Railway•2y ago

pushing changes in deployment

my project is deployed via github and it writes to the working directory, these changes are supposed to be able to get pushed but when i try git push, it says fatal: not a git repository (or any of the parent directories): .git

Solution:

not the best approach, you would want to be using object storage like cloudflare r2

Jump to solution

54 Replies

Percy•2y ago

Project ID: f1e43e9b-5ebf-4488-a4aa-65930dcaccd1

Brody•2y ago

im sorry but i dont see how issues with git on your local computer has anything to do with railway

epik!OP•2y ago

? i'm talking about railway none of this is on my computer

Brody•2y ago

git push is a command you run on your computer am i wrong there?

epik!OP•2y ago

it's part of my code my code runs git push

Brody•2y ago

why would you want to do that?

epik!OP•2y ago

to sync a specific file with the repo on github?

Solution

Brody•2y ago

not the best approach, you would want to be using object storage like cloudflare r2

epik!OP•2y ago

thanks i will check it out can i integrate it with railway

Brody•2y ago

you would intigrate it with your code using the amazon s3 sdk

epik!OP•2y ago

thanks

Brody•2y ago

https://developers.cloudflare.com/r2/examples/aws/

epik!OP•2y ago

works beautifully, thanks so much

epik!OP•2y ago

used rclone

Brody•2y ago

now hold on are you using rclone in the same fashion that you where using git push?

epik!OP•2y ago

i'm using rclone-python, so no

Brody•2y ago

i would recommend against that, the best way would be using the r2 api directly with boto3 using the api directly with the boto3 package is way less error prone, faster, more portable then calling out to the shell to run rclone commands

epik!OP•2y ago

boto3 seems complicated

Brody•2y ago

it's worth it, trust me

epik!OP•2y ago

i'll give it a shot later @Brody i think i figured it out

epik!OP•2y ago

epik!OP•2y ago

but when i upload the type isn't text/plain

Brody•2y ago

I think you can specify the mime type in extra args

epik!OP•2y ago

hmm

epik!OP•2y ago

contenttype?

Brody•2y ago

content type

epik!OP•2y ago

will try works what about charset?

Brody•2y ago

just set ContentType = text/plain; charset=utf-8

epik!OP•2y ago

oh ok tysm

Brody•2y ago

no problem! you are using environment variables for the storage credentials right?

epik!OP•2y ago

was doing a local test, but for railway i'll use the environment variables

Brody•2y ago

perfect!

epik!OP•2y ago

to download objects, should i use copy_object or get_object?

Brody•2y ago

get

epik!OP•2y ago

ok hmm format is confusing

Brody•2y ago

format?

epik!OP•2y ago

to upload it's s3.Bucket("bucket").put_object(Key="file.txt", Body=open("file.txt", "rb"), ContentType="text/plain; charset=utf-8") i don't know how it is for downloading though

Brody•2y ago

download_object, get_object something along those lines, but that's what the documentation is for

epik!OP•2y ago

i'm getting a AttributeError: 's3.Bucket' object has no attribute 'get_object'. Did you mean: 'put_object'?

Brody•2y ago

link me the documentation you've been using

epik!OP•2y ago

oh figured it out i think https://boto3.amazonaws.com/v1/documentation/api/latest/reference/services/s3.html

Brody•2y ago

https://boto3.amazonaws.com/v1/documentation/api/latest/reference/services/s3/client/get_object.html it's an attribute on the client, not bucket

epik!OP•2y ago

how do i use it then?

Brody•2y ago

scroll down on this page, it shows you all the available arguments

epik!OP•2y ago

oh i had a typo in my filename lol

Brody•2y ago

lol

epik!OP•2y ago

this works

s3.Bucket("bucket").download_file(Key="file.xt", Filename="file.txt")

s3.Bucket("bucket").download_file(Key="file.xt", Filename="file.txt")

Brody•2y ago

perfect

epik!OP•2y ago

s3 is defined as

s3 = boto3.resource('s3',
    endpoint_url = 'https://<accountid>.r2.cloudflarestorage.com',
    aws_access_key_id = '<access_key_id>',
    aws_secret_access_key = '<access_key_secret>'
)

s3 = boto3.resource('s3',
    endpoint_url = 'https://<accountid>.r2.cloudflarestorage.com',
    aws_access_key_id = '<access_key_id>',
    aws_secret_access_key = '<access_key_secret>'
)

Brody•2y ago

now just to swap over to service variables then you can run your app locally with the railway service variables available by using railway run

epik!OP•2y ago

alright can i also set endpoint_url using an environment variable?

Brody•2y ago

yes of course

epik!OP•2y ago

alright railway run works well

Brody•2y ago

awesome

Gaming

Programming

pushing changes in deployment