Redis sentinel mode realizes one master, two slave and three Sentinels

Catalog

One 、redis Environmental Science ：

Two 、 The sentry introduces ：

3、 ... and 、 install redis:

Four 、 Use Redis The role of master-slave replication ：

5、 ... and 、 To configure redis One master and two subordinates ：

6、 ... and 、 To configure redis Three sentinels ：

One 、redis Environmental Science ：

Environmental Science ：redis6.2.6
linux One virtual machine ,contos7;

Two 、 The sentry introduces ：

1. One master, two slaves and three sentinels ：

One master, two slaves, three sentinels , When master When the node goes down , Through the Sentinels (sentinel) Re elect a new master node , Ensure cluster availability .

2. The main function of a sentry is ：
1. Cluster monitoring ： Responsible for monitoring Redis master and slave Is the process working .
2. Notice of news ： If a Redis The instance is faulty , Then the sentry is responsible for sending a message as an alarm to the Administrator .
3. Fail over ： If master node Hang up , Will automatically transfer to slave node On .
4. Configuration center ： If a fail over occurs , notice client Client new master Address .

PS： According to the recommendation mechanism , The number of sentinels in the cluster should be odd (3、5…)

3. Sentinels are used to achieve redis High availability of clusters , It's also distributed , Run as a sentinel cluster , Working together .

When failing over , Judge a master node Is it down , It needs the consent of most of the Sentinels , It's about distributed elections .

Even if some of the sentinel nodes are down , The Sentinels are still working , Because if a failover system, which is an important part of high availability mechanism, is single point , Then it's very embarrassing .

4. The core knowledge of sentinels ：

Sentinels need at least 3 An example , To ensure their own robustness .

sentry + redis Master slave deployment architecture , There is no guarantee of zero data loss , Only guarantee redis High availability of clusters .

For the sentry + redis Master slave complex deployment architecture , Try to test the environment and the production environment , They all have enough tests and drills .

3、 ... and 、 install redis:

Omit here redis Installation .

Four 、 Use Redis The role of master-slave replication ：

Master slave copy , It means to put one Redis Data of master node server , Copy to other Redis From the node server .

The primary node is called (master/leader), The slave node is called (slave/follower);

1, data redundancy ： Master-slave replication realizes hot backup of data , This is another way to implement persistence .

2, Fault recovery ： When there is a problem with the master node , Can be served by a slave node , Fast fault recovery ; It's actually a redundancy of services .

3, Read / write separation ：master Services are mainly used to write ,slave Services are mainly used to read data . It can improve the load capacity of the server , According to the change of demand , Add the number of slave nodes .

4, Load balancing ： At the same time, it is combined with read-write separation , The master node provides the write service , Provide read service from node , Share the load on the server . Write less and read more , Sharing read load through multiple slave nodes , Can greatly improve Redis Concurrency and load of services .

5, High availability cornerstone , Master-slave replication is the basis for the implementation of sentinel and cluster modes .

5、 ... and 、 To configure redis One master and two subordinates ：

In normal production, we prepare multiple servers , Like three , Each is in its own linux Installation in system redis.

But here for the convenience of demonstration , I use a computer to install virtual machine , Install three copies on the virtual machine redis service , To simulate the effect of one master and two slaves .

First step ： Modify the original redis.conf The configuration file ：

redis.conf The configuration file adopts the default mixed persistence , Separate RDB Persistence , Because the essence of master-slave replication is RDB The way , So there must be RDB Just participate in the way .

primary redis.conf There are several places in the configuration file that need to be modified ：

1. First, comment out the line of stand-alone connection in the network section , Note out bind 127.0.0.1 2. Similarly, we need to turn on the background operation ：daemonize no, Set to yes.3. take Protection mode is off ：protected-mode yes Change it to ：protected-mode no 4. open RDB Persistent configuration ：#RDB Persistence strategy There are three default methods ,[900 In seconds 1 Time modification ],#[300 In seconds 10 Time modification ],[60 In seconds 10000 Time modification ] Trigger RDB Persistence ,# We can manually modify this parameter or add a new policy save 900 1save 300 10save 60 10000 #RDB file name dbfilename dump.rdb#RDB File storage path dir ./ Policy configuration ：# stay seconds In seconds changes This data modification triggers RDB Persistence 5. Turn on AOF Persistent configuration appendonly yes#AOF file name appendfilename "appendonly.aof"#AOF File storage path And RDB Is the same parameter , Share a file path dir ./ # namely bin Under the table of contents #AOF Strategy , Usually choose the first [always: Every command records ],#[everysec: Once per second ],[no: When you are happy to see the machine, you can record ,linux Generally, it is synchronized once every half an hour ]#appendfsync alwaysappendfsync everysec# appendfsync no#aof The size of the file is larger than the last time it was rewritten , growth 100%( The configuration can be greater than 100%) when , Trigger override .#[ If the size is... After last rewriting 10MB, When AOF The document reaches 20MB The override will also be triggered again , And so on auto-aof-rewrite-percentage 100 #aof The file size exceeds 64MB*2 when , Trigger override ,# Why multiply by 2, because auto-aof-rewrite-percentage 100 Yes, double that 100%,# Rewrite only after doubling auto-aof-rewrite-min-size 64mb 6. Turn on hybrid persistence ：#6.aof-use-rdb-preamble yes # Check whether hybrid persistence is on ,redis5.0 After the default on

The second step ： Will modify the redis.conf Three copies of the configuration file ：

1. Get into redis Default installation path :cd /usr/local/binll # View folder contents cp redis.conf redis_bk.conf # Try to back up a configuration file , Prevent mistakes in modification , It can be reused cp redis.conf redis6379.confcp redis.conf redis6380.confcp redis.conf redis6381.conf

The third step ： Modify three configuration files respectively ：

vim redis6379.conf Revised content ：pidfile /var/run/redis_6379.pidport 6379dbfilename dump6379.rdbvim redis6380.conf Revised content ：pidfile /var/run/redis_6380.pidport 6380dbfilename dump6380.rdbvim redis6381.conf Revised content ：pidfile /var/run/redis_6381.pidport 6381dbfilename dump6381.rdb

redis Preparation before start up ：

Check to see if the fire wall of the virtual machine is closed , Firewall needs to be turned off

1: View fire protection status

systemctl status firewalldservice iptables status

2: Turn off the firewall temporarily

systemctl stop firewalldservice iptables stop

3: Permanently shut down firewall

systemctl disable firewalldchkconfig iptables off

Check the firewall status after closing

Open designated port 6379：

Mode 1 of opening designated ports ：
stay linux In the implementation of ： /sbin/iptables -I INPUT -p tcp --dport 6379 -j ACCEPT
redis Default port number 6379 Remote connection is not allowed , So set in the firewall 6379 Turn on remote service ;

Mode 2 of opening designated ports ：
Open the firewall before opening the specified port ：systemctl start firewalld
firewall-cmd --zone=public --add-port=6379/tcp --permanent

Step four ： Start three redis The server , And check to see if it started successfully ：

redis-server redis6379.confredis-server redis6380.confredis-server redis6381.confps -ef | grep redis

Step five ： Configure the slave library instead of the master library ：

Connect... First redis client ：

redis-cli -p 6380

And then 6380 The machine is configured as a slave ：

slaveof host ip Host port number #slaveof 127.0.0.1 6379info replication

And then 6381 The machine is configured as a slave ：

redis-cli -p 6381#slaveof host ip Host port number slaveof 127.0.0.1 6379info replication

View the slave information on the host ： Discovery configuration succeeded .

redis-cli -p 6379info replication

6、 ... and 、 To configure redis Three sentinels ：

Bring the official sentinel.conf( This file is in redis Under the source directory , That is, under the decompression path ） Copied to the /usr/local/bin：

Then make three copies ：

The changes are as follows ：

vim sentinel0.conf The document is amended as follows ：protected-mode nodaemonize yesport 26379 #sentinel port dir “/usr/local/bin”sentinel monitor mymaster 127.0.0.1 6379 2

And then modify sentinel1.conf The information is as follows ：

protected-mode nodaemonize yesport 26380 # sentinel port , Because we are on a virtual machine , So the ports are different dir “/usr/local/bin”sentinel monitor mymaster 127.0.0.1 6379 2 # there mymaster The two copies should be the same ,mymaster Same hostname

And then modify sentinel2.conf The information is as follows ：

protected-mode nodaemonize yesport 26381 # sentinel port , Because we are on a virtual machine , So the ports are different dir “/usr/local/bin”sentinel monitor mymaster 127.0.0.1 6379 2 # there mymaster The two copies should be the same

Configuration file details
The configuration of Sentry is mainly to modify sentinel.conf Parameters in the configuration file , stay Redis This configuration file can be seen in the installation directory , The detailed explanation of each parameter is as follows :

# sentry sentinel The port on which the instance runs , Default 26379 port 26379# sentry sentinel Working directory of dir ./# Whether to turn on the protection mode , Default on .protected-mode:no# Whether it is set to background startup .daemonize:yes# sentry sentinel Log file logfile:./sentinel.log# sentry sentinel Monitored redis The master node ## ip： host ip Address ## port： Sentry port number ## master-name： You can name the master node by yourself ## quorum: When these quorum Number sentinel The sentry thought master Primary node lost connection Then at this time # It is objectively believed that the primary node is disconnected # sentinel monitor <master-name> <ip> <redis-port> <quorum> sentinel monitor mymaster 127.0.0.1 6379 2# When in Redis In the example, the requirepass, All connections Redis The client of the instance must provide the password .# sentinel auth-pass <master-name> <password> sentinel auth-pass mymaster 123456 # Specify the master node to answer the sentinel sentinel The maximum time interval , Beyond that time , Sentinels subjectively think that the primary node is offline ,# Default 30 second # sentinel down-after-milliseconds <master-name> <milliseconds>sentinel down-after-milliseconds mymaster 30000 # Specifies when failover When switching between active and standby , How many at most slave Simultaneously on the new master To synchronize .# The smaller the number , complete failover The longer it takes ; conversely , But if the number is bigger , Means # The more slave because replication And is not available . You can do this by setting this value to zero 1, To make sure there's only one at a time slave,# In a state where command requests cannot be processed .# sentinel parallel-syncs <master-name> <numslaves>sentinel parallel-syncs mymaster 1 # Timeout for failover failover-timeout, Default three minutes , It can be used in the following ways ：## 1. The same sentinel To the same master two failover The time between . ## 2. When one slave From a wrong one master It starts when you synchronize data there , until slave Be corrected from correct # Of master Where the synchronization ends . ## 3. When you want to cancel an ongoing failover The time required for .## 4. When doing failover when , Configure all slaves Point to the new master Maximum time required . however ,# Even after this timeout ,slaves It will still be correctly configured to point master, But it won't parallel-syncs The configured rules to synchronize data # sentinel failover-timeout <master-name> <milliseconds> sentinel failover-timeout mymaster 180000# When sentinel When any warning level event occurs （ for instance redis Subjective failure and objective failure of examples, etc ）,# Will call this script . The maximum execution time of a script is 60s, If I exceed that time ,# The script will be a SIGKILL Signal termination , And then re execute .# There are the following rules for the result of the script ： ## 1. If the script returns 1, Then the script will be executed again later , The number of repetitions currently defaults to 10.## 2. If the script returns 2, Or 2 A higher return value , The script will not repeat . ## 3. If the script is terminated during execution due to receiving a system interrupt , Then the same return value is 1 The same behavior when .# sentinel notification-script <master-name> <script-path> sentinel notification-script mymaster /var/redis/notify.sh# This script should be generic , Can be called many times , It's not targeted .# sentinel client-reconfig-script <master-name> <script-path>sentinel client-reconfig-script mymaster /var/redis/reconfig.sh

Finally, start three services first , Then get up and send off three sentinels , And then the client ;
Activate the sentry ：

redis-sentinel sentinel0.confredis-sentinel sentinel1.confredis-sentinel sentinel2.conf

View after startup sentinel Information ：redis-cli -p sentinel Port number

redis-cli -p 26379info sentinel

  This is about Redis The sentinel mode realizes one master, two slave and three sentinels, which is introduced here , More about Redis Please search the previous articles of SDN or continue to browse the relevant articles below. I hope you will support SDN more in the future ！