See files generated deployment

[uncut yew]

I was able to get my site deployed on railway but I'm trying to access a file that is generated after deployment. Is it possible to see it (maybe with Railway CLI)?

[vague juniper]

Project ID: N/A

[south sleet]

To help others find answers, you can mark your question as solved via Right click solution message -> Apps -> ✅ Mark Solution

[uncut yew]

N/A

To elaborate, when executing supervisord, all the logs of the child processes will be in a seperate file generated after starting supervisord

im wondering if its possible to view the contents of those files

[spark spire]

for all intensive purposes, it's not possible.
configure supervisord to log to stdout/stderr

[uncut yew]

oh yea the second part would also be useful

what exactly is stdout/stderr through

is that the deploy log?

I can already see the output of supervisor here though

i just want to see the output of the child processes run by supervisor

[spark spire]

I'm sure there's some results from Google that could explain it far better than I could

[spark spire]

configure supervisord to forward the logs of the child process to stdout/stderr

[uncut yew]

oh i see

and how do I view stdout/stderr from railway?

[spark spire]

the deployment logs

[uncut yew]

ok great

thanks i will research into this

[uncut yew]

ok i figured out the command but what would be the directory I should forward the logs to?

I need to set

stdout_logfile

and stderr_logfile

[spark spire]

i dont know off the top of my head, that would be something you could look up though

me telling you would just be me googling it 🤣

[uncut yew]

ah ok

ill try lol

it seems like supervisor only has an option to redirect stderr 😭

theres no option for redirect_stdout

[spark spire]

i googled it

stdout_logfile=/dev/stdout

[uncut yew]

wait what

where did you find that information

[spark spire]

uh in the second link of my google search

[uncut yew]

hmm my stdout is just completely empty now

[uncut yew]

oh i just realized this is actualyl the original code i had

[uncut yew]

so thats why I was able to see this

[spark spire]

new ideas:

• take my original advice and run two railway services, forego supervisord
• use parallel instead of supervisord
• use multirun instead of supervisord

[uncut yew]

I'm trying to get the sdtout of the child processes

[spark spire]

🙂

[uncut yew]

will that allow me to view the stdout of child processes from the deployment log?

[spark spire]

all those options will

the first option being the best

[uncut yew]

hmm wdym by two railway services

what processes am I splitting up

[spark spire]

the things you are trying to get supervisord to run

[uncut yew]

hmm ok i will try

currently its just huey (celery) and the flask app

[spark spire]

yeah a railway service for the flask app and a railway service for the huey thing

[uncut yew]

wait but then I wouldn't need supervisor anymore

[spark spire]

correct

[uncut yew]

since i would just be starting each seperatly

ahh

[spark spire]

therefore no issues with logs

[uncut yew]

i see what you mean

thanks!

[spark spire]

to make it easy, i would have a Dockerfile.huey and a Dockerfile.flask that hopefully are self explanatory, then specify the appropriate dockerfile via the RAILWAY_DOCKERFILE_PATH service variable

[uncut yew]

but just for future reference its not possible to view files deployed inside railway?

[uncut yew]

so for example I wouldn't be able to see the contends of childlogdir

[spark spire]

unless you build that kind of functionality into your app, no

[uncut yew]

ok i see

thanks again!