10 "no no no" redis interview questions

Here are Redis Interview questions , I believe everyone will feel familiar and strange 、 I've seen it. I may forget it immediately after a short interview .JavaPub Sort out these easily forgotten key knowledge and answer , Recommended collection , Review and consult often .

See the comments section

@

• 1. Redis What is it? ？

• 2. What scenarios do you use redis

• 3. Why? Redis It's single threaded ?

• 4. Redis There are several ways to persist ？

• 5. What is cache penetration ？ How to solve ？

• 6. What is a cache avalanche ?

• 7. Redis How to optimize memory in use ？

• 8. You redis Which deployment method to use ？

• 9. redis What should we pay attention to when implementing distributed locks ？

1. Redis What is it? ？

Generally speaking, you should answer the following questions at least , If you don't know the basics , Recommended reading 《 rodert Single row learning redis》

Redis It's a Memory based key-value  The storage system , The data structure includes character string 、list、set、zset（sorted set -- Ordered set ） and hash,bitmap,GeoHash( coordinate ),HyperLogLog,Streams（5.x After the version ）

2. What scenarios do you use redis

You have practical experience , Then go straight to the show . without , Select some of the following classic scenes

1. Use as a queue ,（ Because it's memory based 、 Generally, it will not be used as a consumption queue 、 As a circular queue, it is necessary to apply ）;

2. The simulation is similar to token This kind of scenario needs to set the expiration time , Login failure ;

3. Distributed cache , Avoid a large number of requests to the underlying relational database , Greatly reduce database pressure ;

4. Distributed lock ;

5. be based on bitmap Implement the bloon filter ;

6. Ranking List - be based on zset（ Ordered collection data types ）;

7. Counter - For views 、 The playback volume is high , Use redis incr Realize the counter function ;

8. Distributed session ;

9. The messaging system ;

3. Why? Redis It's single threaded ?

This question gives an official answer

because Redis Is a memory based operation ,CPU No Redis Bottleneck ,Redis The most likely bottleneck is the size of machine memory or network bandwidth . Since a single thread is easy to implement , and CPU Not a bottleneck , So it's natural to adopt a single thread solution .

4. Redis There are several ways to persist ？

redis There are two ways to persist , Namely Snapshot mode （RDB Redis DataBase） And file appending （AOF Append Only File）.

Obvious , Snapshot restart recovery is fast 、 But data is more likely to be lost , The data appended to the file is more complete 、 Slow restart recovery .

Mixed persistence ,Redis 4.0 And then the new way , Hybrid persistence is a combination of RDB and AOF The advantages of , When writing, first put the current data in RDB Is written to the beginning of the file , Then, the following operations are represented by AOF In the file , This will ensure the speed of restart , And reduce the risk of data loss .

On recovery , Restore the files saved in snapshot mode first , Then restore the incremental data in the appended file .

5. What is cache penetration ？ How to solve ？

Cache penetration refers to that the data requested by the user does not exist in the cache, that is, there is no hit , And it doesn't exist in the database , As a result, every time the user requests the data, he / she has to query the database again , Then return to empty .

If a malicious attacker keeps asking for data that doesn't exist in the system , It will result in a large number of requests falling on the database in a short time , Cause too much database pressure , Even the database system .

This is called cache penetration .

How to solve ？

• The query result is also cached if it is empty , The cache time is set a little shorter , Or should key Corresponding data insert Then clean up the cache .

• Yes, it must not exist key To filter . You can put all the possibilities key Put it in a big Bitmap in , Query through the Bitmap Filter .( That is, the principle of the bloom filter ： Blum filter )

6. What is a cache avalanche ?

Cache avalanche refers to the time from bulk to expiration of data in cache , And the amount of query data is huge , The request goes directly to the database , Cause too much pressure on the database and even downtime . Unlike cache breakdown , Cache breakdown refers to concurrent query of the same data , The cache avalanche is that different data has expired , A lot of data can't be found to look up the database .

How to solve ？

Common solutions are ：

• Even expiration

• Add mutex lock

• Cache never expires

• Double layer cache strategy

Even expiration ： Set different expiration times , Make the cache failure time as uniform as possible . It is usually possible to add a random value to the validity period or to plan the validity period uniformly .

Add mutex lock ： It's consistent with the solution of cache breakdown , Let only one thread build the cache at the same time , Other threads block queuing .

Cache never expires ： It's consistent with the solution of cache breakdown , Caches never expire physically , Update the cache with an asynchronous thread .

Double layer cache strategy ： Use the primary and secondary layer cache ：

Main cache ： The validity period is set according to the experience value , Set the cache for primary reads , Load the latest value from the database after the primary cache fails .

Backup cache ： It has a long validity period , Cache read when lock failed , When the primary cache is updated, the backup cache needs to be updated synchronously .

7. Redis How to optimize memory in use ？

1. Shorten the length of key value

• Shortening the length of the value is the key , If the value is a large business object , You can serialize objects into binary arrays ;

• First, we should streamline our business , Remove unnecessary attributes , Avoid storing useless data ;

• The second is the selection of serialization tools , A more efficient sequencer should be chosen to reduce the byte array size ;

• With JAVA For example , The built-in serialization method is not satisfactory in terms of speed or compression ratio , At this point, you can choose a more efficient serialization tool , Such as : protostuff,kryo etc.

1. Shared object pool

Object shared pool refers to Redis Internal maintenance [0-9999] Integer object pool for . Create a large number of integer types redisObject There is memory overhead , Every redisObject Internal structure accounts for at least 16 byte , It even exceeds the space consumption of the integer itself . therefore Redis Memory maintenance one [0-9999] Integer object pool for , Used to save memory . Except for integer value objects , Other types are list,hash,set,zset Internal elements can also use integer object pools . Therefore, in the development, on the premise of meeting the requirements , Try to use integer objects to save memory .

1. String optimization

because redis Lazy deletion mechanism , The space after string reduction is not released , Reserved as preallocated space . Try to add without updating .

1. Coding optimization

The so-called coding is what kind of underlying data structure is used to realize . Different codes will directly affect the memory occupation and read-write efficiency of data .

This needs to be mastered redis Underlying data structure . As a reference ：

image

1. control key The number of

8. You redis Which deployment method to use ？

redis The deployment is divided into single nodes 、 Master-slave deployment （master-slave）、 Sentinel deployment （Sentinel）、 Cluster deployment （cluster）.

A single node ： That is, stand-alone deployment ;

Master-slave deployment ： Divided into one master and one slave or one master and many slaves , The synchronization between master and slave is divided into full or incremental . Quantity synchronization ：master Node passing BGSAVE Generate corresponding RDB file , And send it to slave node ,slave After receiving the write command, the node will master The sent file is loaded and written ; The incremental synchronization ： That is to say master-slave Relationship building begins ,master Every time the data change command is executed, it will be synchronized to slave node . Generally, write requests are forwarded to master, Forward the read request to slave. Improved redis Performance of .

Sentinel deployment ： There are sentinel clusters and Redis The master-slave cluster of , Sentry as a monitoring process in the operating system , Monitor each one accordingly Redis example , If master Service exception （ping pong The node does not reply and has exceeded a certain time ）, It will be confirmed among multiple sentinels , If more than half of them confirm that the service is abnormal , On the other hand master The service goes offline , And elect the current one slave Node to convert to master node ; If slave Node service exception , It was also confirmed by several sentinels , Go offline . Improved redis Cluster high availability features , And the enhancement of horizontal expansion capability .

image

Cluster deployment ： Belong to “ De centralization ” One way , Multiple master The node saves all the data in the entire cluster , And the data is based on key Conduct crc-16 Check the algorithm to hash , take key Hash into corresponding 16383 individual slot, and Redis cluster Each of the clusters master Nodes are responsible for different slot Range . Every master Multiple nodes can be configured under the node slave node , It can also be reused in the cluster sentinel Sentry improves the high availability of the entire cluster .

image

9. redis What should we pay attention to when implementing distributed locks ？

1. The locking process should ensure atomicity ;

2. Ensure that the lock added by who can only be unlocked by who , namely Redis The lock value, You need to pass in the same... When unlocking value To succeed , Guarantee value Uniqueness ;

3. Set lock timeout , Prevent other clients from being unable to acquire the lock when the lock adding party fails to release the lock abnormally , meanwhile , The timeout time should be greater than the business processing time ;

Use Redis command  SET lock_key unique_value NX EX seconds  To lock , Single command operation ,Redis Is a serial execution command , Therefore, only one can be locked successfully .

---

Low valley energy storage

Recommend open source SpringBoot+VUE System Twentythousand 、 Video tutorial attached ：https://gitee.com/rodert/liawan-vue

Open personal wechat , Little buddy, you can add my big size , Quota co., LTD. , First come first served basis , If it's full again, it's really gone

Scan the QR code below to add my wechat ,2022, Huddle for warmth , Let's fight together .