search
HomeBackend DevelopmentPHP TutorialAnalysis of Swoole and Swoft source code of Swoft

The content of this article is to share with you some introduction to Swoole and Swoft about Swoft source code analysis (Task delivery/scheduled tasks). It has certain reference value and friends in need can refer to it.

Preface

Swoft’s task function is based on Swoole’s Task mechanism, or Swoft The essence of the Task mechanism is the encapsulation and enhancement of the Task mechanism of Swoole.

Task delivery

//Swoft\Task\Task.php
class Task
{
    /**
     * Deliver coroutine or async task
     *
     * @param string $taskName
     * @param string $methodName
     * @param array  $params
     * @param string $type
     * @param int    $timeout
     *
     * @return bool|array
     * @throws TaskException
     */
    public static function deliver(string $taskName, string $methodName, array $params = [], string $type = self::TYPE_CO, $timeout = 3)
    {
        $data   = TaskHelper::pack($taskName, $methodName, $params, $type);

        if(!App::isWorkerStatus() && !App::isCoContext()){
            return self::deliverByQueue($data);//见下文Command章节
        }

        if(!App::isWorkerStatus() && App::isCoContext()){
            throw new TaskException('Please deliver task by http!');
        }


        $server = App::$server->getServer();
        // Delier coroutine task
        if ($type == self::TYPE_CO) {
            $tasks[0]  = $data;
            $prifleKey = 'task' . '.' . $taskName . '.' . $methodName;

            App::profileStart($prifleKey);
            $result = $server->taskCo($tasks, $timeout);
            App::profileEnd($prifleKey);

            return $result;
        }

        // Deliver async task
        return $server->task($data);
    }
}

Task deliveryTask::deliver()Package the call parameters and pass them according to the $type parametersSwoole# The $server->taskCo() or $server->task() interface is delivered to the Task process.
Task itself is always executed synchronously, $type only affects the behavior of the delivery operation, Task::TYPE_ASYNC corresponds to $ server->task() is asynchronous delivery, Task::deliver() returns immediately after being called; Task::TYPE_CO corresponds to $server-> ;taskCo() is a coroutine delivery. After delivery, the coroutine control is given up. Task::deliver() will not return from the coroutine until the task is completed or the execution times out.

Task Execution

//Swoft\Task\Bootstrap\Listeners\TaskEventListener 
/**
 * The listener of swoole task
 * @SwooleListener({
 *     SwooleEvent::ON_TASK,
 *     SwooleEvent::ON_FINISH,
 * })
 */
class TaskEventListener implements TaskInterface, FinishInterface
{
    /**
     * @param \Swoole\Server $server
     * @param int            $taskId
     * @param int            $workerId
     * @param mixed          $data
     * @return mixed
     * @throws \InvalidArgumentException
     */
    public function onTask(Server $server, int $taskId, int $workerId, $data)
    {
        try {
            /* @var TaskExecutor $taskExecutor*/
            $taskExecutor = App::getBean(TaskExecutor::class);
            $result = $taskExecutor->run($data);
        } catch (\Throwable $throwable) {
            App::error(sprintf('TaskExecutor->run %s file=%s line=%d ', $throwable->getMessage(), $throwable->getFile(), $throwable->getLine()));
            $result = false;

            // Release system resources
            App::trigger(AppEvent::RESOURCE_RELEASE);

            App::trigger(TaskEvent::AFTER_TASK);
        }
        return $result;
    }
}

Here is the event callback of

swoole.onTask, its responsibility is only to package the Worker process delivered The data is forwarded to TaskExecutor. The essence of

Swoole's Task mechanism is that the Worker process delivers time-consuming tasks to the synchronized Task process( Also known as TaskWorker) processing, so the event callback of swoole.onTask is executed in the Task process. As mentioned above, the Worker process is the environment where most of your HTTP service code is executed, but starting from the TaskEventListener.onTask() method, the execution environment of the code They are all Task process, that is to say, TaskExecutor and the specific TaskBean are executed in the Task process.

//Swoft\Task\TaskExecutor
/**
 * The task executor
 *
 * @Bean()
 */
class TaskExecutor
{
    /**
     * @param string $data
     * @return mixed
    */
    public function run(string $data)
    {
        $data = TaskHelper::unpack($data);

        $name   = $data['name'];
        $type   = $data['type'];
        $method = $data['method'];
        $params = $data['params'];
        $logid  = $data['logid'] ?? uniqid('', true);
        $spanid = $data['spanid'] ?? 0;


        $collector = TaskCollector::getCollector();
        if (!isset($collector['task'][$name])) {
            return false;
        }

        list(, $coroutine) = $collector['task'][$name];
        $task = App::getBean($name);
        if ($coroutine) {
            $result = $this->runCoTask($task, $method, $params, $logid, $spanid, $name, $type);
        } else {
            $result = $this->runSyncTask($task, $method, $params, $logid, $spanid, $name, $type);
        }

        return $result;
    }
}
The task execution idea is very simple. Unpack the data sent by the

Worker process and restore it to the original calling parameters, and find the corresponding ## based on the $name parameters. #TaskBean and call its corresponding task() method. Among them, TaskBean is declared using the class-level annotation @Task(name="TaskName") or @Task("TaskName"). It is worth mentioning that the

@Task

annotation removes the name attribute, and there is also a coroutine attribute. The above code will be based on this parameter. Choose to use coroutine's runCoTask() or synchronization's runSyncTask() to execute Task. However, because the execution of Swoole's Task process is completely synchronous and does not support coroutines, please do not configure this parameter to true in the current version. Similarly, the task code written in TaskBean must be synchronously blocking or it must be able to automatically downgrade asynchronous non-blocking and coroutine to synchronous blocking according to the environment Deliver the task from Process

We mentioned earlier: The essence of

Task

mechanism of

Swoole is that the Worker process delivers time-consuming tasks to Synchronous Task Process (aka TaskWorker) processing. In other words,
Swoole

’s $server->taskCo() or $server->task() can only Used in Worker process. This restriction greatly limits the usage scenarios. How can I deliver tasks in Process
? SwoftIn order to bypass this limitation, the Task::deliverByProcess() method is provided. The implementation principle is also very simple. The call information is delivered from Process to the Worker process through the $server->sendMessage() method of Swoole , and then the Worker process delivers it to the Task process . The relevant code is as follows: <pre class="brush:php;toolbar:false">//Swoft\Task\Task.php /**  * Deliver task by process  *  * @param string $taskName  * @param string $methodName  * @param array  $params  * @param string $type  * @param int    $timeout  * @param int    $workId  *  * @return bool  */ public static function deliverByProcess(string $taskName, string $methodName, array $params = [], int $timeout = 3, int $workId = 0, string $type = self::TYPE_ASYNC): bool {     /* @var PipeMessageInterface $pipeMessage */     $server      = App::$server-&gt;getServer();     $pipeMessage = App::getBean(PipeMessage::class);     $data = [         'name'    =&gt; $taskName,         'method'  =&gt; $methodName,         'params'  =&gt; $params,         'timeout' =&gt; $timeout,         'type'    =&gt; $type,     ];     $message = $pipeMessage-&gt;pack(PipeMessage::MESSAGE_TYPE_TASK, $data);     return $server-&gt;sendMessage($message, $workId); }</pre>After the data is packaged, use

$server->sendMessage()

After delivering to Worker:<pre class="brush:php;toolbar:false">//Swoft\Bootstrap\Server\ServerTrait.php /**  * onPipeMessage event callback  *  * @param \Swoole\Server $server  * @param int            $srcWorkerId  * @param string         $message  * @return void  * @throws \InvalidArgumentException  */ public function onPipeMessage(Server $server, int $srcWorkerId, string $message) {     /* @var PipeMessageInterface $pipeMessage */     $pipeMessage = App::getBean(PipeMessage::class);     list($type, $data) = $pipeMessage-&gt;unpack($message);     App::trigger(AppEvent::PIPE_MESSAGE, null, $type, $data, $srcWorkerId); }</pre>

$server->sendMessage

, the Worker process will trigger a ## when it receives the data. #swoole.pipeMessage event callback, Swoft will convert it into its own swoft.pipeMessage event and trigger.

//Swoft\Task\Event\Listeners\PipeMessageListener.php
/**
 * The pipe message listener
 *
 * @Listener(event=AppEvent::PIPE_MESSAGE)
 */
class PipeMessageListener implements EventHandlerInterface
{
    /**
     * @param \Swoft\Event\EventInterface $event
     */
    public function handle(EventInterface $event)
    {
        $params = $event->getParams();
        if (count($params) <code></code>swoft. The pipeMessage event is ultimately handled by <p>PipeMessageListener<code>. In the relevant monitoring, if it is found that the </code>swoft.pipeMessage<code> event is generated by </code>Task::deliverByProcess()<code>, the </code>Worker process<code> will execute it once</code> Task::deliver()<code>, finally delivers the task data to the </code>TaskWorker process<code>. </code><code>A simple review exercise: from </code>Task::deliverByProcess()</p> to the final execution of a certain <p>TaskBean<code>, what processes have been experienced, and which parts of the call chain have In which processes are they executed? </code></p><h3 id="从Command进程或其子进程中投递任务">从Command进程或其子进程中投递任务</h3><pre class="brush:php;toolbar:false">//Swoft\Task\QueueTask.php
/**
 * @param string $data
 * @param int    $taskWorkerId
 * @param int    $srcWorkerId
 *
 * @return bool
 */
public function deliver(string $data, int $taskWorkerId = null, $srcWorkerId = null)
{
    if ($taskWorkerId === null) {
        $taskWorkerId = mt_rand($this->workerNum + 1, $this->workerNum + $this->taskNum);
    }

    if ($srcWorkerId === null) {
        $srcWorkerId = mt_rand(0, $this->workerNum - 1);
    }

    $this->check();
    $data   = $this->pack($data, $srcWorkerId);
    $result = \msg_send($this->queueId, $taskWorkerId, $data, false);
    if (!$result) {
        return false;
    }

    return true;
}

对于Command进程的任务投递,情况会更复杂一点。
上文提到的Process,其往往衍生于Http/Rpc服务,作为同一个Manager的子孙进程,他们能够拿到Swoole\Server的句柄变量,从而通过$server->sendMessage(),$server->task()等方法进行任务投递。

但在Swoft的体系中,还有一个十分路人的角色: Command
Command的进程从shellcronb独立启动,和Http/Rpc服务相关的进程没有亲缘关系。因此Command进程以及从Command中启动的Process进程是没有办法拿到Swoole\Server的调用句柄直接通过UnixSocket进行任务投递的。
为了为这种进程提供任务投递支持,Swoft利用了SwooleTask进程的一个特殊功能----消息队列

Analysis of Swoole and Swoft source code of Swoft

同一个项目中CommandHttp\RpcServer 通过约定一个message_queue_key获取到系统内核中的同一条消息队列,然后Comand进程就可以通过该消息队列向Task进程投递任务了。
该机制没有提供对外的公开方法,仅仅被包含在Task::deliver()方法中,Swoft会根据当前环境隐式切换投递方式。但该消息队列的实现依赖Semaphore拓展,如果你想使用,需要在编译PHP时加上--enable-sysvmsg参数。

定时任务

除了手动执行的普通任务,Swoft还提供了精度为秒的定时任务功能用来在项目中替代Linux的Crontab功能.

Swoft用两个前置Process---任务计划进程:CronTimerProcess和任务执行进程CronExecProcess
,和两张内存数据表-----RunTimeTable(任务(配置)表)OriginTable((任务)执行表)用于定时任务的管理调度。
两张表的每行记录的结构如下:

\\Swoft\Task\Crontab\TableCrontab.php
/**
 * 任务表,记录用户配置的任务信息
 * 表每行记录包含的字段如下,其中`rule`,`taskClass`,`taskMethod`生成key唯一确定一条记录
 * @var array $originStruct 
 */
private $originStruct = [
    'rule'       => [\Swoole\Table::TYPE_STRING, 100],//定时任务执行规则,对应@Scheduled注解的cron属性
    'taskClass'  => [\Swoole\Table::TYPE_STRING, 255],//任务名 对应@Task的name属性(默认为类名)
    'taskMethod' => [\Swoole\Table::TYPE_STRING, 255],//Task方法,对应@Scheduled注解所在方法
    'add_time'   => [\Swoole\Table::TYPE_STRING, 11],//初始化该表内容时的10位时间戳
];

/**
 * 执行表,记录短时间内要执行的任务列表及其执行状态
 * 表每行记录包含的字段如下,其中`taskClass`,`taskMethod`,`minute`,`sec`生成key唯一确定一条记录
 * @var array $runTimeStruct 
 */
private $runTimeStruct = [
    'taskClass'  => [\Swoole\Table::TYPE_STRING, 255],//同上
    'taskMethod' => [\Swoole\Table::TYPE_STRING, 255],//同上
    'minute'      => [\Swoole\Table::TYPE_STRING, 20],//需要执行任务的时间,精确到分钟 格式date('YmdHi')
    'sec'        => [\Swoole\Table::TYPE_STRING, 20],//需要执行任务的时间,精确到分钟 10位时间戳
    'runStatus'  => [\Swoole\TABLE::TYPE_INT, 4],//任务状态,有 0(未执行)  1(已执行)  2(执行中) 三种。 
    //注意:这里的执行是一个容易误解的地方,此处的执行并不是指任务本身的执行,而是值`任务投递`这一操作的执行,从宏观上看换成 _未投递_,_已投递_,_投递中_描述会更准确。
];

此处为何要使用Swoole的内存Table?

Swoft的的定时任务管理是分别由 任务计划进程任务执行进程 进程负责的。两个进程的运行共同管理定时任务,如果使用进程间独立的array()等结构,两个进程必然需要频繁的进程间通信。而使用跨进程的Table(本文的Table,除非特别说明,都指SwooleSwoole\Table结构)直接进行进程间数据共享,不仅性能高,操作简单 还解耦了两个进程。

为了Table能够在两个进程间共同使用,Table必须在Swoole Server启动前创建并分配内存。具体代码在Swoft\Task\Bootstrap\Listeners->onBeforeStart()中,比较简单,有兴趣的可以自行阅读。

背景介绍完了,我们来看看这两个定时任务进程的行为

//Swoft\Task\Bootstrap\Process\CronTimerProcess.php
/**
 * Crontab timer process
 *
 * @Process(name="cronTimer", boot=true)
 */
class CronTimerProcess implements ProcessInterface
{
    /**
     * @param \Swoft\Process\Process $process
     */
    public function run(SwoftProcess $process)
    {
        //code....
        /* @var \Swoft\Task\Crontab\Crontab $cron*/
        $cron = App::getBean('crontab');

        // Swoole/HttpServer
        $server = App::$server->getServer();

        $time = (60 - date('s')) * 1000;
        $server->after($time, function () use ($server, $cron) {
            // Every minute check all tasks, and prepare the tasks that next execution point needs
            $cron->checkTask();
            $server->tick(60 * 1000, function () use ($cron) {
                $cron->checkTask();
            });
        });
    }
}
//Swoft\Task\Crontab\Crontab.php
/**
 * 初始化runTimeTable数据
 *
 * @param array $task        任务
 * @param array $parseResult 解析crontab命令规则结果,即Task需要在当前分钟内的哪些秒执行
 * @return bool
 */
private function initRunTimeTableData(array $task, array $parseResult): bool
{
    $runTimeTableTasks = $this->getRunTimeTable()->table;

    $min = date('YmdHi');
    $sec = strtotime(date('Y-m-d H:i'));
    foreach ($parseResult as $time) {
        $this->checkTaskQueue(false);
        $key = $this->getKey($task['rule'], $task['taskClass'], $task['taskMethod'], $min, $time + $sec);
        $runTimeTableTasks->set($key, [
            'taskClass'  => $task['taskClass'],
            'taskMethod' => $task['taskMethod'],
            'minute'     => $min,
            'sec'        => $time + $sec,
            'runStatus'  => self::NORMAL
        ]);
    }

    return true;
}

CronTimerProcessSwoft的定时任务调度进程,其核心方法是Crontab->initRunTimeTableData()
该进程使用了Swoole的定时器功能,通过Swoole\Timer在每分钟首秒时执行的回调,CronTimerProcess每次被唤醒后都会遍历任务表计算出当前这一分钟内的60秒分别需要执行的任务清单,写入执行表并标记为 未执行。

//Swoft\Task\Bootstrap\Process
/**
 * Crontab process
 *
 * @Process(name="cronExec", boot=true)
 */
class CronExecProcess implements ProcessInterface
{
    /**
     * @param \Swoft\Process\Process $process
     */
    public function run(SwoftProcess $process)
    {
        $pname = App::$server->getPname();
        $process->name(sprintf('%s cronexec process', $pname));

        /** @var \Swoft\Task\Crontab\Crontab $cron */
        $cron = App::getBean('crontab');

        // Swoole/HttpServer
        $server = App::$server->getServer();

        $server->tick(0.5 * 1000, function () use ($cron) {
            $tasks = $cron->getExecTasks();
            if (!empty($tasks)) {
                foreach ($tasks as $task) {
                    // Diliver task
                    Task::deliverByProcess($task['taskClass'], $task['taskMethod']);
                    $cron->finishTask($task['key']);
                }
            }
        });
    }
}

CronExecProcess作为定时任务的执行者,通过Swoole\Timer0.5s唤醒自身一次,然后把 执行表 遍历一次,挑选当下需要执行的任务,通过sendMessage()投递出去并更新该 任务执行表中的状态。
该执行进程只负责任务的投递,任务的实际实际执行仍然在Task进程中由TaskExecutor处理。

定时任务的宏观执行情况如下:Analysis of Swoole and Swoft source code of Swoft

相关推荐:

关于PHP中token的生成的解析

Parsing of URL access patterns in TP5

The above is the detailed content of Analysis of Swoole and Swoft source code of Swoft. For more information, please follow other related articles on the PHP Chinese website!

Statement
The content of this article is voluntarily contributed by netizens, and the copyright belongs to the original author. This site does not assume corresponding legal responsibility. If you find any content suspected of plagiarism or infringement, please contact admin@php.cn
PHP: An Introduction to the Server-Side Scripting LanguagePHP: An Introduction to the Server-Side Scripting LanguageApr 16, 2025 am 12:18 AM

PHP is a server-side scripting language used for dynamic web development and server-side applications. 1.PHP is an interpreted language that does not require compilation and is suitable for rapid development. 2. PHP code is embedded in HTML, making it easy to develop web pages. 3. PHP processes server-side logic, generates HTML output, and supports user interaction and data processing. 4. PHP can interact with the database, process form submission, and execute server-side tasks.

PHP and the Web: Exploring its Long-Term ImpactPHP and the Web: Exploring its Long-Term ImpactApr 16, 2025 am 12:17 AM

PHP has shaped the network over the past few decades and will continue to play an important role in web development. 1) PHP originated in 1994 and has become the first choice for developers due to its ease of use and seamless integration with MySQL. 2) Its core functions include generating dynamic content and integrating with the database, allowing the website to be updated in real time and displayed in personalized manner. 3) The wide application and ecosystem of PHP have driven its long-term impact, but it also faces version updates and security challenges. 4) Performance improvements in recent years, such as the release of PHP7, enable it to compete with modern languages. 5) In the future, PHP needs to deal with new challenges such as containerization and microservices, but its flexibility and active community make it adaptable.

Why Use PHP? Advantages and Benefits ExplainedWhy Use PHP? Advantages and Benefits ExplainedApr 16, 2025 am 12:16 AM

The core benefits of PHP include ease of learning, strong web development support, rich libraries and frameworks, high performance and scalability, cross-platform compatibility, and cost-effectiveness. 1) Easy to learn and use, suitable for beginners; 2) Good integration with web servers and supports multiple databases; 3) Have powerful frameworks such as Laravel; 4) High performance can be achieved through optimization; 5) Support multiple operating systems; 6) Open source to reduce development costs.

Debunking the Myths: Is PHP Really a Dead Language?Debunking the Myths: Is PHP Really a Dead Language?Apr 16, 2025 am 12:15 AM

PHP is not dead. 1) The PHP community actively solves performance and security issues, and PHP7.x improves performance. 2) PHP is suitable for modern web development and is widely used in large websites. 3) PHP is easy to learn and the server performs well, but the type system is not as strict as static languages. 4) PHP is still important in the fields of content management and e-commerce, and the ecosystem continues to evolve. 5) Optimize performance through OPcache and APC, and use OOP and design patterns to improve code quality.

The PHP vs. Python Debate: Which is Better?The PHP vs. Python Debate: Which is Better?Apr 16, 2025 am 12:03 AM

PHP and Python have their own advantages and disadvantages, and the choice depends on the project requirements. 1) PHP is suitable for web development, easy to learn, rich community resources, but the syntax is not modern enough, and performance and security need to be paid attention to. 2) Python is suitable for data science and machine learning, with concise syntax and easy to learn, but there are bottlenecks in execution speed and memory management.

PHP's Purpose: Building Dynamic WebsitesPHP's Purpose: Building Dynamic WebsitesApr 15, 2025 am 12:18 AM

PHP is used to build dynamic websites, and its core functions include: 1. Generate dynamic content and generate web pages in real time by connecting with the database; 2. Process user interaction and form submissions, verify inputs and respond to operations; 3. Manage sessions and user authentication to provide a personalized experience; 4. Optimize performance and follow best practices to improve website efficiency and security.

PHP: Handling Databases and Server-Side LogicPHP: Handling Databases and Server-Side LogicApr 15, 2025 am 12:15 AM

PHP uses MySQLi and PDO extensions to interact in database operations and server-side logic processing, and processes server-side logic through functions such as session management. 1) Use MySQLi or PDO to connect to the database and execute SQL queries. 2) Handle HTTP requests and user status through session management and other functions. 3) Use transactions to ensure the atomicity of database operations. 4) Prevent SQL injection, use exception handling and closing connections for debugging. 5) Optimize performance through indexing and cache, write highly readable code and perform error handling.

How do you prevent SQL Injection in PHP? (Prepared statements, PDO)How do you prevent SQL Injection in PHP? (Prepared statements, PDO)Apr 15, 2025 am 12:15 AM

Using preprocessing statements and PDO in PHP can effectively prevent SQL injection attacks. 1) Use PDO to connect to the database and set the error mode. 2) Create preprocessing statements through the prepare method and pass data using placeholders and execute methods. 3) Process query results and ensure the security and performance of the code.

See all articles

Hot AI Tools

Undresser.AI Undress

Undresser.AI Undress

AI-powered app for creating realistic nude photos

AI Clothes Remover

AI Clothes Remover

Online AI tool for removing clothes from photos.

Undress AI Tool

Undress AI Tool

Undress images for free

Clothoff.io

Clothoff.io

AI clothes remover

AI Hentai Generator

AI Hentai Generator

Generate AI Hentai for free.

Hot Article

R.E.P.O. Energy Crystals Explained and What They Do (Yellow Crystal)
4 weeks agoBy尊渡假赌尊渡假赌尊渡假赌
R.E.P.O. Best Graphic Settings
4 weeks agoBy尊渡假赌尊渡假赌尊渡假赌
R.E.P.O. How to Fix Audio if You Can't Hear Anyone
1 months agoBy尊渡假赌尊渡假赌尊渡假赌
R.E.P.O. Chat Commands and How to Use Them
1 months agoBy尊渡假赌尊渡假赌尊渡假赌

Hot Tools

Dreamweaver Mac version

Dreamweaver Mac version

Visual web development tools

DVWA

DVWA

Damn Vulnerable Web App (DVWA) is a PHP/MySQL web application that is very vulnerable. Its main goals are to be an aid for security professionals to test their skills and tools in a legal environment, to help web developers better understand the process of securing web applications, and to help teachers/students teach/learn in a classroom environment Web application security. The goal of DVWA is to practice some of the most common web vulnerabilities through a simple and straightforward interface, with varying degrees of difficulty. Please note that this software

Safe Exam Browser

Safe Exam Browser

Safe Exam Browser is a secure browser environment for taking online exams securely. This software turns any computer into a secure workstation. It controls access to any utility and prevents students from using unauthorized resources.

ZendStudio 13.5.1 Mac

ZendStudio 13.5.1 Mac

Powerful PHP integrated development environment

SublimeText3 English version

SublimeText3 English version

Recommended: Win version, supports code prompts!