Get Deployd running on Heroku!

package.json

{
  "name": "app_name",
  "version": "0.0.1",
  "description": "Super awesome app",
  "keywords": [],
  "homepage": "",
  "author": "You!",
  "contributors": [],
  "dependencies": {
    "deployd": ">= 0"
  },
  "scripts": {
    "start": "node server"
  },
  "engines": {
    "node": "0.8.x",
    "npm":  "1.2.x"
  },
  "license": "MIT"
}

Procfile

web: node server

server.js

// require deployd
var deployd = require('deployd');

// configure database etc. 
var server = deployd({
  port: process.env.PORT || 5000,
  env: 'production',
  db: {
    host: 'something.mongolab.com',
    port: 27857,
    name: 'database_name',
    credentials: {
      username: 'username',
      password: 'password'
    }
  }
});

// heroku requires these settings for sockets to work
server.sockets.manager.settings.transports = ["xhr-polling"];

// start the server
server.listen();

// debug
server.on('listening', function() {
  console.log("Server is listening on port: " + process.env.PORT);
});

// Deployd requires this
server.on('error', function(err) {
  console.error(err);
  process.nextTick(function() { // Give the server a chance to return an error
    process.exit();
  });
});

Great work. I don't knew where to "configure" the io object. Thanks for sharing

Very useful, thanks!

Thank you!

Thanks for the script!

For websockets on heroku it is possible to enable websockets (beta)
heroku labs:enable websockets

I have problems with memory leaks. Every get request takes up to 15 MByte and never frees the memory.
To see the memory usage I enabled this option on heroku:

heroku labs:enable log-runtime-metrics

and then

heroku logs --tail

When i try this i get
remote: ! Push rejected, no Cedar-supported app detected
! [remote rejected] master -> master (pre-receive hook declined)

Anyone else get this?

@robotnic did you solve your memory leaks issue? In affirmative case, how?

Thanks

FYI, I had trouble having this working. I solved my issues using

 "engines": {
    "node": "0.10.x",
    "npm":  "2.2.x"
  }

And I also had to create a "resources" folder in the root of my heroku app folder.
Note that the resources folder, has it is empty might be ignored by git, so I added an empty file into it.

I hope it help someone has I lost a couple of hours solving this.

I created a demo to use Deployd on Heroku with the deploy button:
https://github.com/NicolasRitouet/deployd-heroku-demo
It doesn't work with the key yet.
The resources folder is not needed anymore, if Deployd doesn't find it, it will create it.
I haven't tested everything, but the basic features seems to work.

TY @NicolasRitouet

Not able to run the server. How can I access my apis and dashboard?

If I understand this correctly the following may not work in socket.io 1.0 and later...

` server.sockets.manager.settings.transports = ["xhr-polling"];`

In any case that line throws an error for me: Cannot read property 'settings' of undefined.

The socket.io docs say to do this...

var socket = require('socket.io')({
  // options go here
});

But I'm not clear how to apply that approach in a script like the one above.

Any advice is very much appreciated.

---

UPDATE:
Got it working by replacing line 20 with...

server.sockets.server.set('transports', ["xhr-polling"]);

These were also suggested in a stackoverflow answer, but I haven't tried them yet...

var server = deployd({
    socketIo: {
        options: { transports : ['xhr-polling'] }
    }
});

or

server.sockets.server.eio.transports = ['xhr-polling'];

got this working just now (4/22/18) using the following server.js:

var deployd = require('deployd');

// configure database etc.
var server = deployd({
  port: process.env.PORT || 5000,
  env: 'production',
  db: {
    connectionString: "mongodb://<username>:<password>@<hostname>:<password>/<db name>"
  }
});

server.sockets.server.set('transports', ['xhr-polling']);

// start the server
server.listen();

// debug
server.on('listening', function() {
  console.log("Server is listening on port: " + process.env.PORT);
});

// Deployd requires this
server.on('error', function(err) {
  console.error(err);
  process.nextTick(function() { // Give the server a chance to return an error
    process.exit();
  });
});```

Also seemed to work without using the "server.sockets.server.set" line - unsure if that's truly needed or not, but kept it in anyways.