Starting NodeJS on OpenShift without EADDRINUSE errors - node.js

Question

I am trying to get NodeJS to run on OpenShift, do encounter the same issue though no matter what. Here is an example:
UPDATE with more details <<<<<
FIRST:
This piece of code works fine straight out of Cloud9
var port = process.env.PORT || "127.0.0.1";
var ipaddress = process.env.IP || 8080;
console.log("Getting started here!");
var http = require('http');
var server = http.createServer(function(request, response) {
console.log((new Date()) + ' Received request for ' + request.url);
response.writeHead(200, {'Content-Type': 'text/plain'});
response.write("Welcome to Node.js on OpenShift!\n\n");
response.end("Thanks for visiting us!!!! \n");
});
server.listen( port, ipaddress, function() {
console.log((new Date()) + ' Server is listening on port 8080');
});
console.log("Listening to " + ipaddress + ":" + port + "...");
However as soon as I push it to my OpenShift account changing the port and ipaddress variables to:
var ipaddress = process.env.OPENSHIFT_NODEJS_IP || "127.0.0.1";
var port = process.env.OPENSHIFT_NODEJS_PORT || 8080;
... I am getting this on the OpenShift VM when I SSH over, change to app-root/runtime/repo and run node server.js:
events.js:72
throw er; // Unhandled 'error' event
^
Error: listen EADDRINUSE
at errnoException (net.js:901:11)
at Server._listen2 (net.js:1020:19)
at listen (net.js:1061:10)
at Server.listen (net.js:1129:5)
That would indicate the port is in use already, right? I changed the port to 15550 as I have read somewhere here on stackoverflow I assume that a specific port range is not directly available - only via port forwarding.
Guess what, I just get a different error (EACCES).
Here are the processes running on my OpenShift VM (as far as I can see with my account):
PID TTY STAT TIME COMMAND
177800 ? S 0:00 sshd: XXXX#pts/1
177801 pts/1 Ss 0:00 /bin/bash --init-file /usr/bin/rhcsh -i
210176 pts/1 R+ 0:00 ps ax
SECOND:
I wonder how I can cut down the build process? When I push my code to OpenShift this is what I get:
remote: /bin/sh: warning: setlocale: LC_ALL: cannot change locale (C.UTF-8)
remote: /bin/bash: warning: setlocale: LC_ALL: cannot change locale (C.UTF-8)
remote: /bin/bash: warning: setlocale: LC_ALL: cannot change locale (C.UTF-8)
remote: sh: warning: setlocale: LC_ALL: cannot change locale (C.UTF-8)
remote: sh: warning: setlocale: LC_ALL: cannot change locale (C.UTF-8)
remote: Stopping NodeJS cartridge
remote: Wed Mar 18 2015 07:27:49 GMT-0400 (EDT): Stopping application 'nodejs' ...
remote: Wed Mar 18 2015 07:27:50 GMT-0400 (EDT): Stopped Node application 'nodejs'
remote: /bin/sh: warning: setlocale: LC_ALL: cannot change locale (C.UTF-8)
remote: /bin/bash: warning: setlocale: LC_ALL: cannot change locale (C.UTF-8)
remote: /bin/bash: warning: setlocale: LC_ALL: cannot change locale (C.UTF-8)
remote: Saving away previously installed Node modules
remote: Building git ref 'master', commit e8239d1
remote: Building NodeJS cartridge
The whole process takes like 10 seconds or more. I am not sure what exactly is happening behind the scenes but it seems to be a lot.
So my second questions, is there a way to cut that build time down?
I tried to create an empty file in the .opengit/markers directory as I have read somewhere, but it did not help to improve anything.
Thanks for your help!
z4c

Answer 1

.. I am getting this on the OpenShift VM when I SSH over, change to app-root/runtime/repo and run node server.js
Don't manually run the server yourself. When you push your code to OpenShift, it stops the server, does the build, deploys the code, and starts the server.
So the server is already running. That's why you're getting Error: listen EADDRINUSE try running it yourself. It won't let you run on your own port. That's why you're getting (EACCES) trying to run it on 15550.
If you need to start/stop/restart the server, use OpenShift's rhc command or from the admin console.
For your 2nd question: is there a way to cut that build time down?
The answer is yes. Setting up hot deploy will save you few seconds (by creating an empty file .openshift/markers/hot_deploy) or change the script not to do the build but that's not a good idea. It's doing what it's supposed to do. The right way to cut down the down time to 0 (unless OpenShift is down) is to use the load balancing feature OpenShift offers. It can run 2+ servers at a time. So when you push your code, it stops 1 server while the other servers are still taking requests. Once the 1st server is back up, it'll rebounce the 2nd (and the rest of the servers)