The content of this article is to share some knowledge and experience about PHP interviews. It has a certain reference value. Friends in need can refer to it

• ##CORS (cross origin resource sharing) cross-origin resource sharing, which allows cross-site resource requests. The client needs to use a specific method to request resources (such as xmlhttprequest) and needs to bring specific header information. The server must also support cross-domain resource requests

• memcache can use multi-threading, high throughput, suitable for large visits

• memcache only supports simple key/value structure

• memcache cannot persist data and cannot be backed up. It can only be used for caching. Data will be lost after restarting.

• Because memcache supports multiple Thread operation, so the issue of data consistency must be considered, and cas (check and set) optimistic locking is used

• Since redis is single-formed, it can ensure the orderliness of operations

• Redis supports a variety of data structures, key/value, list, set, zset, dict, etc.

• Redis data can be stored persistently. Store the data in the disk and can read it out when restarting again

• redis supports data backup, that is, data backup in master-slave mode

• Static html, staticize commonly used but rarely updated data

• Separate image servers to avoid Affects the application server. At the same time, individual configuration optimization, cache settings, etc. can be performed for the picture server.

• Database cluster, library table hashing

  
  

• Database cluster can realize reading and writing separation and improve the database response speed. It is generally one master and multiple slaves, or x master and y slave. For example, when writing less and reading more, one server is dedicated to writing operations, and other servers are used for reading operations. This can avoid the impact of read-write locks. However, after the write operation is completed, the data needs to be synchronized.

• Another benefit of a database cluster is that when one database server goes down, there is still complete data on other servers

• In a distributed database, there are multiple nodes in the system, and each node performs different functions. If a node dies, the corresponding function cannot be completed.

• The difference between database cluster and distributed database. Suppose a task takes one hour to complete on one node, and there are now 10 such tasks and 10 nodes. The distributed database can split the task into 10 tasks. Each node completes a different task, regardless of the dependencies of the subtasks. All tasks will be completed after an hour. Each node in the database cluster can complete a complete task. It can average 10 tasks to each node, so that all tasks are completed after one hour.

• sql optimization

• If there is too much data in the table, you can consider splitting the table and then using hash mapping

• Index optimization

• Cache

• Mirror, used to improve access speed

• Load balancing, solving high concurrency and massive access problems

• CDN, allowing users to access the nearest cdn server and get the fastest response

• Data compression

##__set/__get handles attributes that do not exist in the class
#__call/_callStatic is triggered when a method/static method that does not exist in the class is called. __callStatic itself must also be declared as a static method
__toString Convert the object into a string and output it
__invoke Treat the object as a function to execute
The difference between require and include

include function: will read the specified file and execute the program inside;
require function: will read the contents of the target file and replace itself with the read contents;
include_once function: include and Run the specified file. This behavior is similar to the include statement, the only difference is that if the file has already been included, it will not be included again. As the name of this statement implies, it will only be included once;
require_once function: exactly the same as the require statement, the only difference is that PHP will check whether the file has already been included, and if so will not be included again.
Using include_once or require_once will reduce the efficiency of the program, because before loading the file, it will first search whether the file has been loaded in the loaded file table
Include will only generate an alarm when a file that does not exist. When require a file that does not exist, a fatal error will be thrown directly and the script will stop.
include is a conditional inclusion function, and require is Unconditionally contains functions. That is, if there is an if condition in the outer layer of include or require, when if is false, include will not include it. It will only include the file when it is true, and require will include the file regardless of whether it is true or false.

Why can’t you use parentheses when returning?

• return is just a language structure, not a function. There is no need to enclose the return value in parentheses. Reduce efficiency

• If the return value does not provide parameters, null will be returned at this time, and parentheses must not be included at this time. If you include parentheses, it will cause a parsing error

• You must not use parentheses when returning a reference to a variable. Otherwise, the value of the reference will be returned instead of the reference itself

PHP weak type variable mechanism

fast_cgi

php_fpm

php5 php7 difference

The above is the detailed content of Sharing some knowledge and experience about PHP. For more information, please follow other related articles on the PHP Chinese website!