)，可以在一個(gè)架構(gòu)中運(yùn)行多個(gè)哨兵(sentinel) 進(jìn)程，這些進(jìn)程使用流言協(xié)議(gossip protocols)來(lái)接收關(guān)于Master主服務(wù)器是否下線的信息，這篇文章主要介紹了redis實(shí)現(xiàn)sentinel哨兵架構(gòu),需要的朋友可以參考下
1、redis哨兵（Sentinel）
1.1、redis集群介紹
前面文章講的主從復(fù)制集群是無(wú)法實(shí)現(xiàn)master和slave角色的自動(dòng)切換的，如果master節(jié)點(diǎn)出現(xiàn)現(xiàn)redis服務(wù)異常、主機(jī)斷電、磁盤損壞等問(wèn)題導(dǎo)致master無(wú)法使用，而redis主從復(fù)制無(wú)法實(shí)現(xiàn)自動(dòng)的故障轉(zhuǎn)移(將slave 自動(dòng)提升為新master)，需要手動(dòng)修改環(huán)境配置,才能切換到slave redis服務(wù)器，另外當(dāng)單臺(tái)Redis服務(wù)器性能無(wú)法滿足業(yè)務(wù)寫入需求的時(shí)候,也無(wú)法橫向擴(kuò)展Redis服務(wù)的并行寫入性能。
需要解決以上的兩個(gè)核心問(wèn)題：
- master和slave角色的無(wú)縫切換，讓業(yè)務(wù)無(wú)感知從而不影響業(yè)務(wù)使用；
- 可橫向動(dòng)態(tài)擴(kuò)展Redis服務(wù)器，從而實(shí)現(xiàn)多臺(tái)服務(wù)器并行寫入以實(shí)現(xiàn)更高并發(fā)的目的。
Redis集群實(shí)現(xiàn)的方式：
- 客戶端分片: 由應(yīng)用決定將不同的KEY發(fā)送到不同的Redis服務(wù)器
- 代理分片: 由代理決定將不同的KEY發(fā)送到不同的Redis服務(wù)器,代理程序如:codis,twemproxy等
- Redis Cluster
1.2、redis哨兵（Sentinel）的工作原理
Sentinel可以管理多個(gè)redis主從集群
Sentinel 進(jìn)程是用于監(jiān)控redis集群中Master主服務(wù)器工作的狀態(tài)，在Master主服務(wù)器發(fā)生故障的時(shí)候，可以實(shí)現(xiàn)Master和Slave服務(wù)器的切換，保證系統(tǒng)的高可用，此功能在redis2.6+的版本已引用，Redis的哨兵模式到了2.8版本之后就穩(wěn)定了下來(lái)。一般在生產(chǎn)環(huán)境也建議使用Redis的2.8版本的以后版本
哨兵(Sentinel) 是一個(gè)分布式系統(tǒng)，可以在一個(gè)架構(gòu)中運(yùn)行多個(gè)哨兵(sentinel) 進(jìn)程，這些進(jìn)程使用流言協(xié)議(gossip protocols)來(lái)接收關(guān)于Master主服務(wù)器是否下線的信息，并使用投票協(xié)議(Agreement Protocols)來(lái)決定是否執(zhí)行自動(dòng)故障遷移,以及選擇哪個(gè)Slave作為新的Master
每個(gè)哨兵(Sentinel)進(jìn)程會(huì)向其它哨兵(Sentinel)、Master、Slave定時(shí)發(fā)送消息，以確認(rèn)對(duì)方是否”活”著，如果發(fā)現(xiàn)對(duì)方在指定配置時(shí)間(此項(xiàng)可配置)內(nèi)未得到回應(yīng)，則暫時(shí)認(rèn)為對(duì)方已離線，也就是所謂的”主觀認(rèn)為宕機(jī)” (主觀:是每個(gè)成員都具有的獨(dú)自的而且可能相同也可能不同的意識(shí))，英文名稱：Subjective Down，簡(jiǎn)稱SDOWN
有主觀宕機(jī)，對(duì)應(yīng)的有客觀宕機(jī)。當(dāng)“哨兵群”中的多數(shù)Sentinel進(jìn)程在對(duì)Master主服務(wù)器做出SDOWN 的判斷，并且通過(guò) SENTINEL is-master-down-by-addr 命令互相交流之后，得出的Master Server下線判斷，這種方式就是“客觀宕機(jī)”(客觀:是不依賴于某種意識(shí)而已經(jīng)實(shí)際存在的一切事物)，英文名稱是：Objectively Down， 簡(jiǎn)稱 ODOWN
通過(guò)一定的vote算法，從剩下的slave從服務(wù)器節(jié)點(diǎn)中，選一臺(tái)提升為Master服務(wù)器節(jié)點(diǎn)，然后自動(dòng)修改相關(guān)配置，并開啟故障轉(zhuǎn)移（failover）
Sentinel 機(jī)制可以解決master和slave角色的自動(dòng)切換問(wèn)題，但單個(gè) Master 的性能瓶頸問(wèn)題無(wú)法解決,類似于MySQL中的MHA功能
Redis Sentinel中的Sentinel節(jié)點(diǎn)個(gè)數(shù)應(yīng)該為大于等于3且最好為奇數(shù)
客戶端初始化時(shí)連接的是Sentinel節(jié)點(diǎn)集合，不再是具體的Redis節(jié)點(diǎn)，但Sentinel只是配置中心不是代理。
Redis Sentinel節(jié)點(diǎn)與普通redis沒有區(qū)別,要實(shí)現(xiàn)讀寫分離依賴于客戶端程序
redis 3.0之前版本中,生產(chǎn)環(huán)境一般使用哨兵模式,3.0后推出redis cluster功能,可以支持更大規(guī)模的生產(chǎn)環(huán)境sentinel中的三個(gè)定時(shí)任務(wù)：
1.2.1sentinel中的三個(gè)定時(shí)任務(wù)：
- 每10秒每個(gè)sentinel對(duì)master和slave執(zhí)行info，發(fā)現(xiàn)slave節(jié)點(diǎn)，確認(rèn)主從關(guān)系。
- 每2秒每個(gè)sentinel通過(guò)master節(jié)點(diǎn)的channel交換信息(pub/sub)通過(guò)sentinel__:hello頻道交互，交互對(duì)節(jié)點(diǎn)的“看法”和自身信息
- 每1秒每個(gè)sentinel對(duì)其他sentinel和redis執(zhí)行ping
1.3、實(shí)現(xiàn)哨兵
環(huán)境準(zhǔn)備：
準(zhǔn)備三臺(tái)主機(jī)搭建主從集群，再在每個(gè)機(jī)器上搭建sentinel
| IP | redis版本 | 主機(jī)名 |
| 10.0.0.101 | redis-5.0.14 | node1 |
| 10.0.0.102 | redis-5.0.14 | node2 |
| 10.0.0.103 | redis-5.0.14 | node3 |
1.3.1、實(shí)現(xiàn)哨兵需要先實(shí)現(xiàn)一下主從復(fù)制的架構(gòu)
哨兵的前提是已經(jīng)實(shí)現(xiàn)了一個(gè)redis的主從復(fù)制的運(yùn)行環(huán)境，從而實(shí)現(xiàn)一個(gè)一主兩從基于哨兵的高可用redis架構(gòu)
注意: master的配置文件中masterauth和slave都必須相同
主從復(fù)制上篇文章已經(jīng)實(shí)現(xiàn)了，這里繼續(xù)用上一篇的環(huán)境
#主節(jié)點(diǎn) [root@master etc]#redis-cli -a 123456 Warning: Using a password with '-a' or '-u' option on the command line interface may not be safe. 127.0.0.1:6379> info replication # Replication role:master connected_slaves:2 slave0:ip=10.0.0.102,port=6379,state=online,offset=14,lag=1 slave1:ip=10.0.0.103,port=6379,state=online,offset=14,lag=0 master_replid:e9310fdb8dd9f91d265d4c9a8621a6879e0262ff master_replid2:0000000000000000000000000000000000000000 master_repl_offset:14 second_repl_offset:-1 repl_backlog_active:1 repl_backlog_size:1048576 repl_backlog_first_byte_offset:1 repl_backlog_histlen:14 127.0.0.1:6379> #slave1節(jié)點(diǎn) [root@slave1 ~]#redis-cli -a 123456 Warning: Using a password with '-a' or '-u' option on the command line interface may not be safe. 127.0.0.1:6379> info replication # Replication role:slave master_host:10.0.0.101 master_port:6379 master_link_status:up master_last_io_seconds_ago:8 master_sync_in_progress:0 slave_repl_offset:224 slave_priority:100 slave_read_only:1 connected_slaves:0 master_replid:e9310fdb8dd9f91d265d4c9a8621a6879e0262ff master_replid2:0000000000000000000000000000000000000000 master_repl_offset:224 second_repl_offset:-1 repl_backlog_active:1 repl_backlog_size:1048576 repl_backlog_first_byte_offset:1 repl_backlog_histlen:224 #slave2節(jié)點(diǎn) [root@slave2 ~]#redis-cli -a 123456 Warning: Using a password with '-a' or '-u' option on the command line interface may not be safe. 127.0.0.1:6379> info replication # Replication role:slave master_host:10.0.0.101 master_port:6379 master_link_status:up master_last_io_seconds_ago:7 master_sync_in_progress:0 slave_repl_offset:2016 slave_priority:100 slave_read_only:1 connected_slaves:0 master_replid:ee9d24063f9d79d698e875634517923d6a9c2a10 master_replid2:0000000000000000000000000000000000000000 master_repl_offset:2016 second_repl_offset:-1 repl_backlog_active:1 repl_backlog_size:1048576 repl_backlog_first_byte_offset:1 repl_backlog_histlen:2016
1.3.2、編輯哨兵的配置文件
Sentinel實(shí)際上是一個(gè)特殊的redis服務(wù)器,有些redis指令支持,但很多指令并不支持.默認(rèn)監(jiān)聽在26379/tcp端口.所有redis節(jié)點(diǎn)使用相同的以下的配置文件.
#這里因?yàn)槲沂窃创a編譯安裝的，所以要人源碼的路徑下找到sentinel.conf將其復(fù)制到redis安裝的目錄下 [root@master]# cd /usr/local/src/redis-5.0.14[root@master /usr/local/src/redis-5.0.14# ls 00-RELEASENOTES CONTRIBUTING INSTALL README.md runtest-cluster sentinel.conf TLS.md BUGS COPYING Makefile redis.conf runtest-moduleapi src utils CONDUCT deps MANIFESTO runtest runtest-sentinel tests [root@master etc]#sed -i "/^#/d" redis-sentinel.conf [root@master etc]#sed -i "/^$/d" redis-sentinel.conf [root@master etc]#vim redis-sentinel.con bind 0.0.0.0 port 26379 daemonize no pidfile /apps/redis/run/redis-sentinel.pid logfile /apps/redis/log/sentinel.log dir /tmp sentinel monitor mymaster 10.0.0.100 6379 2 #mymaster是集群的名稱，此行指定當(dāng)前mymaster集群中master服務(wù)器的地址和端口 #2為法定人數(shù)限制(quorum)，即有幾個(gè)sentinel認(rèn)為master down了就進(jìn)行故障轉(zhuǎn)移，一般此值是所有sentinel節(jié)點(diǎn)(一般總數(shù)是>=3的 奇數(shù),如:3,5,7等)的一半以上的整數(shù)值，比如，總數(shù)是3，即3/2=1.5，取整為2,是master的ODOWN客觀下線的依據(jù) sentinel auth-pass mymaster wm521314 #mymaster集群中master的密碼，注意此行要在上面行的下面 sentinel down-after-milliseconds mymaster 3000 #(SDOWN)判斷mymaster集群中所有節(jié)點(diǎn)的主觀下線的時(shí)間，單位：毫秒，建議3000 sentinel parallel-syncs mymaster 1 #發(fā)生故障轉(zhuǎn)移后，可以同時(shí)向新master同步數(shù)據(jù)的slave的數(shù)量，數(shù)字越小總同步時(shí)間越長(zhǎng)，但可以減輕新master的負(fù)載壓力 sentinel failover-timeout mymaster 180000 #所有slaves指向新的master所需的超時(shí)時(shí)間，單位：毫秒 sentinel deny-scripts-reconfig yes #禁止修改腳本 [root@master etc]# scp /apps/redis/etc/redis-sentinel.conf 10.0.0.102:/apps/redis/etc/[root@master etc]# scp /apps/redis/etc/redis-sentinel.conf 10.0.0.103:/apps/redis/etc/
三個(gè)哨兵服務(wù)器的配置都如下：
[root@slave2 etc]#cat redis-sentinel.conf port 26379 daemonize no pidfile /apps/redis/run/redis-sentinel.pid logfile /apps/redis/log/sentinel.log dir /tmp sentinel monitor mymaster 10.0.0.101 6379 2 sentinel down-after-milliseconds mymaster 30000 sentinel parallel-syncs mymaster 1 sentinel failover-timeout mymaster 180000 sentinel deny-scripts-reconfig yes
1.3.2、啟動(dòng)哨兵
#我這里是編譯安裝，所以在所有節(jié)點(diǎn)生成新的service文件 [root@master ~]#vim /lib/systemd/system/redis-sentinel.service [Unit] Description=Redis Sentinel After=network.target [Service] ExecStart=/apps/redis/bin/redis-sentinel /apps/redis/etc/redis-sentinel.conf --supervised systemd ExecStop=/bin/kill -s QUIT $MAINPID User=redis Group=redis RuntimeDirectory=redis RuntimeDirectoryMode=0755 [Install] WantedBy=multi-user.target [root@master ~]#scp /lib/systemd/system/redis-sentinel.service 10.0.0.102:/lib/systemd/system/ [root@master ~]#scp /lib/systemd/system/redis-sentinel.service 10.0.0.103:/lib/systemd/system/ #在所有節(jié)點(diǎn)修改一下文件的所有者所屬組，不然無(wú)法啟動(dòng) chown -R redis.redis /apps/redis/ #在所有節(jié)點(diǎn)啟動(dòng)哨兵 [root@master ~]#systemctl daemon-reload [root@master ~]#systemctl enable --now redis-sentinel.service
1.3.3、驗(yàn)證哨兵端口
[root@master ~]#ss -nlt State Recv-Q Send-Q Local Address:Port Peer Address:Port Process LISTEN 0 511 0.0.0.0:26379 0.0.0.0:* #已經(jīng)啟動(dòng)，另外兩個(gè)節(jié)點(diǎn)同樣查看 LISTEN 0 511 0.0.0.0:6379 0.0.0.0:* LISTEN 0 1024 127.0.0.53%lo:53 0.0.0.0:* LISTEN 0 128 0.0.0.0:22 0.0.0.0:* LISTEN 0 128 127.0.0.1:6010 0.0.0.0:* LISTEN 0 511 [::]:26379 [::]:* LISTEN 0 128 [::]:22 [::]:* LISTEN 0 128 [::1]:6010 [::]:*
1.3.4、查看哨兵日志
[root@master ~]#tail -f /apps/redis/log/sentinel.log #查看主節(jié)點(diǎn)日志 37184:X 04 Nov 2022 23:32:07.023 # Configuration loaded 37184:X 04 Nov 2022 23:32:07.023 # systemd supervision requested, but NOTIFY_SOCKET not found 37184:X 04 Nov 2022 23:32:07.024 * Increased maximum number of open files to 10032 (it was originally set to 1024). 37184:X 04 Nov 2022 23:32:07.024 * Running mode=sentinel, port=26379. 37184:X 04 Nov 2022 23:32:07.032 # Sentinel ID is 2f9b5e2cb38fc5b3db19417b9626f0ebcc3a7120 37184:X 04 Nov 2022 23:32:07.032 # +monitor master mymaster 10.0.0.101 6379 quorum 2 37184:X 04 Nov 2022 23:32:07.033 * +slave slave 10.0.0.102:6379 10.0.0.102 6379 @ mymaster 10.0.0.101 6379 37184:X 04 Nov 2022 23:32:07.036 * +slave slave 10.0.0.103:6379 10.0.0.103 6379 @ mymaster 10.0.0.101 6379 37184:X 04 Nov 2022 23:32:07.595 * +sentinel sentinel f505346a3640b4942963b26829e5d99a1fa6691e 10.0.0.102 26379 @ mymaster 10.0.0.101 6379 37184:X 04 Nov 2022 23:32:12.779 * +sentinel sentinel b6aebc495fe5d6e81e57ffd5d4f5bef795b3819a 10.0.0.103 26379 @ mymaster 10.0.0.101 6379 #查看從節(jié)點(diǎn)日志 [root@slave1 ~]#tail -f /apps/redis/log/sentinel.log 6376:X 04 Nov 2022 23:31:59.446 # Configuration loaded 6376:X 04 Nov 2022 23:31:59.446 # systemd supervision requested, but NOTIFY_SOCKET not found 6376:X 04 Nov 2022 23:31:59.446 * Increased maximum number of open files to 10032 (it was originally set to 1024). 6376:X 04 Nov 2022 23:31:59.447 * Running mode=sentinel, port=26379. 6376:X 04 Nov 2022 23:31:59.455 # Sentinel ID is f505346a3640b4942963b26829e5d99a1fa6691e 6376:X 04 Nov 2022 23:31:59.455 # +monitor master mymaster 10.0.0.101 6379 quorum 2 6376:X 04 Nov 2022 23:31:59.457 * +slave slave 10.0.0.102:6379 10.0.0.102 6379 @ mymaster 10.0.0.101 6379 6376:X 04 Nov 2022 23:31:59.458 * +slave slave 10.0.0.103:6379 10.0.0.103 6379 @ mymaster 10.0.0.101 6379 6376:X 04 Nov 2022 23:32:08.989 * +sentinel sentinel 2f9b5e2cb38fc5b3db19417b9626f0ebcc3a7120 10.0.0.101 26379 @ mymaster 10.0.0.101 6379 6376:X 04 Nov 2022 23:32:12.716 * +sentinel sentinel b6aebc495fe5d6e81e57ffd5d4f5bef795b3819a 10.0.0.103 26379 @ mymaster 10.0.0.101 6379
#查看從節(jié)點(diǎn)日志
[root@slave1 ~]#tail -f /apps/redis/log/sentinel.log 6376:X 04 Nov 2022 23:31:59.446 # Configuration loaded 6376:X 04 Nov 2022 23:31:59.446 # systemd supervision requested, but NOTIFY_SOCKET not found 6376:X 04 Nov 2022 23:31:59.446 * Increased maximum number of open files to 10032 (it was originally set to 1024). 6376:X 04 Nov 2022 23:31:59.447 * Running mode=sentinel, port=26379. 6376:X 04 Nov 2022 23:31:59.455 # Sentinel ID is f505346a3640b4942963b26829e5d99a1fa6691e 6376:X 04 Nov 2022 23:31:59.455 # +monitor master mymaster 10.0.0.101 6379 quorum 2 6376:X 04 Nov 2022 23:31:59.457 * +slave slave 10.0.0.102:6379 10.0.0.102 6379 @ mymaster 10.0.0.101 6379 6376:X 04 Nov 2022 23:31:59.458 * +slave slave 10.0.0.103:6379 10.0.0.103 6379 @ mymaster 10.0.0.101 6379 6376:X 04 Nov 2022 23:32:08.989 * +sentinel sentinel 2f9b5e2cb38fc5b3db19417b9626f0ebcc3a7120 10.0.0.101 26379 @ mymaster 10.0.0.101 6379 6376:X 04 Nov 2022 23:32:12.716 * +sentinel sentinel b6aebc495fe5d6e81e57ffd5d4f5bef795b3819a 10.0.0.103 26379 @ mymaster 10.0.0.101 6379
1.3.5、當(dāng)前sentinel狀態(tài)
[root@slave1 ~]#redis-cli -p 26379 127.0.0.1:26379> 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=mymaster,status=ok,address=10.0.0.101:6379,slaves=2,sentinels=3
1.3.6、驗(yàn)證數(shù)據(jù)是否同步
[root@master ~]#redis-cli -a 123456 Warning: Using a password with '-a' or '-u' option on the command line interface may not be safe. 127.0.0.1:6379> set k1 v1 OK 127.0.0.1:6379> set k2 v2 OK [root@slave1 ~]#redis-cli -a 123456 Warning: Using a password with '-a' or '-u' option on the command line interface may not be safe. 127.0.0.1:6379> get k1 "v1" [root@slave2 ~]#redis-cli -a 123456 Warning: Using a password with '-a' or '-u' option on the command line interface may not be safe. 127.0.0.1:6379> get k1 "v1"
1.3.7、制造Redis Master節(jié)點(diǎn)故障故障轉(zhuǎn)移并查看其過(guò)程
#停掉master節(jié)點(diǎn)的redis服務(wù)來(lái)模擬節(jié)點(diǎn)故障 [root@master ~]#killall redis-server #這時(shí)候查看日志發(fā)現(xiàn)主節(jié)點(diǎn)已經(jīng)變成了10.0.0.102機(jī)器 [root@master ~]#tail -f /apps/redis/log/sentinel.log 800:X 06 Nov 2022 14:36:38.502 # -sdown sentinel b6aebc495fe5d6e81e57ffd5d4f5bef795b3819a 10.0.0.103 26379 @ mymaster 10.0.0.101 6379 800:X 06 Nov 2022 14:36:38.604 # -sdown slave 10.0.0.103:6379 10.0.0.103 6379 @ mymaster 10.0.0.101 6379 800:X 06 Nov 2022 15:21:26.435 # +sdown master mymaster 10.0.0.101 6379 800:X 06 Nov 2022 15:21:26.500 # +new-epoch 1 800:X 06 Nov 2022 15:21:26.501 # +vote-for-leader f505346a3640b4942963b26829e5d99a1fa6691e 1 800:X 06 Nov 2022 15:21:27.019 # +config-update-from sentinel f505346a3640b4942963b26829e5d99a1fa6691e 10.0.0.102 26379 @ mymaster 10.0.0.101 6379 800:X 06 Nov 2022 15:21:27.019 # +switch-master mymaster 10.0.0.101 6379 10.0.0.102 6379 800:X 06 Nov 2022 15:21:27.019 * +slave slave 10.0.0.103:6379 10.0.0.103 6379 @ mymaster 10.0.0.102 6379 800:X 06 Nov 2022 15:21:27.019 * +slave slave 10.0.0.101:6379 10.0.0.101 6379 @ mymaster 10.0.0.102 6379 800:X 06 Nov 2022 15:21:30.032 # +sdown slave 10.0.0.101:6379 10.0.0.101 6379 @ mymaster 10.0.0.102 6379 #查看各節(jié)點(diǎn)sentinel信息 [root@master ~]#redis-cli -p 26379 127.0.0.1:26379> 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=mymaster,status=ok,address=10.0.0.102:6379,slaves=2,sentinels=3 #故障轉(zhuǎn)移后的sentinel配置文件會(huì)被自動(dòng)修改 [root@slave1 ~]#grep "^[a-Z]" /apps/redis/etc/redis-sentinel.conf bind 0.0.0.0 port 26379 daemonize no pidfile "/apps/redis/run/redis-sentinel.pid" logfile "/apps/redis/log/sentinel.log" dir "/tmp" sentinel myid f505346a3640b4942963b26829e5d99a1fa6691e sentinel deny-scripts-reconfig yes sentinel monitor mymaster 10.0.0.102 6379 2 #這里發(fā)現(xiàn)主節(jié)點(diǎn)已經(jīng)變成10.0.0.102 sentinel down-after-milliseconds mymaster 3000 sentinel auth-pass mymaster 123456 sentinel config-epoch mymaster 1 protected-mode no supervised systemd sentinel leader-epoch mymaster 1 sentinel known-replica mymaster 10.0.0.101 6379 sentinel known-replica mymaster 10.0.0.103 6379 sentinel known-sentinel mymaster 10.0.0.103 26379 b6aebc495fe5d6e81e57ffd5d4f5bef795b3819a sentinel known-sentinel mymaster 10.0.0.101 26379 2f9b5e2cb38fc5b3db19417b9626f0ebcc3a7120 sentinel current-epoch 1 #進(jìn)入redis查看現(xiàn)在info replication信息 [root@slave1 ~]#redis-cli -a 123456 Warning: Using a password with '-a' or '-u' option on the command line interface may not be safe. 127.0.0.1:6379> info replication # Replication role:master connected_slaves:1 slave0:ip=10.0.0.103,port=6379,state=online,offset=905513,lag=0 master_replid:2824cc99085826d751e848cd7b04c9f4abc398cb master_replid2:488fa34f2c093be4886741a38323dd1a0ecf9e03 master_repl_offset:905648 second_repl_offset:537903 repl_backlog_active:1 repl_backlog_size:1048576 repl_backlog_first_byte_offset:1 repl_backlog_histlen:905648 [root@slave2 ~]#redis-cli -a 123456 Warning: Using a password with '-a' or '-u' option on the command line interface may not be safe . 127.0.0.1:6379> INFO replication # Replication role:slave master_host:10.0.0.102 master_port:6379 master_link_status:up master_last_io_seconds_ago:0 master_sync_in_progress:0 slave_repl_offset:901300 slave_priority:100 slave_read_only:1 connected_slaves:0 master_replid:2824cc99085826d751e848cd7b04c9f4abc398cb master_replid2:488fa34f2c093be4886741a38323dd1a0ecf9e03 master_repl_offset:901300 second_repl_offset:537903 repl_backlog_active:1 repl_backlog_size:1048576 repl_backlog_first_byte_offset:564 repl_backlog_histlen:900737
1.3.8、查看選舉新主的過(guò)程日志監(jiān)測(cè)
[root@slave2 ~]#tail -f /apps/redis/log/sentinel.log 798:X 06 Nov 2022 15:21:26.509 # +sdown master mymaster 10.0.0.101 6379 798:X 06 Nov 2022 15:21:26.534 # +new-epoch 1 798:X 06 Nov 2022 15:21:26.535 # +vote-for-leader f505346a3640b4942963b26829e5d99a1fa6691e 1 798:X 06 Nov 2022 15:21:26.609 # +odown master mymaster 10.0.0.101 6379 #quorum 3/2 798:X 06 Nov 2022 15:21:26.609 # Next failover delay: I will not start a failover before Sun Nov 6 15:27:27 2022 798:X 06 Nov 2022 15:21:27.053 # +config-update-from sentinel f505346a3640b4942963b26829e5d99a1fa6691e 10.0.0.102 26379 @ mymaster 10.0.0.101 6379 798:X 06 Nov 2022 15:21:27.053 # +switch-master mymaster 10.0.0.101 6379 10.0.0.102 6379 798:X 06 Nov 2022 15:21:27.053 * +slave slave 10.0.0.103:6379 10.0.0.103 6379 @ mymaster 10.0.0.102 6379 798:X 06 Nov 2022 15:21:27.053 * +slave slave 10.0.0.101:6379 10.0.0.101 6379 @ mymaster 10.0.0.102 6379 798:X 06 Nov 2022 15:21:30.094 # +sdown slave 10.0.0.101:6379 10.0.0.101 6379 @ mymaster 10.0.0.102 6379
1.3.9、恢復(fù)故障的原master重新加入redis集群
#重新其他原master的redis服務(wù)，會(huì)發(fā)現(xiàn)會(huì)自動(dòng)指向新的master [root@master ~]#systemctl restart redis #在原master上觀察狀態(tài) [root@master ~]#redis-cli -a 123456 Warning: Using a password with '-a' or '-u' option on the command line interface may not be safe. 127.0.0.1:6379> info replication # Replication role:slave master_host:10.0.0.102 master_port:6379 master_link_status:up master_last_io_seconds_ago:1 master_sync_in_progress:0 slave_repl_offset:942522 slave_priority:100 slave_read_only:1 connected_slaves:0 master_replid:2824cc99085826d751e848cd7b04c9f4abc398cb master_replid2:0000000000000000000000000000000000000000 master_repl_offset:942522 second_repl_offset:-1 repl_backlog_active:1 repl_backlog_size:1048576 repl_backlog_first_byte_offset:936383 repl_backlog_histlen:6140 [root@master ~]#redis-cli -p 26379 127.0.0.1:26379> 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=mymaster,status=ok,address=10.0.0.102:6379,slaves=2,sentinels=3 #再回到新master節(jié)點(diǎn)觀察狀態(tài) [root@slave1 ~]#redis-cli -a 123456 Warning: Using a password with '-a' or '-u' option on the command line interface may not be safe. 127.0.0.1:6379> info replication # Replication role:master connected_slaves:2 slave0:ip=10.0.0.103,port=6379,state=online,offset=976915,lag=0 slave1:ip=10.0.0.101,port=6379,state=online,offset=977185,lag=0 master_replid:2824cc99085826d751e848cd7b04c9f4abc398cb master_replid2:488fa34f2c093be4886741a38323dd1a0ecf9e03 master_repl_offset:977185 second_repl_offset:537903 repl_backlog_active:1 repl_backlog_size:1048576 repl_backlog_first_byte_offset:1 repl_backlog_histlen:977185 #在新主查看一下日志 [root@slave1 ~]#tail -f /apps/redis/log/sentinel.log 786:X 06 Nov 2022 15:21:27.030 * +slave-reconf-sent slave 10.0.0.103:6379 10.0.0.103 6379 @ mymaster 10.0.0.101 6379 786:X 06 Nov 2022 15:21:27.580 * +slave-reconf-inprog slave 10.0.0.103:6379 10.0.0.103 6379 @ mymaster 10.0.0.101 6379 786:X 06 Nov 2022 15:21:27.639 # -odown master mymaster 10.0.0.101 6379 786:X 06 Nov 2022 15:21:28.621 * +slave-reconf-done slave 10.0.0.103:6379 10.0.0.103 6379 @ mymaster 10.0.0.101 6379 786:X 06 Nov 2022 15:21:28.672 # +failover-end master mymaster 10.0.0.101 6379 786:X 06 Nov 2022 15:21:28.672 # +switch-master mymaster 10.0.0.101 6379 10.0.0.102 6379 786:X 06 Nov 2022 15:21:28.672 * +slave slave 10.0.0.103:6379 10.0.0.103 6379 @ mymaster 10.0.0.102 6379 786:X 06 Nov 2022 15:21:28.672 * +slave slave 10.0.0.101:6379 10.0.0.101 6379 @ mymaster 10.0.0.102 6379 786:X 06 Nov 2022 15:21:31.738 # +sdown slave 10.0.0.101:6379 10.0.0.101 6379 @ mymaster 10.0.0.102 6379 786:X 06 Nov 2022 15:54:28.253 # -sdown slave 10.0.0.101:6379 10.0.0.101 6379 @ mymaster 10.0.0.102 6379
到此這篇關(guān)于redis實(shí)現(xiàn)sentinel哨兵架構(gòu)的文章就介紹到這了,更多相關(guān)redis哨兵架構(gòu)內(nèi)容請(qǐng)搜索腳本之家以前的文章或繼續(xù)瀏覽下面的相關(guān)文章希望大家以后多多支持腳本之家！
相關(guān)文章
Redis?的常用的集群方式主要有以下三種，分別是主從復(fù)制模式、哨兵模式、Redis-Cluster集群模式，那么下面我們就分別了解一下這三種集群模式的優(yōu)點(diǎn)與缺點(diǎn)2023-06-06
本文主要介紹了Redis 分片集群的實(shí)現(xiàn)，文中通過(guò)示例代碼介紹的非常詳細(xì)，對(duì)大家的學(xué)習(xí)或者工作具有一定的參考學(xué)習(xí)價(jià)值，需要的朋友們下面隨著小編來(lái)一起學(xué)習(xí)學(xué)習(xí)吧2023-01-01
這篇文章主要介紹了springboot使用Redis作緩存使用入門教程，本文通過(guò)實(shí)例代碼給大家介紹的非常詳細(xì)，對(duì)大家的學(xué)習(xí)或工作具有一定的參考借鑒價(jià)值,需要的朋友可以參考下2021-07-07
在分布式系統(tǒng)中,實(shí)現(xiàn)一個(gè)可靠的鎖機(jī)制是非常重要的,本文主要介紹了redis實(shí)現(xiàn)紅鎖的示例代碼,具有一定的參考價(jià)值,感興趣的可以了解一下2025-04-04
本篇文章主要介紹了深入理解redis分布式鎖和消息隊(duì)列，小編覺得挺不錯(cuò)的，現(xiàn)在分享給大家，也給大家做個(gè)參考。一起跟隨小編過(guò)來(lái)看看吧2017-07-07
文章講解了Redis哨兵機(jī)制的基本原理、主庫(kù)和從庫(kù)自動(dòng)切換的過(guò)程、如何減少誤判、哨兵集群的組成和通信機(jī)制,以及哨兵在故障發(fā)生時(shí)如何選舉Leader進(jìn)行主從切換2025-01-01
這篇文章主要介紹了redis單節(jié)點(diǎn)安裝與配置方式,具有很好的參考價(jià)值,希望對(duì)大家有所幫助,如有錯(cuò)誤或未考慮完全的地方,望不吝賜教2024-01-01
Redis中的數(shù)據(jù)結(jié)構(gòu)跳表詳解
跳表是一種基于并聯(lián)的鏈表結(jié)構(gòu),用于在有序元素序列中快速查找元素的數(shù)據(jù)結(jié)構(gòu),本文給大家介紹Redis中的數(shù)據(jù)結(jié)構(gòu)跳表,感興趣的朋友跟隨小編一起看看吧2024-06-06
Redis作為一款內(nèi)存數(shù)據(jù)庫(kù)，因?yàn)槭莾?nèi)存讀寫，所以性能很強(qiáng)，但內(nèi)存存儲(chǔ)是易失性的，斷電或系統(tǒng)奔潰都會(huì)導(dǎo)致數(shù)據(jù)丟失，因此Redis也需要將其數(shù)據(jù)持久化到磁盤上面，當(dāng)Redis服務(wù)重啟時(shí)，會(huì)把磁盤上的數(shù)據(jù)再加載進(jìn)內(nèi)存，Redis提供了兩種持久化機(jī)制-RDB快照和AOF日志2022-11-11
redis客戶端連接錯(cuò)誤 NOAUTH Authentication required
本文主要介紹了redis客戶端連接錯(cuò)誤 NOAUTH Authentication required，詳細(xì)的介紹了解決方法，感興趣的可以了解一下2021-07-07