jinyung2
commented
3 years ago Time Spent: 4 hours
Took a long time figuring out how to work with running flask applications on the linode server. Initially thought it would be easy to just spin up the flask run script, but this script is actually only intended for development purposes. It is not optimized for a production build. Found out using some kind of WSGI compliant web server is the better way to go, popular option was Gunicorn. Tried to use gunicorn in conjunction with pm2 but was unable to figure out a solution. Tried to use gunicorn by supplying the --daemon flag which did work but wanted a solution that manages the service with a bit more output (instead of having to check processes running on server). Looked at a guide from linode regarding deploying flask apps and it utulized NGINX reverse proxy alongside a program called supervisor. Set up supervisor as per instructions, ended up having to supply a command that initially fires up a bash shell followed by commands to start the venv and start the gunicorn server. Ended up working.
Todo:
- Configure front end application for development/production env variables and supply different commands to the scripts in
package.jsonto supply a production version of the application. - Refactor API calls within the application to have a central location for API calls so there is a single source where the main application needs to trigger to be served up
- Reconfigure react-app to make sure its being served to the correct location (maybe?)
Estimate: 2 hours