This time I will bring you an analysisNode.js Detailed explanation of the use of cluster module, detailed explanation of the use of Node.js cluster moduleWhat are the precautions, the following is a practical case, let's take a look take a look.

cluster module overview

The node instance is a single-threaded job. In server-side programming, multiple node instances are usually created to handle client requests to improve system throughput. For such multiple node instances, we call it a cluster.

With the help of node's cluster module, developers can obtain the benefits of cluster services without modifying the original project code with almost no modifications.

The cluster has the following two common implementation solutions, and the cluster module that comes with node adopts the second solution.

Option 1: Multiple node instances and multiple ports

The node instances in the cluster each listen to different ports, and then the reverse proxy implements the request to multiple Distribution of ports.

1. Advantages: Simple implementation, each instance is relatively independent, which is good for service stability.

2. Disadvantages: increased port occupancy, communication between processes is more troublesome.

Option 2: The main process forwards the request to the child process

In the cluster, create a main process (master) and several child processes ( worker). The master monitors client connection requests and forwards them to workers according to specific policies.

1. Advantages: Usually only one port is occupied, communication is relatively simple, and the forwarding strategy is more flexible.

2. Disadvantages: The implementation is relatively complex and requires high stability of the main process.

Getting started example

In the cluster module, the main process is called master and the child process is called worker.

The example is as follows, create server instances with the same number of CPUs to handle client requests. Note that they are all listening on the same port.

//?server.js
var?cluster?=?require('cluster');
var?cpuNums?=?require('os').cpus().length;
var?http?=?require('http');
if(cluster.isMaster){
?for(var?i?=?0;?i?<?cpuNums;?i++){
??cluster.fork();
?}
}else{
?http.createServer(function(req,?res){
??res.end(`response?from?worker?${process.pid}`);
?}).listen(3000);
?console.log(`Worker?${process.pid}?started`);
}

Create batch script: ./req.sh.

#!/bin/bash
#?req.sh
for((i=1;i<=4;i++));?do??
?curl?http://127.0.0.1:3000
?echo?""
done

The output is as follows. As you can see, the responses come from different processes.

response from worker 23735
response from worker 23731
response from worker 23729
response from worker 23730

cluster Module implementation principle

To understand the cluster module, we mainly need to understand three questions:

1. How do master and workers communicate?

2. How to achieve port sharing for multiple server instances?

3. With multiple server instances, how do requests from clients be distributed to multiple workers?

The following will be introduced based on the schematic diagram. For source code level introduction, you can refer to the author's github.

Question 1: How to communicate between master and worker

This question is relatively simple. The master process creates worker processes through cluster.fork(). Cluster.fork() internally creates child processes through child_process.fork().

In other words:

1. The master process and the worker process are the relationship between parent and child processes.

2. The master process and the worker process can communicate through the IPC channel. (Important)

Question 2: How to implement port sharing

In the previous example, servers created in multiple wokers listened to the same port. port 3000. Generally speaking, if multiple processes listen to the same port, the system will report an error.

Why is our example okay?

The secret is that in the net module, the listen() method is specially processed. Depending on whether the current process is a master process or a worker process:

1. master process: listen to requests normally on this port. (No special treatment)

2. worker進(jìn)程：創(chuàng)建server實(shí)例。然后通過IPC通道，向master進(jìn)程發(fā)送消息，讓master進(jìn)程也創(chuàng)建 server 實(shí)例，并在該端口上監(jiān)聽請求。當(dāng)請求進(jìn)來時，master進(jìn)程將請求轉(zhuǎn)發(fā)給worker進(jìn)程的server實(shí)例。

歸納起來，就是：master進(jìn)程監(jiān)聽特定端口，并將客戶請求轉(zhuǎn)發(fā)給worker進(jìn)程。

如下圖所示：

問題3：如何將請求分發(fā)到多個worker

每當(dāng)worker進(jìn)程創(chuàng)建server實(shí)例來監(jiān)聽請求，都會通過IPC通道，在master上進(jìn)行注冊。當(dāng)客戶端請求到達(dá)，master會負(fù)責(zé)將請求轉(zhuǎn)發(fā)給對應(yīng)的worker。

具體轉(zhuǎn)發(fā)給哪個worker？這是由轉(zhuǎn)發(fā)策略決定的?？梢酝ㄟ^環(huán)境變量NODE_CLUSTER_SCHED_POLICY設(shè)置，也可以在cluster.setupMaster(options)時傳入。

默認(rèn)的轉(zhuǎn)發(fā)策略是輪詢（SCHED_RR）。

當(dāng)有客戶請求到達(dá)，master會輪詢一遍worker列表，找到第一個空閑的worker，然后將該請求轉(zhuǎn)發(fā)給該worker。

master、worker內(nèi)部通信小技巧

在開發(fā)過程中，我們會通過 process.on('message', fn) 來實(shí)現(xiàn)進(jìn)程間通信。

前面提到，master進(jìn)程、worker進(jìn)程在server實(shí)例的創(chuàng)建過程中，也是通過IPC通道進(jìn)行通信的。那會不會對我們的開發(fā)造成干擾呢？比如，收到一堆其實(shí)并不需要關(guān)心的消息？

答案肯定是不會？那么是怎么做到的呢？

當(dāng)發(fā)送的消息包含cmd字段，且改字段以NODE_作為前綴，則該消息會被視為內(nèi)部保留的消息，不會通過message事件拋出，但可以通過監(jiān)聽'internalMessage'捕獲。

以worker進(jìn)程通知master進(jìn)程創(chuàng)建server實(shí)例為例子。worker偽代碼如下：

//?woker進(jìn)程
const?message?=?{
?cmd:?'NODE_CLUSTER',
?act:?'queryServer'
};
process.send(message);

master偽代碼如下：

worker.process.on('internalMessage',?fn);

相信看了本文案例你已經(jīng)掌握了方法，更多精彩請關(guān)注php中文網(wǎng)其它相關(guān)文章！

推薦閱讀：

設(shè)計(jì)模式的策略模式怎樣在前端中使用

如何對微信小程序進(jìn)行開發(fā)

The above is the detailed content of Analyzing the use of Node.js cluster module in detail. For more information, please follow other related articles on the PHP Chinese website!