From code along by the Net Ninja on youtube. Check out my live site at http://books.anhonestobserver.com Stack: graphql, mongodb atlas, react on aws amplify front end, node on ec2 backend, I'm for hire. email [email protected]

trying https://www.youtube.com/watch?v=Buh3GjHPmjo for cicd from github to ec2 cors anywhere creator is having problems. had to click a button to temporarily enable it. just realized node server had to have a running process for application stop to not crash and theres a delay in the deploment logs pkill kills forever and serve node adds another running server everytime i deploy without pkill deployment stops at application stop if there is pkill but no running node cicd goes all the way until the last step. had to manually go into ec2 and serve client maybe try updating dependencies via npx npm-check-updates -u and npm install no more need for amplify because of cors expiration. so weird. the whole cicd works even though it says timeout during the application start phase npm ci is the key, and old dependencies i guess. the stdout logs for the deployment are not in order. but overall its working. had to put > app.out.log 2> app.err.log < /dev/null & at the end of application start so that deployment succeeds. everything working now. need to make timeout at least 300 seconds because application start takes around 4 minutes just found out the application stop script is from the previous successful deployment https://docs.aws.amazon.com/codedeploy/latest/userguide/reference-appspec-file-structure-hooks.html#appspec-hooks-server it runs the application stop from /opt/codedeploy-agent/deployment-root/deployment-instructions https://docs.aws.amazon.com/codedeploy/latest/userguide/troubleshooting-deployments.html#troubleshooting-deployments-lifecycle-event-failures to troubleshoot application stop troubles

trying https://www.youtube.com/watch?v=oykl1Ih9pMg to take out port at the end of url amazon linux might be like centos to install nginx on ec2, had to do sudo amazon-linux-extras install nginx1 had to do sudo systemctl start nginx to start nginx if not found, check path by echo $PATH, check where nginx is by which nginx if it's not in the path, add nginx to path by adding export PATH="/usr/sbin/nginx:$PATH" to nano ~/.bashrc then load to current shell source ~/.bashrc https://linuxize.com/post/how-to-add-directory-to-path-in-linux/ now when i go to ip address, nginx shows up in etc/nginx/conf.d/ i made a file called books.config i put this in it server { listen 80; listen [::]:80; server_name books.anhonestobserver.com www.books.anhonestobserver.com; location / { proxy_pass http://localhost:5000/; 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; } } from https://medium.com/@ramavathvinayak/deploy-a-node-js-application-on-aws-ec2-with-nginx-as-a-reverse-proxy-e8f41f1edaef then sudo service nginx restart now it really works without the port at the end of the url dont forget the .env or else it will just flash or crash. very important, the client has to point to 13.250.105.63 instead of local host

now trying ubuntu because couldnt get nginx to run just found out i cant copy paste into the terminal from vs code. could work around by copy paste into somewhere else like the search bar, then copy paste into ec2 shell https://www.youtube.com/watch?v=oykl1Ih9pMg https://gist.github.com/bradtraversy/cd90d1ed3c462fe3bddd11bf8953a896 Check NGINX config sudo nginx -t Restart NGINX sudo service nginx restart or should it be sudo systemctl reload nginx to check status, sudo systemctl status nginx NOW IT'S REALLY REALLY WORKING. well not really. forever isnt starting. thats because my path was wrong. now it's REALLY REALLY REALLY WORKING. need to use ubuntu instance, not stupid amazon linux then follow brad traversy didnt do the firewall step though

now for enabling https again following traversy video but certbot didnt work so looking in the comments in the git repo, trying this "As I was using Ubuntu 18.04, I had to follow these instructions before using sudo add-apt-repository ppa:certbot/certbot.

sudo apt-get update sudo apt-get install software-properties-common sudo add-apt-repository universe sudo add-apt-repository ppa:certbot/certbot sudo apt-get update

And then, install certbot using:

sudo apt-get install certbot python-certbot-nginx" still certbot command not found the certificate manager in aws doesnt seem to do anything. google cloud doesnt do anything when trying to create an ssl for books. maybe because i didnt buy the domain name from them. trying to just put https in app.js but doesnt work. says POST https://54.169.202.39:4000/graphql net::ERR_SSL_PROTOCOL_ERROR app engine, firebase, hostgator come with ssl certification, not amazon. trying the free ssl https://certbot.eff.org/lets-encrypt/ubuntufocal-nginx other paid ssls cost $5 to hundreds per year. certificate expires 2021-12-18 but theres auto renew. it will be mixed content https from http so now you have to change client app to https protocol error important commands: sudo systemctl status nginx sudo nano /etc/nginx/sites-available/books.anhonestobserver.com.conf sudo systemctl reload nginx sudo nginx -t sudo service nginx restart followed this https://www.digitalocean.com/community/questions/configure-nginx-for-nodejs-backend-and-react-frontend-app need to put this uri: "https://www.books.anhonestobserver.com/graphql", in apollo in client app, then in ec2 in /etc/nginx/sites-available/default, add another location location /graphql { proxy_pass http://localhost:4000/graphql; } this makes requests that are going to https://www.books.anhonestobserver.com/graphql go to the express server listening to 4000/graphql. other stuff in the default file is server_name books.anhonestobserver.com www.books.anhonestobserver.com; location / { proxy_pass http://localhost:5000; #whatever port your front end app runs on 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; } dont even need the root stuff that locates the html file, because the front end server on 5000 is serving that. everything now really really really really working. no port at the end of the url because of nginx. forever working. https working because of certbot and nginx. cicd working. maybe put proxy option in package.json in client, so apollo wont be the long url, itll just be ':4000/graphql' or something. https://www.newline.co/fullstack-react/articles/using-create-react-app-with-a-server/

NOW MIGRATING TO EC2 WITH DOCTORDB

cant specify a port when pm2 start app.js. only if serving react build folder.

just figured out the .5gb memory server is too small to npm install, so work around is to create a swap file https://hinty.io/devforth/npm-install-killed-without-a-reason-fixed/ npm install works with 1gb memory though.

copy the other conf file with sudo cp /etc/nginx/sites-available/books.anhonestobserver.com.conf sudo nano /etc/nginx/sites-available/books.anhonestobserver.com.conf have to do the sim link thing sudo ln -s /etc/nginx/sites-available/books.anhonestobserver.com.conf /etc/nginx/sites-enabled/ sudo systemctl reload nginx to make sure it's working because certbot was previously installed, it redirected books.anhonest to doctordb.anhonest. to fix, just run sudo certbot --nginx again to expand the certificates. dont be alarmed if the pem name is still doctordb. it still works.

nginx now looks like server { listen 80; root /home/ubuntu/books/client/build;

    index index.html index.htm index.nginx-debian.html;
    server_name books.anhonestobserver.com www.books.anhonestobserver.com;

    location / {
            try_files $uri /index.html;
    }

    location /graphql {
            proxy_pass http://localhost:4000/graphql;
            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;
    }

}

sudo systemctl reload nginx

i guess the workflow is npm run build locally, then push to gh. it should automatically deploy.

pm2 stuff: ss -tnlp | grep "node /" to see what ports pm2 processes are running on.

to run server, with auto restart when files change, put --watch, pm2 start server.js --watch --time --name heat

then to auto restart when instance reboots, pm2 startup pm2 save

to check logs , pm2 logs heat