I deploy a nodejs application on the aws beanstalk servers and want to use socket.io feature based on WebSocket protocol. I know there's a discussion here to directly connect to nodejs servers instead of using nginx as an proxy server. But if I still want to have the nginx as proxy server because of extra features provide by nginx, such as static files, ...etc.
I find it's already support WebSocket proxying on nginx 1.3.13 and I found it seems aws elastic-beanstalk still use the 1.2.x nginx.
So I am wondering if there's any way to upgrade nginx version under beanstalk and how to enable WebSocket proxying to nodejs server.
Thanks
We use elastic beanstalk with multiple docker containers(allows you custom nginx version) with following
1.Nginx config
location /ws/
{
proxy_http_version 1.1;
proxy_set_header Upgrade $http_upgrade;
proxy_set_header Connection "upgrade";
proxy_pass http://unix:/<<socket>>;
}
Enable TCP mode load balancing in elastic load balancer if you are using one.
You would need additional module enabled, which can be done during nginx compilation.
To do that you would need to add below line to your configuration script.
--add-module=/root/nginx_patched/nginx_tcp_proxy_module
It is required if you would like to get sockets enabled, for example for node.js socket.io. Full tutorial can be found here.
Sorry for the link but it is quite broad topic. You might need step by step guide if you starting from the scratch.
Hope it helps.
Related
I wanted to move a project I have working on Windows to run on a Unix machine.
The machine running on Debian 9 with Nginx.
This project runs absolutely fine on Windows with IIS.
I've followed all the instructions on here created a service for this to run on the start of the machine and Nginx configuration to proxy the connection from the port I want to use to port 5000.
When I start the application running Dotnet Myddl.dll it starts and says it is only listening on port 5000.
Then when I try to access it, I can see a warning.
warn: Microsoft.AspNetCore.HttpsPolicy.HttpsRedirectionMiddleware[3]
Failed to determine the HTTPS port for redirect.
I know it is related to my app redirecting to HTTPS and not knowing where to redirect it, but how do I resolve this?
My service
[Unit]
Description=Myapp API
[Service]
WorkingDirectory=/var/www/myapp/publish
ExecStart=/usr/bin/dotnet /var/www/myapp/publish/myapp.dll
Restart=always
# Restart service after 10 seconds if the dotnet service crashes:
RestartSec=10
KillSignal=SIGINT
SyslogIdentifier=dotnet-example
User=www-data
Environment=ASPNETCORE_ENVIRONMENT=Production
Environment=DOTNET_PRINT_TELEMETRY_MESSAGE=false
[Install]
WantedBy=multi-user.target
server {
listen 6969;
server_name mysite.net *.mysite.net;
location / {
proxy_pass http://localhost:5000;
proxy_http_version 1.1;
proxy_set_header Upgrade $http_upgrade;
proxy_set_header Connection keep-alive;
proxy_set_header Host $host;
proxy_cache_bypass $http_upgrade;
proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
proxy_set_header X-Forwarded-Proto $scheme;
}
}
EDIT:
I've been trying to resolve this and still can't. When I start the app on the unix machine I get the following
root#myhost:/var/www/myapp/publish# dotnet Myapp.dll info: Microsoft.Hosting.Lifetime[0] Now listening on: localhost:5000 info: Microsoft.Hosting.Lifetime[0] Application started. Press Ctrl+C to shut down. info: Microsoft.Hosting.Lifetime[0] Hosting environment: Production info: Microsoft.Hosting.Lifetime[0] Content root path: /var/www/myapp/publish
Obviously it is missing https option, and I can't figure why.
EDIT2:
I've published the app as self contained for Linux-x64, and now I do not get the warning saying that it couldn't determine the https port, on my browser I get redirected to https://mydomain:5001 when I access it on http://mydomain:6969
Still I do not get the app listening on https on Unix, just on Windows.
EDIT3:
Noticed that if I go to one of my endpoints e.g. http://IP:6969/api/users I get a 500 response.
EDIT4:
When I was loading my application locally, I was getting straight through to the swagger page, such as /swagger/index.html, for some reason my API when complied for Linux does not accept this URL and returns me a 404, but if I get to one of my endpoints e.g. /api/users, it does return me the data I was expecting for.
Default https port should be 5001.
You can set https port following offical docs.
Or disable relatied middlware. Use nginx for https termination if needed.
I was trying to Migrate an API, and as on Windows it was returning me to my swagger page located on /swagger/index.html I was expecting the same to happen, which for some reason, doesn't.
So if I access one of my endpoints (e.g. /api/users) it does work fine.
I have installed Nginx for Windows (64-bit) from here because the official binaries are 32-bit. The aim is to use Nginx for load balancing NodeJS applications. I am following instructions from here where, the link to sample basic configuration file also exists.
The following configuration file works successfully on Linux where nginx was installed via Ubuntu PPA. The servers are themselves started via pm2.
map $http_upgrade $connection_upgrade {
default upgrade;
'' close;
}
upstream top_servers {
# Use IP Hash for session persistence
ip_hash;
# Least connected algorithm
least_conn;
# List of Node.JS Application Servers
server 127.0.0.1:3001;
server 127.0.0.1:3002;
server 127.0.0.1:3003;
server 127.0.0.1:3004;
}
server {
listen 80;
server_name ip.address;
location /topserver/ {
proxy_set_header X-Real-IP $remote_addr;
proxy_set_header Host $http_host;
proxy_pass http://top_servers;
proxy_set_header X-Request-Id $request_id;
}
}
However, this file does not work with Windows. I am getting an error as 'No such file or directory' under the html folder of Nginx installation on Windows. I haven't done any such setting for Linux.
Can you please help me convert the above configuration file for Windows?
NOTE I don't have a choice - Windows is a must for this project.
So, I over-wrote the contents of conf/nginx.conf with the contents shown above. First, I got an error as "map" directive is not allowed here. Then, after removing this directive, I got another error as "upstream" directive is not allowed here". I think, the binaries I am using does not support load balancing.
Locally, I use ExpressJS on port 3001 and then start my react app with npm start which runs the development server on port 3000. This allows me to route requests as a proxy from 3000 to 3001.
For production, I installed Ubuntu NodeJS 6.12.13 on 16.04 on a DigitalOcean Droplet and then installed Nginx and PM2.
In my Nginx default file I have set the following:
location / {
proxy_pass http://localhost:3001;
proxy_http_version 1.1;
proxy_set_header Upgrade $http_upgrade;
proxy_set_header Connection 'upgrade';
proxy_set_header Host $host;
proxy_cache_bypass $http_upgrade;
}
I've moved over my Express and React setup and added the Express server to the PM2 startup. Nginx is being used as a reverse proxy server to use Express on port 3001. Here is the PM2 startup (www is the name of the server file which runs Express).
When I load my domain, I receive the Express default page:
Now I'm not sure how to start the react app, because it doesn't seem logical to start it using npm start and keep the terminal open for a production server. I need to see my React app when I visit the domain instead of the Express message.
I've found articles which mention to use npm run build but they don't explain how to then run the React app. Sorry I'm new to this, but any help would be appreciated. Thank you.
You won't run the React app because there is no such a thing :) After building your app all your files bundled in a single Javascript file. You are using start for your React app in development for development purposes.
After doing:
npm run build
you will have a build directory in your app directory. Just copy all the files and directories from this build directory to your server where your Nginx's default directory points.
If you don't want to open your regular app codes in developer tools of browsers, delete build/static/js/some_file.js.map and build/static/css/some_file.css.map before uploading your files to server. Those are source map files which are for debugging purposes. If you include them, in developer tools everyone can see you files directly. Your code actually open to world, to anybody right now but with a bundled, uglified and minified way. If you include source map files, they will be opened as they are.
This is how you run a static app. Without a backend, means here without Express, just using a web server.
But, since your question involves Express I assume you are using a backend server. So, one method is copying all your project to your server again with all backend and frontend code as you are using in development. Build your React app. But this time instead of starting both an Express server and React development server, on your server you will only run Express. Express will be the one serving your frontend. You should have already configured this in your development and done some production tests.
So, if you don't use a backend server you don't need Express or any other thing apart from a single web server. If you use a backend server then you need something like Express to serve both your backend requests (like to API's) and your React app. In addition you will need something like PM2 to run Express and optionally Express to use proxies for different apps.
I have a node js application in Elastic Bean stalk.We are considering using socket io for a feature .
I read in some places that socket io support has to be manually enabled in AWS elasticbeanstalk. Specifically when it uses the default NGINX proxy.
I read By default an elastic beanstalk instance has an nginx proxy in front of it that is not configured to allow webSockets.
Is this correct information? If so , how to enable socket io support in AWS EB?
This is correct information. You'll need to do some additional configuration for your Elastic Beanstalk deployment to get WebSockets(Socket.io or otherwise) to work.
Once you create your Elastic Beanstalk Environment, you'll need to configure your load balancer to accept TCP connections, and add a configuration file to your node project's root directory:
Configure Load Balancer:
Head over to your EC2 console and select the Load Balancers tab
Select the load balancer that belongs to your ELB environment from
the list
Select the Listeners tab
Change the default entries' Instance Protocol to TCP
Add the Configuration File:
In the root directory of your node project, create a folder called
.ebextensions
Create a file called enable-websockets.config in your new .ebextensions folder with the following contents:
container_commands:
enable_websockets:
command: |
sed -i '/\s*proxy_set_header\s*Connection/c \
proxy_set_header Upgrade $http_upgrade;\
proxy_set_header Connection "upgrade";\
proxy_pass_request_headers on;\
' /tmp/deployment/config/#etc#nginx#conf.d#00_elastic_beanstalk_proxy.conf
This file tells the NGINX reverse proxy how to handle the HTTP 101 upgrade status code that WebSockets need to communicate with your application server.
I've been trying to get Ghost.io installed on my web server for quite sometime. I have a VPS with Centos 6 and Cpanel.
Today I found a script at http://www.allaboutghost.com/one-click-ghost-install-script/ that said you could just enter a command into your ssh terminal and have it all installed for you.
Command
wget -O - https://raw.github.com/howtoinstallghost/installghost.sh/master/installGhost.sh | sudo bash
I did this and it appears to have worked, I didn't get any errors but now I am not able to find the install in either FileZilla or by using my web browser. The website says that it installs in the /var/www/ghost/ directory but I can't find that. If I use cd /var/www/ghost/ in the ssh it takes me right to it and even lets me edit the config.example.js file.
If I direct my browser to www.mydomain.com:80 since the site says it installs on port 80 it just takes me back to my home page.
What am I missing and what do I need to do?
As Per the comments, I did follow the instructions on the github page. Now All I get when I visit mydomainname.com/ghost/
Ghost is installed really easy. You'd better don't use 3rd party scripts for that as it might really vary from system to system. All you need is to get Node.js installed and then follow instruction for example from here: https://www.digitalocean.com/community/tutorials/how-to-install-and-configure-ghost-on-ubuntu-16-04. They are very detailed and must work for Centos as well.
Most common errors are:
- Not configuring a reverse proxy (nginx or apache) to link to your ghost install on port 2368. Here is an example for Nginx:
server {
listen 80;
server_name your_domain_or_ip_address;
location / {
proxy_set_header HOST $host;
proxy_set_header X-Forwarded-Proto $scheme;
proxy_set_header X-Real-IP $remote_addr;
proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
proxy_pass http://127.0.0.1:2368;
}
}
If you now see default home page it means some web server is already running and intercept all the requests and parks them to a default location (e.g. /var/www)
- If you want Ghost to be the one and the only web server on your VPS, you have to remove or shut down currently installed web server and try to configure ghost to answer on port 80 like this:
server: {
host: '0.0.0.0',
port: '80'
}
I didn't test it but has to work. This kind of install is not recommended and insecure. I suppose you can configure reverse proxy from Cpanel, but not sure.
The best and easiest way to set up Ghost is using SSH.
Hope it helps. For any further help, you have to provide more details and possibly logs and configs. Most of the possible errors you'll catch when installing or starting your blog with npm
sudo npm install --production
sudo npm start --production
Good luck with your deployment.
Hey you don't need to do that much you can just launch ghost from digitalpress for free...
Know how to host ghost blog for free out here : https://treanches.digitalpress.blog/hosting-ghost-blog/
You can think this as self promotion but this article is so much understanding you won't regret reading it