Self modifying code in node.js, would cluster work?

Question

I am asking this since I don t have the tool or time to test this right now, but the idea is bothering me. I ll answer this myself when I ll have the time to play with it.

In node.js, how does require() work? Does it keep the required function in memory? or doest it read the file anew?

Exemple:

launcher.js

var cluster = require('cluster');

if (cluster.isMaster) {
    cluster.fork();
    cluster.on('exit', function () {
        cluster.fork();
    }
}
if (cluster.isWorker) {
    var self = require('self_modifying.js');
    self.start()
}

As long as self_modifying.js have a start() function which is the 'main' method, it could self-update just by modifying it s own source file, and the process.exit(0), and so restart with it new code?

Answer 1

To answer:

In node.js, how does require() work? Does it keep the required function in memory? or doest it read the file anew?

In node.js when a require is performed it will cache the module being loaded so each further require call will load this from memory, rather than from disk as an optimisation. See: http://nodejs.org/api/modules.html#modules_caching

Answer 2

As pointed by @Tom Grant, module are cached. So you need to deference your application before starting it anew, like explained here

This work, but require self_modifying.js to export a function start

var cluster = require('cluster');

if (cluster.isMaster) {
    cluster.fork();
    cluster.on('exit', function () {
        delete require.cache[require.resolve('/full/path/to/self_modifying.js')];
        cluster.fork();
    }
}
if (cluster.isWorker) {
    var self = require('self_modifying.js');
    self.start()
}