Question about Node.js Integration and Zero downtime deployment

[tommyc]

Hi all,

I have tried to setup a Node.js application with OpenLiteSpeed. After browsing the doc and reading source code on Github, I found documentation is very limited and have some questions.

In documentation, it recommends serving Node.js application via Web Server Context.

1. Are there any performance difference vs serving via external app and Proxy Context?
2. During deployment, will the existing user being handled with the old process or they disconnected?
3. During deployment, e.g. via scp, will the server route traffic to the incomplete (coping) scripts?
4. In the node.js integration scripts, it seems handle the destination scripts with file description via read_fd.
   • Assuming the project have the below structure, will the application throw error during upgrading dependency?
   • Destination file is app.js and it remains unchange.
     • project_root
       • node_modules
         • express.js
       • app.js (this require express.js)
       • package.json
5. When will the Node.js Application restart, after LiteSpeed server restart only or destination file change? Or it always use the latest file?
6. Will using it with External App (with pm2) and proxy context a better fit than using App Server Context in a production environment?

Many many thanks.