Basic knowledge of redis

Redis So widely used in Internet technology storage , Almost all back-end technology interviewers are Redis The use and principle of 360° It's difficult . Take it one at a time as an Internet company offer Face bully of （ Allow me to use the rhetoric of exaggeration ）, Beat countless competitors , Each time can only see countless lonely figure disappointed to leave , A little guilty , On a lonely night , I thought over the pain , Decided to start writing 《 Condole interviewer 》 series , I hope it can help you in your future interview , To the interviewer 360° The counterattack , Ask your interviewer , Let the colleagues in the interview come back , Crazy harvest factory offer！

The interview begins

A paunch , A middle-aged man in a plaid shirt , Take one full of scratches mac Coming to you , Look at the Balding Hair , I thought it must be NIMA's top architect ！ But we are full of poetry and calligraphy , Empty is not empty .

Hello, young man , Look at your resume. It's used in your project Redis, Why do you use Redis？

I can't help but scold , What's the problem , We all use this one , But you can't say .

I mean it ： Hello, handsome and charming interviewer , Because traditional relational databases like Mysql It can't be applied to all the scenes , For example, the stock deduction of seckill ,APP Home page traffic peak and so on , It's easy to crack the database , So cache middleware is introduced , At present, the commonly used cache Middleware in the market is Redis and Memcached But he took their advantages and disadvantages into account , Finally, I chose Redis.

As for the more detailed comparison, friends remember to check Redis and Memcached The difference between , For example, the advantages and disadvantages of the two and their respective scenes , I will write it out later when I have time .

The young man , I'll ask you again ,Redis What are the data structures ？

String、Hash、List、Set、SortedSet.

Here I believe 99% All of our readers can answer Redis Of 5 Basic data types . If we can't answer the question, we need to make up for it , You know that the five most suitable scenes are better .

however , If you are Redis Middle and advanced users , And you need to highlight the difference between you and other candidates in this interview , You need to add the following data structures HyperLogLog、Geo、Pub/Sub.

If you want more points , Then you said you played Redis Module, image BloomFilter,RedisSearch,Redis-ML, At this time the interviewer's eyes start to shine , I thought this young man There's something .

notes ： I answered in an interview Redis When it comes to relevant questions , Often mentioned BloomFilter（ The bloon filter ） There are a lot of real scenarios for this thing , And it's really fragrant , The principle is easy to understand , Take a look at the article and you can talk in front of the interviewer .

If there's a lot of key You need to set the expiration date at the same time , What should we pay attention to in general ？

If a large number of key Expiration time is set too centrally , By the time it expires ,Redis There may be a short-term stumbling . In serious cases Cache avalanche , We usually need to add a random value to the time , Spread the expiration time a little bit .

E-commerce home page often uses timed tasks to refresh the cache , It's possible that a large number of data failure times are very concentrated , If the failure time is the same , And just at the point of failure, a large number of users swarmed in , It is possible to cause Cache avalanche

Then you used Redis Distributed locks , What is it ？

First take setnx To fight for the lock , After grabbing it , Reuse expire Add an expiration time to the lock to prevent it from forgetting to release .

At this time, the other party will tell you that you have a good answer , Then ask if you are setnx After performing expire The previous process was unexpected crash Or restart maintenance , What then ？

You have to give surprising feedback ： alas , Yeah , The lock will never be released . And then you need Grab your head , Pretend to think for a moment , It seems that the next result is the result of your active thinking , Then answer ： I remember set Instructions have very complex parameters , This should be able to put setnx and expire To synthesize an instruction for use ！

The other person will smile , I began to meditate ： Um. , This boy is not bad , It's beginning to be interesting . If Redis There are 1 One hundred million key, Among them is 10w individual key It starts with a fixed known prefix , How to find them all ？

Use keys The command can scan out the key list .

The other side then asked ： If this redis Providing services for online business , That use keys What's wrong with the instructions ？

At this time, you have to answer Redis A key feature ：Redis Of single thread .keys Instructions can cause threads to block for a period of time , Online services will pause , Until the command is executed , Service can be restored . It can be used at this time scan Instructions ,scan Instructions It can extract the specified patterns without blocking key list , But there is a certain probability of repetition , In the client to do a duplicate can , But the overall time spent will be more than the direct use of keys Commander .

however , Incremental iteration commands are not without drawbacks ： for instance , Use SMEMBERS The command returns all the elements currently contained in the set key , But for SCAN For this kind of incremental iteration command , Because in the process of incrementally iterating on the bond , The key may be modified , So the incremental iteration command can only provide limited guarantee for the returned elements .

Have used Redis Do asynchronous queues , How do you use ？

In general use list Structure as a queue ,rpush Production news ,lpop News consumption . When lpop When there is no news , Be appropriate sleep Try again later .

If the other side asks, can we not use sleep Well ？

list There's also a directive called blpop, When there is no news , It will block until the news comes .

If the other side continues to ask if it can produce one consumption many times ？

Use pub/sub Theme subscriber mode , Can achieve 1:N Message queue of .

If the other side continues to ask pub/sub What's the downside ？

When consumers are offline , Production messages will be lost , You have to use a professional message queue like RocketMQ etc. .

If the other side is inquisitive TM Questioning Redis How to implement delay queue ？

This is a series of moves , I guess now you'd like to beat the interviewer to death （ The interviewer wanted to kill himself. Why did he ask so many questions he didn't know ）, If you have a baseball bat in your hand , But you are very restrained . Calm down the excited heart , And then he replied with an air of self-respect ： Use sortedset, Take the time stamp as score, The content of the message is key call zadd To produce news , Consumers use zrangebyscore Instructions for N Second before the data polling for processing .

Come here , The interviewer has secretly thumbed you up . And I have given it to you in silence A+, But what he doesn't know is that at the moment you've raised your middle finger , Behind the chair .

Redis How to persist ？ How does the master and slave data interact ？

RDB Mirror full persistence ,AOF Do incremental persistence . because RDB It will take a long time , Not real time enough , During shutdown, a lot of data will be lost , So we need to AOF To work with . stay redis When the instance is restarted , Will use RDB Persistent files rebuild memory , Reuse AOF Replay the recent operation instructions to fully recover the state before restart .

It's easy to understand , hold RDB It can be understood as the data of the whole table ,AOF It is better to understand the log of each operation , When the server restarts, first get all the data in the table , But he may not be complete , You play back the log again , The data is not complete . however Redis Its own mechanism is AOF Persistence is on and exists AOF When you file , Priority load AOF file ;AOF Close or AOF When the file does not exist , load RDB file ; load AOF/RDB Behind the document city ,Redis Successful launch ;AOF/RDB When there is an error in the file ,Redis Failed to start and print error message

The other side asked what would happen if the machine suddenly lost power ？

Depending on AOF journal sync Configuration of properties , If performance is not required , Every time I write an instruction sync Here's the disk , There's no loss of data . But with high performance, every time sync It's unrealistic , Usually use timing sync, such as 1s1 Time , The most you can lose at this time 1s The data of .

The other side asked RDB What is the principle of ？

Just give me two words ,fork and cow.fork Refer to redis By creating subprocesses RDB operation ,cow refer to copy on write, After the subprocess is created , Parent child processes share data segments , The parent process continues to provide read and write Services , Writing dirty page data will gradually separate from subprocesses .

notes ： To answer this question , If you can still say AOF and RDB Advantages and disadvantages , I think I'm the interviewer. I'll give you some compliments on this issue , In fact, there is a big difference between the two , And it's about Redis Data synchronization of cluster and so on . You can also leave a message if you want to know , I will write a special article to introduce .

Pipeline What are the benefits , Why use pipeline？

It can be many times IO Round trip time is reduced to one time , Premise is pipeline There is no causal correlation between the instructions executed . Use redis-benchmark We can find the influence when we do the pressure test redis Of QPS An important factor in the peak value is pipeline Number of batch instructions .

Redis Do you understand the synchronization mechanism of ？

Redis You can use master-slave synchronization , From slave synchronization . The first synchronization , The master node does it once bgsave, At the same time, the subsequent modification operations are recorded in memory buffer, To be completed RDB Full file synchronization to replication node , When the replication node acceptance is complete, it will RDB The image is loaded into memory . After loading , Then inform the main node to synchronize the modified operation records to the replication node for replay, and the synchronization process is completed . The subsequent incremental data passes through AOF Log synchronization , It's a bit like a database binlog.

Have you ever used Redis colony , How to guarantee the high availability of cluster , What is the principle of clustering ？

Redis Sentinal  Focus on high availability , stay master It will automatically slave Upgrade to master, Continued provision of services .

Redis Cluster  Focus on scalability , In a single redis When there is not enough memory , Use Cluster For fragmentation storage .

The end of the interview

Young man, you can , When can I come to work , Or come tomorrow ？

You try to be calm , I still need to rent a house in such a hurry , Maybe next Monday .

well Think this kid is so NB Is it a lot of Offer On the hand , No, I have to call hr Give him more money .

To last , You can't help but praise yourself !

summary

In a technical interview , Whether it's Redis Or something , If you can give a real example , Or directly say that the problems and gains of the development process will add a lot to the impression of the interviewer , The answer should be more logical , Don't be a little East and a little west , It's easy to get dizzy .

Another point is that I ask you why Redis You don't have to answer the questions as soon as you come up , You can answer this way ：

Hello, handsome interviewer , First of all, our project DB There's a bottleneck , Especially the scene of seckill and hot data DB Basically, I can't carry it , Then we need to add caching middleware , At present, there are some cache middleware on the market  Redis  and  Memcached , Their advantages and disadvantages ……, Combine these and then combine the characteristics of our project , In the end, who we chose when we chose the technology .

If you are so methodical , I have answered my questions reasonably and said so many knowledge points beyond my questions , I think you're not just someone who can write code , Your logic is clear , You choose technology , I have my own understanding and thinking about middleware and projects , To put it bluntly, it's your offer There's a play. .

All right, everyone , The above is the whole content of this article , I'll update a few later every week 《 Condole interviewer 》 Series and Java Technology stack related articles . If you have anything to know , Or leave me a message , I'll write it as soon as I have time , We make progress together .