[redis series] redis learning 13. Redis often asks simple interview questions

Let's see redis Often ask common interview questions

Redis What is a ？

http://www.redis.cn/ redis The Chinese website gives a very clear and clear definition

[ Failed to transfer the external chain picture , The origin station may have anti-theft chain mechanism , It is suggested to save the pictures and upload them directly (img-5HZnf7He-1650460558149)(C:\Users\Administrator\AppData\Roaming\Typora\typora-user-images\image-20210828092258505.png)]

• Redis It's open source （BSD The license ） Of , Data structure storage system in memory

• Redis Can be used as a database 、 Caching and message middleware

• The supported data structures are 8 Kind of

  character string （strings）, hash （hashes）, list （lists）, aggregate （sets）, Ordered set （sorted sets）, bitmaps, hyperloglogs and geospatial

Use redis What are the benefits ？

• The data accessed is very fast ,redis The data is operated directly in memory
• Rich features ： Available for caching , Message queue , wait
• Support 8 Kind of data type

Redis Why so fast ？

First redis It's single threaded , Many people think that a single thread must be slow , Multithreading must be fast , This is a misunderstanding

• redis Using multiple channels I/O Reuse model , Non blocking Of IO
• Using single thread , There will be no consumption of context switching , Don't think about locking and unlocking , There is no performance consumption of competing resources
• The data structure is simple , Above 8 Kind of data type , Master something without too much effort
• operation redis The data of , Basically in the proper operation memory , Must be quick

Redis Can it be persistent ？ If you can , What are the ways of persistence ？

It can persist , Yes 2 Ways of planting , RDB and AOF , redis The default is to support RDB Persistence

• RDB Persistence mode

Writes a dataset snapshot in memory to disk within a specified time interval , This is the snapshot snapshot , When restoring a snapshot , Is to read the snapshot file into memory .

redis adopt fork Create a child process to do persistent actions

advantage

• Suitable for large-scale data recovery
• The requirement of data integrity is not high

Inferiority

• Operations need to be performed at certain intervals , If redis Unexpected downtime , The last written data will be lost

• fork The subprocess needs to occupy a certain space

• AOF Persistence mode

AOF yes redis Another way of persistence , Log every write operation , take redis All write operations performed are recorded , Only additional files are allowed , Rewriting of files is not allowed

advantage

• Every operation reids Will be recorded , The integrity of the file is good
• Synchronize once per second , One second of data could be lost
• No synchronization , This is the most efficient

Inferiority

• Relative to data files ,aof The file will be much larger than rdb file , It's faster than rdb The files are slow
• aof Operating efficiency ratio rdb slow , therefore redis The default configuration is rdb Persistence

Redis What are the application scenarios ？

• Simple message queuing

Simple message queue for publishing and subscribing , Use to redis in List data structure , Can pass lpush and rpop Write and read messages , This only applies to simple message queues , If you have a lot of data , Require data consistency , Good performance, etc , Of course, professional message queuing components are preferred

• Used to cache

Access data that you often need to access redis in , And set up redis Maximum memory limit and deletion key The strategy of

• For counter

For example, web page visits , Like and so on ,redis Support self increasing and self decreasing operations , The data is all in memory , Very suitable for counting

• Do distributed locks

redis The order that comes with you SETNX , Naturally, it can be used as a distributed lock , It is used to control the orderly operation of multiple nodes , Of course redis I have provided RedLock For everyone to use

RedLock What is a

Redis The information provided by the official website is based on Redis How to implement distributed locks , It has the following characteristics ：

• Fault tolerance

As long as most Redis If the node survives, it can provide services normally

• Safety features

RedLock Mutually exclusive access , There will always be only one redis The client can get the lock , Perform the operation

• Avoid deadlock

Final redis The client may get the lock , There will be no deadlock

• Make a leaderboard

Use zset Make a leaderboard ,top xx And so on

• Do common concern , Be a good friend around you

redis Provide a collection set , Can calculate the common friends , Common interests , And so on

Redis Delete expired key How is it handled ？

Yes 3 Ways of planting ：

• Timed expiration

Timed expiration , Every time you set a key When , All for this key Set the corresponding timer , When this key ttl At maturity , Will this key Delete the

Every key When the expiration time is set, we need to get a timer , such It's very exhausting cpu resources , And then affect redis performance

• Expire regularly

Expire regularly , Is to set every other period of time , scanning redis Set the expiration time in key, And delete the expired key.

• Inertia expires

Inertia expires , Whenever you visit this key When , To judge this key Is it invalid , Delete when it fails , If you don't visit , Then this key It's always in memory , It takes up a lot of memory

Redis What is the essence of affairs ？

Is a collection of commands , All commands in a transaction are serialized , In the process of transaction execution , The commands are executed in order , They have

• Disposable
• Sequence
• exclusive

redis There is no concept of isolation level for transactions

redis Transaction , Execute the command like this

Put commands in transactions , Not immediately , But only when the execution order is initiated , adopt exec Trigger

redis It's a single instruction that guarantees atomicity , But transactions don't guarantee atomicity

The process of executing a transaction is like this ：

• multi Open transaction
• Join the team with various orders
• exec Perform transactions

The transaction ACID What do you mean by separation ？

• Atomicity Atomicity

Atomicity refers to , A business is an indivisible whole , Or make it together , Or fail together

• Consistency Uniformity

Before and after the execution of the transaction , The integrity of the data must be consistent

• Isolation Isolation,

When multiple transactions are processed at the same time , Mutual interference , The complementary effect , Let's play

• Durability persistence

refer to , Once the transaction is committed , So the data affected is persistent

Redis What is the principle of master-slave replication ？

Slave Start successfully connected to master A... Will be sent later sync command

master After receiving the order , Start the background save process , At the same time, collect all commands received to modify the database set , After the background process is executed ,master The process will transfer the entire data file to slave , And complete a synchronization

Copy in full

slave Service received master After the data transmitted , Save it and load it into memory

Incremental replication

master Pass all the newly collected modified commands to slave , And complete synchronization

Once reconnected master node , A full synchronization is performed

What does sentinel mode do ？

Sentinel mode , Mainly for the realization of redis High availability of clusters

Because using master-slave replication , When redis When the host goes down , There is no way to elect a host , Sentinel mode can

The role of a sentry ：

• By sending commands ,Redis The server returns monitoring status information , Including master and slave servers
• If the sentry detects that the primary server is down , Will automatically slave Switch master, Then notify other slave servers by publishing and subscribing , Modify the configuration file , Let him be the host

Subjective offline

If master Server down , Then one of the Sentinels will detect , The system does not execute immediately failover The process of , Just the current sentinel , Judge master Unavailable , This is Subjective offline

Objective offline

The general sentry in the cluster is also a cluster , If deployed 3 A sentinel

When the other two sentinels found master When the server is unavailable , Then there will be a vote between the Sentinels , We will write the specific voting algorithm later , The voting structure was initiated by a sentry , Conduct failover Failover operations , After successful switching , Through the publish and subscribe mode , Let each monitoring sentry switch its monitoring server to this master Server , This is Objective offline

redis Of through , breakdown , Avalanche problem can be seen in the previous article ：【Redis series 】redis Learn twelve ,redis Cache penetration , Cache breakdown , Cache avalanche

Reference material ：

redis_doc

Welcome to thumb up , Focus on , Collection

friends , Your support and encouragement , I insist on sharing , The drive to improve quality

Okay , That's it this time

Technology is open , Our mindset , It should be more open . Embrace change , Born in the sun , Try to move on .

I am a Little Devil boy Nezha , Welcome to like, pay attention to collection , See you next time ~