4cd0f196-cd60-4ec1-bf0d-51e5ec7fdf63

Try setting the engines.node in your package.json

you can also provide a NIXPACKS_NODE_VERSION env var, but you can only set it to major versions: 14, 16, 18, or 20

you can read about it here: https://nixpacks.com/docs/providers/node#setup

@Zerkia pinging for visibility

I will take a look, thanks 🙂

Ended up exposing the file instead and transferring sensitive data as an .env and that seems to have worked.

However, when trying to make an optimized production build, it seems after a while it just gives up and crashes. the error code doesn't really tell me much either that I know of either. Could you perhaps take a look if possible ?

If u send ur build logs here, sure

Sure, one moment

1330900713534132345

This is the error area, I don't know how to copy the full build log, couldn't see a copy button anywhere

this is the start if needed

Cntrl K -> Download Logs

Just to make sure, you aren't commiting your node_modules to the repo, right?

Nono, it's a company app I need to upload a copy of to my exam. Looking up some StackOverflow posts, it may be the CI seeing the eslint and prettier errors as actual errors

yeah, that's a possibility

Trying that real quick, will report back

Completely forgot to report back sorry.

Missing some authentication with sentry which my company makes use of. My supervisor asked if skipping the build portion is possible and going directly to the "npm run start" portion of the deployment ?

Completely forgot to report back sorry.

Missing some authentication with sentry which my company makes use of. My supervisor asked if skipping the build portion is possible and going directly to the "npm run start" portion of the deployment ?

Little update on that: Tried setting the custom build command to "npm run start", but that just seems to have it stuck in "building…" for forever

while actually seemingly running

Albeit the site does not seems accessible due to it not getting to the deploy stage

you cannot start your app during build

build is for building

Yeah I figured, just wanted to try. Is there no way to omit the build stage whatsoever and just have it directly go to "npm run start" or similar ?

you mean omit running npm run build

Yes

I obviously don't know much about your project, but are you sure you can skip it? what does the build script actually run?

"build": "CI=false react-scripts build && npm run sentry:sourcemaps",

sentry being the issue due to auth

you definitely need to run build for react scripts

so you can just replace the build command in the service settings with that

you will also need to copy the Caddyfile and nixpacks.toml from this repo into yours, otherwise you'll be asking Railway to run a development mode server for your app

and sentry probably has some docs about a ci/cd environment

maybe u just miss some extra config?

My supervisor suggested removing the sentry part of the build command and running it build + run like it should be

I haven't touched sentry ever, so this is out of my depth

please do this now

Haven't done that yet, will do so now if still required, just wanted to mention:

It seems to run, deploy and such but then crashes very shortly after, stating:

Starting the development server...

The build failed because the process exited too early. This probably means the system ran out of memory or someone called `kill -9` on the process.

Just wanted to mention it before adding the files, in case I need to do something else as well on the railway server

Are you on railway trial?

or do you have the hobby plan?

this is still needed

I have a railway trial, and I'm wroking on the files now

trial has 4.82 left

It doesn't seem to crash now, after adding the files, but I don't seem to be able to access the page, stating a 502 bad gateway. the railway url is set to run via port 3000

your target port is wrong

Well how do I know what target port is right? locally it ran via 3000

click the edit icon and choose the correct port

It does show 8080 (Caddy), I assume that's the one you're referring to ?

Page seems to be running now, just need to add the new url to the whitelist for my backend, so keeping this open for another moment, just in case something Railway-relevant pops up. If not, I'll let you guys know, thanks 🙂

Does not seem to be talking with the backend properly despite the whitelist. I remember I had a similar issue when trying to deploy the code locally for production, compared to development. Is there any way to have the code running in development on the host at all? Not sure if it's smart or anything of course, but considering I only need to showcase some crud actions that runs through redis and postgresql tomorrow for my exam, would it be possible ?

nope, please do not run a development server, on any cloud server

Does not seem to be talking with the backend properly
please explain more, this amount of information makes it impossible for us to provide support

Of course, I'm trying to troubleshoot it myself at the moment, as it seems to potentially be internal, but I'm trying to ensure that, that is the case first

is there a link we can goto to see the error?

the error happens on the platform itself when trying to log in to it. my api does not seem to be properly reached, and I'm trying to figure out why that is. It's kinda hard to explain as I'm not sure what causes it. My guess is maybe some development mode vs production mode stuff but I'm (very slowly) trying to get contact with my supervisor to get some answers, as this is not something I've developed, and it's the last thing I need for a hosted application.

I have already updated the various areas that needed updates in regards to postgres, redis, front and backend urls', as well as whitelisting them, so I don't think it's something as simple as a bad url when running the flask api request

have you tried looking at your networking tab in the dev tools for your browser?

I have, I see no payload, but the request goes through as a 200. the console showcases a CORS with the api call mentioned

does it go to the correct url though

It does not, trying to figure out where to change it now as well

Thanks for pointing it out, I'm admittedly a little tunnel visioned atm

are you using a REACT_ env var to set the api url?

yes, set as REACTAPPHOSTAPIKEY with a reference to my backend application

that is also hosted on railway

the one you helped me with yesterday

why is that called an API key instead of a url?

an API key is a completely different thing

Don't ask me, I didn't make it. I'm looking at where the apitokens comes from and it's lead me to my company's JWTcontext file, but I can't really seem to see the wrong url being hardcoded or anything anywhere. But my supervisor told me to look there.

None of the url stuff is my code though so it's kinda confusing for me to look through honestly.

In general also kinda hard to help you help me when I don't know how their server is set up really

let me know when you know

Honestly I don't know if I ever will. I can try calling you and screen sharing to try to explain, but honestly their app is huge so it's kinda hard for me to explain since I didn't make this from scratch, but merely added some functionality to it.

I'm sorry but we would not be able to offer that level of support to your class of plan

That's alright

1. update code to call the backend with a url defined in a variable

2. set that variable in your service variables

I managed to get it working thanks to a friend who had some expertise in the codebase.

Turned out to be a multitude of issues:

1. Build folder was not gitignored so it defaulted to an old build

2. some env variables were slightly outdated

3. whitelist had trailing slashes when they shouldn't have had

Passed my exam though, so thank you all for your help, I greatly appreciate it 🙂

I'm marking this as solved, feel free to close it whenever

awesome!

!s