I'm doing a little OJT on my first node project and, while I can stand up a simple server, the app is going to get hammered so using cluster seems like a good idea. I've cobbled together some code snippets that I've found in various searches (including SO), but the server won't start. I'm sure my inexperience with Node has me doing something stupid, but I don't see it.
var express = require( 'express' );
var cluster = require( 'cluster' );
var path = require( 'path' );
var cCPUs = require( 'os' ).cpus().length;
var port = 3000;
var root = path.dirname( __dirname );
if( cluster.isMaster ) {
for( var i = 0; i < cCPUs; i++ ) {
cluster.fork();
}
cluster.on( 'death', function( worker ) {
console.log( 'Worker ' + worker.pid + ' died.' );
});
}
else {
// eyes.inspect( process.env );
console.log( 'Worker: %s', process.env.NODE_WORKER_ID );
var app = express();
var routes = require( './routes' )( app );
app
.use( cluster.repl( root + 'cluster.repl' ) )
.use( cluster.stats({ connections: true, requests: true }) )
.use( cluster.reload( root ) )
.listen( port );
}
RESULT:
TypeError: Object #<Cluster> has no method 'repl'
If I remove the use
calls, the workers start up correctly, but process.env.NODE_WORKER_ID
is undefined
. Inspecting process.env
shows me that it's definitely not defined. Maybe the snippet I used was from an old version, but I'm not sure how to identify the worker thread in any other way.
How can I unscramble this?
repl
,stats
andreload
methods you're calling don't exist oncluster
. Start with the canonical example from thecluster
docs and go from there instead. – Selfabsorptioncluster
has been around since 0.6, but it was reworked a bit in 0.8. – Selfabsorption