Tearful eyes, it's not easy to change jobs. Three rounds of interviews, four hours of soul torture

I am a graduate student in the science class of double non University ,Java Development 3 Years of working experience , The following is a summary of recent interviews ：

First, let's talk about my interview preparation experience , In order to ensure that your resume has a high probability of passing the screening , I am here 2018 After graduation, I interviewed many companies , I went to a small company in Shanghai and worked until this year 5 month .

Compared with previous years , This year's gold, silver and four don't look so smooth , It may have something to do with my lack of preparation , But to my surprise, the first company to give me an interview was Alibaba . When I first received the interview notice , The mood is very complicated , Nervous and excited , Alibaba is a famous deep excavation project , In fact, my project is relatively weak . This interview , There are three rounds , About four hours in total , These three rounds of interviews are for me , It's soul torture , As for the ending , Look down ！

Ants are basically on the phone . First side leader Noodles , Second face , The third intersection .

First round interview ：1 Hours 28 minute

The interview questions asked by the interviewer ：

• Recent projects
• What problems have you encountered in the process of doing , How to solve
• redis Cluster high availability principle , One is down , How to switch to another
• How to ensure redis and mysql The data is consistent
• Database transactions
• How to optimize large tables
• limit The limitations of ,limit What's the problem when it's worth a lot
• mysql Lock introduction , What are the application scenarios
• Index failure scenario
• The life cycle of a thread
• How to deal with the high thread pool
• In distributed tasks , Many threads of a microservice report errors , How to handle and locate
• JVM Which memory is divided into
• Introduce the garbage collection mechanism
• Did you read JDK Source code habits

Second round interview ：1 Hours 11 minute

1. ThreadLocal principle , On the heap or stack . Have you used it ？
2. Sync Where can it work ？ What's the difference? ？Sync The underlying principle of ？
3. Upgrading a lightweight lock to a heavyweight lock has overhead ？ Is there any way to keep the lock from upgrading ？ Come up with a heavyweight lock ？
4. aop principle , Have you actually used it ？
5. Springboot In the course of using , Where do you find it more convenient ？
6. Yes spring What do you know about the core principles of ？ Will make in-depth questions according to the answers .
7. Spring Instantiation bean Will instantiate all bean Do you ？
8. Spring Do you need to build an instance first when instantiating ？
9. Bean How to solve the circular dependency of ？
10. Why do you use Redis？
11. Redis How do you achieve performance optimization ？
12. From the perspective of underlying implementation Redis Why fast ?
13. on-line Redis What should I pay attention to ?
14. How to make Redis 6.x Maximize performance ?
15. on-line Redis How to choose a persistence solution ? Why? ?
16. Talk about Redis Implementation details of underlying multiplexing
17. Is it used Redis I've done design ?
18. JVM General tuning steps ？
19. Jvm What are your collectors ？ Their advantages and disadvantages ？ And their respective recycling algorithms ？
20. Sql The optimization of the ？sql Optimization of the statement itself ？sql The optimization of the index used ？
21. What will lead to sql The index of is invalid ？
22. The underlying structure of the index and its advantages and disadvantages ？

The third round of interview intersection ：1 Hours 15 minute

1. Project related issues
2. Second kill project needs to consider
3. The second kill scenario uses distributed locks ,
4. How service isolation is done , Assuming that there is no sub database and sub table , How to isolate services
5. database CPU The reason for the high occupancy rate
6. Total disk space 50G, The statistical space is 30G, The remaining 20G May have gone somewhere
7. How to limit current , no need redis, How else can we achieve ( Token bucket , The sliding window )
8. Problems encountered after the project is transformed into a micro service , How to solve ,
9. Implementation principle of distributed things （ I brought it up myself TCC Distributed things ）
10. RocketMQ Ask more ( oooo .. Didn't answer well )
11. stay RocketMQ In the group consumption of , Will each consumer group consume a full message of the theme ？ Why? ？
12. Suppose a RocketMQ The cluster is deployed in two computer rooms , Each machine room has some NameServer、Broker And client nodes , When the link between the two machine rooms is interrupted ,NameServer、Broker Can you provide services normally ？ Why? ？
13. If strict order is not required , In order to improve consumption performance , Can you put Consumer Set to automatically confirm the consumption location , Then batch pull messages into the memory queue , Then asynchronous multithreading executes consumption business logic in parallel ？ Why? ？
14. RocketMQ Can you guarantee that in “ From message production to consumption completion ” In the process , The message is not heavy or lost (Exactly once) ?
15. stay RocketMQ One of the Broker On , Consume messages from a queue using a message sequence number . among , Message file (commitlog) The number is m individual , The number of messages in all message files is k strip , The index file corresponding to the target queue (consumerQueue) The number is n individual , Shared in queue j Bar message . What is the fastest time complexity to find a message ？O(n) still O(1), Why? ？
16. RocketMQ What kind of zero copy technology is used ？
17. RocketMQ How to solve repeated consumption ？
18. Dubbo What are the core components ？ Underlying principle ？Spi What is it? ？spi What are the benefits of
19. What are you learning recently ？ Tell me about SpringCloud alibaba Something about

Summary

Project related issues

Ants don't test algorithms , There are also some tests for basic knowledge , I feel that I pay more attention to the eye edge and the ability to talk about projects .

Big factory interview , It's the process of hanging and being fished . After not meeting a few favorite companies in the second half of last year , I have made up my mind to continue polishing myself . I started a new interview in the first half of this year . But my current resume skills module focuses on my understanding of redis、 Multithreading 、jvm、MySQL These knowledge points are more skillfully mastered （ After the back two sides, I was afraid of being dug at the bottom , So that my resume is now familiar with ）

Another reason is RocketMQ No in-depth study of , Answer haltingly , After talking RockerMQ after , I'll turn to Kafka when , Listen to the interviewer's tone and pause , I guess I'm a little disappointed ！ But this experience gave me a very different feeling

And programmers need to go through layers of interviews when applying . It is said that , Sharpening a knife never misses a woodcutter , Good preparation before the interview can help you better deal with the interviewer's questions and emergencies in the interview .

in general , If you have a big company you want to go to , We must improve ourselves , Match your ability and quality with the company , I've always believed in , Opportunity is always for those who are prepared . The following I often brush the interview manual , Share with you ~

Data acquisition method ：