主從複製，讀寫分離 Master/Slave

是什麼

master寫入

slave讀取

能幹嘛

讀寫分離，更加安全，效能提升

怎麼玩

一主二僕、薪火相傳、反客為主

周明老師，能夠把長篇大論總結的很精闢。

1. 配從不配主
2. slaveof 主庫ip 主庫埠

準備三臺機器

一主，二從

66是主機，61、62作為從機。

通過info replication 進行檢視身份。

192.168.1.66:6379> info replication # Replication role:master connected_slaves:0 master_replid:ba9a0c9d5cbeb6f7ce375b4c3559f5e848fc7025 master_replid2:0000000000000000000000000000000000000000 master_repl_offset:0 second_repl_offset:-1 repl_backlog_active:0 repl_backlog_size:1048576 repl_backlog_first_byte_offset:0 repl_backlog_histlen:0 
 

設定61、62跟隨66

192.168.1.61:6379> slaveof 192.168.1.66 6379 OK 192.168.1.61:6379> info replication # Replication role:slave master_host:192.168.1.66 master_port:6379 master_link_status:down master_last_io_seconds_ago:-1 master_sync_in_progress:0 slave_repl_offset:1 master_link_down_since_seconds:1513056594 slave_priority:100 slave_read_only:1 connected_slaves:0 master_replid:ba9a0c9d5cbeb6f7ce375b4c3559f5e848fc7025 master_replid2:0000000000000000000000000000000000000000 master_repl_offset:0 second_repl_offset:-1 repl_backlog_active:0 repl_backlog_size:1048576 repl_backlog_first_byte_offset:0 repl_backlog_histlen:0 
 

192.168.1.62:6379> slaveof 192.168.1.66 6379 OK 192.168.1.62:6379> info replication # Replication role:slave master_host:192.168.1.66 master_port:6379 master_link_status:down master_last_io_seconds_ago:-1 master_sync_in_progress:0 slave_repl_offset:1 master_link_down_since_seconds:1513056636 slave_priority:100 slave_read_only:1 connected_slaves:0 master_replid:ff1f0f120165c6673a797e65aa0d82e3ccbe9a6c master_replid2:0000000000000000000000000000000000000000 master_repl_offset:0 second_repl_offset:-1 repl_backlog_active:0 repl_backlog_size:1048576 repl_backlog_first_byte_offset:0 repl_backlog_histlen:0 
 

tips:MASTER aborted replication with an error: NOAUTH Authentication required. 出現了這個錯誤。需要從伺服器中新增配置。masterauth 123456 (123456是主redis的密碼，此引數是當與主連線時的密碼驗證)

設定了主從複製之後，此時在主機中，

192.168.1.66:6379> set k4 v4 OK 

在從機中檢視，

192.168.1.61:6379> get k4 "v4" 

192.168.1.62:6379> get k4 "v4" 

只要變為從機，主機中資料都會被弄過來！

周陽老師語錄：學的不是指令，而是探尋知識的過程！指令分分鐘講完！但是思考知識更重要！

不停的破壞，不停的做實驗，不停的嘗試！折騰！才能更好的理解知識！

周明老師語錄：讀書或者看視訊，產生爭議和思考了，比不思考要有收穫。帶著思考學習，帶著問題讀書。

在主從複製中，如果主機宕機了，從機還是從機！原地待命中！

主機恢復之後，主從體系不會被破壞！

薪火相傳 66 傳61，61傳62

192.168.1.66:6379> info replication
# Replication
role:master
connected_slaves:1
slave0:ip=192.168.1.61,port=6379,state=online,offset=3288,lag=1 master_replid:dddb7b383d4153816d28377f65c6ed7d688ac8bf master_replid2:0000000000000000000000000000000000000000 master_repl_offset:3288 second_repl_offset:-1 repl_backlog_active:1 repl_backlog_size:1048576 repl_backlog_first_byte_offset:1 repl_backlog_histlen:3288 

192.168.1.61:6379> info replication
# Replication
role:slave
master_host:192.168.1.66
master_port:6379 master_link_status:up master_last_io_seconds_ago:7 master_sync_in_progress:0 slave_repl_offset:6197 slave_priority:100 slave_read_only:1 connected_slaves:1 slave0:ip=192.168.1.62,port=6379,state=online,offset=6197,lag=0 master_replid:dddb7b383d4153816d28377f65c6ed7d688ac8bf master_replid2:0000000000000000000000000000000000000000 master_repl_offset:6197 second_repl_offset:-1 repl_backlog_active:1 repl_backlog_size:1048576 repl_backlog_first_byte_offset:573 repl_backlog_histlen:5625 

192.168.1.62:6379> info replication # Replication role:slave master_host:192.168.1.61 master_port:6379 master_link_status:up master_last_io_seconds_ago:3 master_sync_in_progress:0 slave_repl_offset:5931 slave_priority:100 slave_read_only:1 connected_slaves:0 master_replid:dddb7b383d4153816d28377f65c6ed7d688ac8bf master_replid2:0000000000000000000000000000000000000000 master_repl_offset:5931 second_repl_offset:-1 repl_backlog_active:1 repl_backlog_size:1048576 repl_backlog_first_byte_offset:5819 repl_backlog_histlen:113 

注意了，master_link_status:up表示成功，master_link_status:down表示失敗。原因可能是主機的埠號沒有開啟！

iptables -I INPUT 4 -p tcp -m state --state NEW -m tcp --dport 6379 -j ACCEPT

反客為主命令

192.168.1.61:6379> SLAVEOF no one OK 192.168.1.61:6379> set k11 v11 OK

192.168.1.62:6379> get k11 "v11" 192.168.1.62:6379> set k12 v12 (error) READONLY You can't write against a read only slave.

前提是，你下面要有小弟，你才能反客為主！

主從複製，第一次會全量複製，之後是增量複製。只要重新連線master，都會執行一次全量複製。

哨兵模式，反客為主的自動版本！主機掛了，會自動從從機中選擇一個牛人，作為新的主機。在哨兵模式中，主機迴歸之後，變成從機了。

怎麼玩

在從機中新建一個哨兵配置sentinel.conf

sentinel monitor mymaster 192.168.1.61 6379 1 sentinel auth-pass mymaster 123456

開啟哨兵監控！

redis-sentinel /myredis/sentinel.conf 

[[email protected] bin]# redis-sentinel /myredis/sentinel.conf 
4880:X 12 Dec 16:15:34.969 # oO0OoO0OoO0Oo Redis is starting oO0OoO0OoO0Oo 4880:X 12 Dec 16:15:34.969 # Redis version=4.0.1, bits=64, commit=00000000, modified=0, pid=4880, just started 4880:X 12 Dec 16:15:34.969 # Configuration loaded 4880:X 12 Dec 16:15:34.970 * Increased maximum number of open files to 10032 (it was originally set to 1024). _._ _.-``__ ''-._ _.-`` `. `_. ''-._ Redis 4.0.1 (00000000/0) 64 bit .-`` .-```. ```\/ _.,_ ''-._ ( ' , .-` | `, ) Running in sentinel mode |`-._`-...-` __...-.``-._|'` _.-'| Port: 26379 | `-._ `._ / _.-' | PID: 4880 `-._ `-._ `-./ _.-' _.-' |`-._`-._ `-.__.-' _.-'_.-'| | `-._`-._ _.-'_.-' | http://redis.io `-._ `-._`-.__.-'_.-' _.-' |`-._`-._ `-.__.-' _.-'_.-'| | `-._`-._ _.-'_.-' | `-._ `-._`-.__.-'_.-' _.-' `-._ `-.__.-' _.-' `-._ _.-' `-.__.-' 4880:X 12 Dec 16:15:34.973 # WARNING: The TCP backlog setting of 511 cannot be enforced because /proc/sys/net/core/somaxconn is set to the lower value of 128. 4880:X 12 Dec 16:15:34.992 # Sentinel ID is 5bdc4c724103019a7d987848ce6a8af91341ee1d 4880:X 12 Dec 16:15:34.992 # +monitor master mymaster 192.168.1.66 6379 quorum 1 4880:X 12 Dec 16:15:46.522 * +slave slave 192.168.1.61:6379 192.168.1.61 6379 @ mymaster 192.168.1.66 6379 4880:X 12 Dec 16:15:46.593 * +slave slave 192.168.1.62:6379 192.168.1.62 6379 @ mymaster 192.168.1.66 6379 4880:X 12 Dec 16:16:16.655 # +sdown slave 192.168.1.62:6379 192.168.1.62 6379 @ mymaster 192.168.1.66 6379 4880:X 12 Dec 16:17:02.014 # +sdown master mymaster 192.168.1.66 6379 4880:X 12 Dec 16:17:02.014 # +odown master mymaster 192.168.1.66 6379 #quorum 1/1 4880:X 12 Dec 16:17:02.014 # +new-epoch 1 4880:X 12 Dec 16:17:02.014 # +try-failover master mymaster 192.168.1.66 6379 4880:X 12 Dec 16:17:02.024 # +vote-for-leader 5bdc4c724103019a7d987848ce6a8af91341ee1d 1 4880:X 12 Dec 16:17:02.024 # +elected-leader master mymaster 192.168.1.66 6379 4880:X 12 Dec 16:17:02.024 # +failover-state-select-slave master mymaster 192.168.1.66 6379 4880:X 12 Dec 16:17:02.125 # +selected-slave slave 192.168.1.61:6379 192.168.1.61 6379 @ mymaster 192.168.1.66 6379 4880:X 12 Dec 16:17:02.125 * +failover-state-send-slaveof-noone slave 192.168.1.61:6379 192.168.1.61 6379 @ mymaster 192.168.1.66 6379 4880:X 12 Dec 16:17:02.184 * +failover-state-wait-promotion slave 192.168.1.61:6379 192.168.1.61 6379 @ mymaster 192.168.1.66 6379 4880:X 12 Dec 16:17:03.085 # +promoted-slave slave 192.168.1.61:6379 192.168.1.61 6379 @ mymaster 192.168.1.66 6379 4880:X 12 Dec 16:17:03.085 # +failover-state-reconf-slaves master mymaster 192.168.1.66 6379 4880:X 12 Dec 16:17:03.184 * +slave-reconf-sent slave 192.168.1.62:6379 192.168.1.62 6379 @ mymaster 192.168.1.66 6379 4880:X 12 Dec 16:17:03.184 # +failover-end master mymaster 192.168.1.66 6379 4880:X 12 Dec 16:17:03.184 # +switch-master mymaster 192.168.1.66 6379 192.168.1.61 6379 4880:X 12 Dec 16:17:03.185 * +slave slave 192.168.1.62:6379 192.168.1.62 6379 @ mymaster 192.168.1.61 6379 4880:X 12 Dec 16:17:03.185 * +slave slave 192.168.1.66:6379 192.168.1.66 6379 @ mymaster 192.168.1.61 6379 4880:X 12 Dec 16:17:33.238 # +sdown slave 192.168.1.66:6379 192.168.1.66 6379 @ mymaster 192.168.1.61 6379 4880:X 12 Dec 16:17:33.238 # +sdown slave 192.168.1.62:6379 192.168.1.62 6379 @ mymaster 192.168.1.61 6379 4880:X 12 Dec 16:22:14.045 # -sdown slave 192.168.1.62:6379 192.168.1.62 6379 @ mymaster 192.168.1.61 6379 4880:X 12 Dec 16:25:14.815 * +fix-slave-config slave 192.168.1.62:6379 192.168.1.62 6379 @ mymaster 192.168.1.61 6379 

哨兵配置檔案可以放在任意一個從伺服器中。

周陽語錄：這些牛逼的技術，公司不一定讓你去做，但是你要懂！懂了才有機會去做！

主從複製的缺點，是有一定的延遲，主資料更新到從資料庫有一定的延遲。