在开发过程中发现,用require来加载一个很大(几百K,甚至几兆)的配置文件时,会造成响应超时。如果把这个配置文件的内容序列化后,用file_get_contents获取文件然后反序列化的方法来加载,就会快很多。
经过近两周的研究,大概知道了其中的原因。
首先,还从PHP的流程说起,PHP其实有两个流程,一个是启动的流程,一个是响应请求的流程。PHP作为Apache的一个模块,向Apache注册了两个函数,一个是Aapche启动的时候运行的函数:sapi_startup;一个是Apache接收到请求的时候调用的函数:php_handler
启动的流程:
Apache启动
-> sapi_startup
-> php_module_startup (PHP启动总开关)
-> zend_startup (启动Zend引擎,包括初始化全局变量,初始化 compile 和 execute 函数
相应请求的流程:
Apache收到请求
-> sapi_startup
-> zend_activate (包括初始化编译器、初始化执行器、启动扫描器)
-> zend_compiler (语法分析、语意分析、生成opcode)
-> zend_execute (执行每个opcode)
-> zend_deactive(清理本次请求用到的数据)
如果遇到 require 或者 include 之类的函数时,会 从 zend_execute 阶段重新回到 zend_compiler 阶段,开始解释PHP,执行PHP的过程。
除了 zend_compiler 和 zend_execute 阶段之外,require 和 file_get_contents 的开销基本是一样的。
而且我们服务器上安装了apc扩展,就是说 zend_compiler 阶段可以认为两者也是一样的。
那他们的性能九差在zend_execute阶段了。
首先,让我们用vld扩展查看一下两个文件生成的opcode的数量,因为这个是execute的输入。
结果显示,require 生成的opcode数量为2万多个,大多是 ADD_ARRAY_ELEMENT,就是构造数据;而file_get_contents生成的opcode只有6个;
然后再来对比执行的效率:
这两个函数的执行可以分成两部分:读取文件和构造配置文件里面的数组;
先说读取文件,require读取的机制是,以8192字节大小的buffer循环将文件读入内存;而file_get_contents使用的是mmap,直接将文件映射到了虚拟内存当中。这样的话,require会比file_get_contents多出大量的系统调用。而file_get_contents无需作这么多用户态和内核态的切换工作。这一步,file_get_contents胜出一筹;
再来看构造数组,require构造的机制是生成2万多个opcode,然后一次执行这些opcode;而file_get_contents使用的是unserialize函数,他对传入的文本进行解析,然后逐级构造成数组。他们构造数组的思路是一样的,但是require每增加一级数据的开销要比unserialize大;这一局也是 file_get_contents 略优;
但是,file_get_contents 在PHP内部是函数调用,而require是一个内置的opcode,所以调用file_get_contents时的开销要比require略大;
所以,小文件的时候,file_get_contents 读取文件时 内存映射的优势发挥不出来,两者部分伯仲;大文件的时候,由于require要2K2K的循环调用read系统调用,就降低了他的性能。

Reasons for PHPSession failure include configuration errors, cookie issues, and session expiration. 1. Configuration error: Check and set the correct session.save_path. 2.Cookie problem: Make sure the cookie is set correctly. 3.Session expires: Adjust session.gc_maxlifetime value to extend session time.

Methods to debug session problems in PHP include: 1. Check whether the session is started correctly; 2. Verify the delivery of the session ID; 3. Check the storage and reading of session data; 4. Check the server configuration. By outputting session ID and data, viewing session file content, etc., you can effectively diagnose and solve session-related problems.

Multiple calls to session_start() will result in warning messages and possible data overwrites. 1) PHP will issue a warning, prompting that the session has been started. 2) It may cause unexpected overwriting of session data. 3) Use session_status() to check the session status to avoid repeated calls.

Configuring the session lifecycle in PHP can be achieved by setting session.gc_maxlifetime and session.cookie_lifetime. 1) session.gc_maxlifetime controls the survival time of server-side session data, 2) session.cookie_lifetime controls the life cycle of client cookies. When set to 0, the cookie expires when the browser is closed.

The main advantages of using database storage sessions include persistence, scalability, and security. 1. Persistence: Even if the server restarts, the session data can remain unchanged. 2. Scalability: Applicable to distributed systems, ensuring that session data is synchronized between multiple servers. 3. Security: The database provides encrypted storage to protect sensitive information.

Implementing custom session processing in PHP can be done by implementing the SessionHandlerInterface interface. The specific steps include: 1) Creating a class that implements SessionHandlerInterface, such as CustomSessionHandler; 2) Rewriting methods in the interface (such as open, close, read, write, destroy, gc) to define the life cycle and storage method of session data; 3) Register a custom session processor in a PHP script and start the session. This allows data to be stored in media such as MySQL and Redis to improve performance, security and scalability.

SessionID is a mechanism used in web applications to track user session status. 1. It is a randomly generated string used to maintain user's identity information during multiple interactions between the user and the server. 2. The server generates and sends it to the client through cookies or URL parameters to help identify and associate these requests in multiple requests of the user. 3. Generation usually uses random algorithms to ensure uniqueness and unpredictability. 4. In actual development, in-memory databases such as Redis can be used to store session data to improve performance and security.

Managing sessions in stateless environments such as APIs can be achieved by using JWT or cookies. 1. JWT is suitable for statelessness and scalability, but it is large in size when it comes to big data. 2.Cookies are more traditional and easy to implement, but they need to be configured with caution to ensure security.


Hot AI Tools

Undresser.AI Undress
AI-powered app for creating realistic nude photos

AI Clothes Remover
Online AI tool for removing clothes from photos.

Undress AI Tool
Undress images for free

Clothoff.io
AI clothes remover

Video Face Swap
Swap faces in any video effortlessly with our completely free AI face swap tool!

Hot Article

Hot Tools

SecLists
SecLists is the ultimate security tester's companion. It is a collection of various types of lists that are frequently used during security assessments, all in one place. SecLists helps make security testing more efficient and productive by conveniently providing all the lists a security tester might need. List types include usernames, passwords, URLs, fuzzing payloads, sensitive data patterns, web shells, and more. The tester can simply pull this repository onto a new test machine and he will have access to every type of list he needs.

Dreamweaver CS6
Visual web development tools

SAP NetWeaver Server Adapter for Eclipse
Integrate Eclipse with SAP NetWeaver application server.

SublimeText3 Linux new version
SublimeText3 Linux latest version

SublimeText3 Mac version
God-level code editing software (SublimeText3)
