PM2 and Nginx: 502 Bad Gateway

8
votes

I've been trying to deploy my Node project on a brand new DO droplet, but i'm having some problems with PM2.

My steps are a follows:

  1. Node came installed on the Droplet image (Ubuntu, Node v4.4.4)
  2. Installed PM2 globally
  3. Setup Nginx to reverse proxy 127.0.0.1:3000
  4. Cloned my project and did npm install

All i get is Nginx complaining about a 502 Bad Gateway.

If i look at the Nginx error.log i get this:

connect() failed (111: Connection refused) while connecting to upstream, client: client.ip, server: my.server, request: "GET / HTTP/1.1", upstream: "http://127.0.0.1:3000/", host: "my.server"

PM2 doesn't have much to say about anything. Nothing in pm2 logs and status is online.

I tried skipping PM2 and just doing npm start which worked perfectly. I also tried setting up a dummy hello world application instead, and using that with PM2 - it also worked.

So this is currently where i'm at:

  • My project + PM2: doesn't work.
  • My project without PM2: works.
  • Hello World app + PM2: works.

I'm not really sure where to go from here.. I could just skip PM2 and use node, but i do want the features of PM2.

Any ideas?

4
node.jsubuntunginxpm2
Are you sure your app is actually starting on port 3000? – Ben Fortune
Well.. it's probably isn't as Nginx is complaining. I just have no idea how i can check. PM2 says that's everything is all good and online and the logs are clean. – Jacob K
My problem actually turned out to so silly actually, I noticed that 502s were given only for POST requests that do modify the database, which was a sqllite file within the app directory which caused PM2 to restart the app before sending the response when that file was written because it was run with --watch flag. on other cases it can be the --kill-timeout that is too short for you response time, which also causes the app restart. – OsMinOsM

4 Answers

12
votes

I just had to start PM2 with bin/www instead of app.js. Express generator and everything...

0
votes

Nginx has a directive called proxy_read_timeout which defaults to 60 secs. It determines how long nginx will wait to get the response to a request. In nginx. conf file, setting proxy_read_timeout to 120 secs solved our problem.

-1
votes

It may be a problem with pm2 --watch argument if you use it. It may detect updating logs and keeping restarting server.

-1
votes

This is some issue with the node installation. Delete the node_modules folder, install again using npm and start the project using pm2.

pm2 kill

rm -rf node_modules

npm i

pm2 start bin/www