I have 3 years of experience in the development, by the end of 18 years began to run away to find work, to get success in the Internet winter Alibaba, headlines today, drops and other companies offer, the post is to develop Java back-end, the final choice went to Ali Baba.
I interviewed a number of companies, visit the company feels most points are the same, so their experiences down, hoping to give'm looking for a friend looking for work or prepare a little help. In addition, Ali also currently doing the interviewer's job seekers from identity becomes the interviewer, look at problems in many ways not the same, so below both the perspective of job seekers, but also the interviewer's perspective
Interview Process
Let me talk about the interview process, large companies generally have a 3-4 round technical side, one side of HR. On Ali, I experienced a total of four technical side, the first two rounds is the main basis for asking and project implementation, the first three are intersecting faces, two interviewers, mainly to ask project implementation and development. The first four are departments face the boss, asked to understand some of the major infrastructure, technology and business, personal development more abstract things.
Note that the point
- Resume one to two is the best, the project experience to write about 2-3 on the same subject, be sure to write the most highlights of the project
- Starting time work experience to write clearly, while large companies have back tone, do not merge or omit some relatively short period of work experience, the impact may not be just the interview, but later may not get into this company
- Blog nothing articles, github project did not write a good resume
For the interviewer's question, to think clearly before you answer, if you feel that the time required may be relatively long, can be said that the idea of my thinking, my next class finishing with the interviewer
During the interview pay attention to speech rate and articulation, after I do the interviewer, found that many people tend to speak quickly, or slurred speech, also cause problems that would otherwise make the interviewer feel that you did not answer the point on (in particular, telephone surface)
Interview after interview results do not directly ask
Technical inspection
Overall, the technology-related study is divided into two blocks, one foundation, the second is experience.
Basis including java-based, database, middleware, from preparation before the daily accumulation and interviews.
Previous experience includes projects done, solve problems, and some scenes problems (such as your project if the traffic flow tenfold how to ensure available).
本文主要说一些大概的问题，整体的进阶的技术点比较多，下面放了下链接，如果失效就可以进群获取，有些是我自己总结的，有些是收（抄）集（袭）了别人的。
set
- The collection is divided into two blocks: Thread safety in non-thread-safe collection and under java.util java.util.concurrent package collection.
List
- ArrayList and LinkedList implementations and differences
Map
HashMap: to understand the data structures, how to hash conflict resolution (list and red-black tree), the timing of expansion, optimization rehash of time to avoid expansion
LinkedHashMap: understand the basic principles, what two orderly, how to use it to achieve LRU
TreeMap: understand the data structure to see why the key target must be to achieve Compare the interface, how to use it to achieve consistent hashing
Set
- Set basically achieved by a corresponding map, look like simple
common problem
- How to hash hashmap resolve the conflict, why hashmap in the list need to turn into a red-black tree?
- hashmap When will trigger the expansion?
- Why concurrency problem when there is infinite loop hashmap before jdk1.8?
- When hashmap expansion need to be calculated once for each entry hash it?
- Why hashmap array length to ensure a power of 2?
- How to implement LRU with LinkedHashMap?
- How to achieve consistency with the hash TreeMap?
Middleware, storage, and other frameworks
Spring
- bean life cycle, circular dependencies, spring cloud (such as projects have used) to achieve the AOP, spring transaction propagation
common problem
- java dynamic proxy cglib difference and dynamic agent (often used in conjunction with the spring so they asked up here)
- spring bean in the life cycle of what?
- What kind of problems have circular dependency injection and constructor injection property?
Dubbo (or other frame Rpc)
RPC framework as a common understanding of Dubbo implementation: service discovery, routing, asynchronous calls, limiting downgrade, failure retry
common problem
- How Dubbo do load balancing?
- How do Dubbo limiting downgrade?
- How Dubbo elegant service offline?
- How Dubbo asynchronous calls?
RocketMq (or other message middleware)
- A common understanding of messaging middleware such as RocketMq implementation: how to ensure high availability and high throughput, message sequence, repeated consumption, transaction messages, delay messages, dead letter queue
common problem
- How RocketMq ensure high availability?
- How RocketMq ensure high throughput?
- RocketMq news is ordered it?
- How to order local news RocketMq is guaranteed?
- Implementation Mechanism RocketMq transaction messages?
- RocketMq be a problem repeated consumption of it? How to solve?
- What level of delayed messages RocketMq support? How to achieve it?
- RocketMq is a push model or pull model?
- Consumer load balancing is kind of how?
Redis (or other caching system)
- redis working model, redis persistence, redis expired elimination mechanism, redis common form of distributed clusters, distributed lock, buffer breakdown, avalanche cache, cache coherency problem
common problem
- Why redis high performance?
- How to use single-threaded multi-core cpu redis machine?
- redis cache elimination strategy?
- How redis persistent data?
- redis which has several data structures?
- redis clusters, which has several forms?
- There are massive key and value are relatively small data, how to store it in a more efficient memory in redis?
- How to ensure data consistency and redis DB in?
- How to solve the cache and cache penetrate the avalanche?
- How to implement a distributed lock with redis?
- Mysql
- Transaction isolation level, locks, data structures, indexes, clustered indexes and non-clustered index, the leftmost matching principle, query optimization (explain commands)
- Recommended Reading:
- http://hedengcheng.com/?p=771
- https://tech.meituan.com/2014/06/30/mysql-index.html
- http://hbasefly.com/ 2017/08/19 / mysql- transaction /
common problem
- Mysql (innondb the same below), which has several transaction isolation level?
- What are the different transaction isolation level locks will be added?
- Mysql row lock, the table lock, the lock gap, intent locks are doing what?
- Talk about what is the leftmost match?
- How to optimize slow queries?
- mysql index Why is b + tree instead of b tree, red-black tree
- How to select the sub-sub-sub-table library table key
- In the case of sub-library sub-table, the query is generally how do sort?
zk
- zk general principle (can understand the principle under similar Raft algorithm), zk achieve distributed lock, zk do cluster master election
common problem
- How to implement a distributed lock with zk, and there are advantages and disadvantages redis distributed lock