#metabrainz

Hi bitmap: yes the production.toml stays in the config folder of root dir.

For sentry, you may have to add a project URL. If you want to not use it, I'll add a

* Hi

@bitmap:chatbrainz.org: yes the production.toml stays in the config folder of root dir. For sentry, you may have to add a project URL. If you want to not use it, I'll add a callback option for it

* Hi

@bitmap:chatbrainz.org: yes the production.toml stays in the config folder of root dir. For sentry, you may have to add a project URL. If you want to not use it, I'll add a fallback option for it

Also, in .env file add RUN_MODE=production so that the app can identify which toml file to read

yellowhatpro: thanks, it'd be best if the project could run without a Sentry instance (ours doesn't even work currently). regarding `RUN_MODE=production`, shouldn't that already be set in the Dockerfile.prod image?

okk, I actually was adding it in docker-compose file

I shall add it in dockerfile itself

btw, bitmap how is the app being run in prod?

I am a bit clueless about it

I'm building an image from Dockerfile.prod and running that directly

ohh okay

the config file is to be mounted into the container when it's run

yeah right, and that was my next question

basically I was thinking to add gha for pushing image to dockerhub

but for that, I would need the production.yaml too

so I was unsure how we will add the config when the gha will be running

A couple other small things I noticed:

* If you're running the project outside of Docker, are you supposed to edit the files under config/ directly? If so, I'd change these to example files, and gitignore the real ones (otherwise changing the config will cause unstaged changes in the repo)

* I shouldn't have to specify `database_url` in the config, the app should construct it from the other settings :)

Ohh cool

I'll rename them as example

yellowhatpro[m]: I'd just require that a config be mounted in the container

it'll mainly be used in production anyway, and it doesn't make much sense to run without providing your own config

of course you could still have a default config in the image that connects to local services on their standard ports

umm a bit of confusion, how will we mount the config while we build it in github action?

you don't need to mount a config file when building the image

coz I was thinking to push the image to dockerhub

sure, just copy a default config into the image

I mean it to work with prod config

for example, if you want to publish the prod image of the app

to dockerhub

idk i might be sounding stupid

or pushing prod image to dockerhub is not the case we are dealing with

that's definitely useful, we do the same for other prod images. but we'll store the prod config file in a private/internal repo, not in the image itself (as it contains a secret key for accessing the wayback machine API)

ohh

i see

so that's when you mount files

yes, we have a script in the same private repo which executes the docker run command and has access to the config file

you have private repos in github?

correct