I've reported exactly the same issue yesterday.

Hello!

We're acknowledging your issue and attaching a ticket to this thread.

We don't have an ETA for it, but, our engineering team will take a look and you will be updated as we update the ticket.

Please reply to this thread if you have any questions!

Hi there, I have raised the issue to our engineering team.

Hello,

The deployment failed due to unable to pull the image from ghcr.io:
```
We were unable to connect to the registry for this image. If you are using a private image, please make sure your credentials are valid. If this error persists, please reach out to the Railway team.
```
Can you make sure that the credentials you've provided are valid - e.g. PAT is not expired and has read:packages permissions?

Hi @edgnkv, thanks for the response. this one is fixable and understandeable.

can you pls check also deployment id ce8b8eaa-7912-487c-af13-c31618409b94 ?

Heya, same issue there, we're getting a 403 from the ghrc.

and what about 0c21014a-8cef-4d58-a066-0829221178bc ? stuck for 17 minutes in "creating containers" mode. GH token is 100% valid. The fact it doesnt show distinct error (and I have it when token is invalid), is very strange.

Hello,
This deployment failed with 403 Forbidden error - which means the token is valid but it doesn't have permissions. Can you make sure the token has correct scopes selected? See the doc here https://docs.github.com/en/packages/working-with-a-github-packages-registry/working-with-the-container-registry#authenticating-with-a-personal-access-token-classic

I am attaching my token permissions. Again, it works in local scenarios (push/pull). I believe it's railway bug (token is not passed for some reason, that's why there is no error - token is passed on pull stage, but not on deployment) - I am for sure out of context how it's made, but intuition says

afaiu, 403 means that token is passed, but doesn't have permissions. The errors happens on the pull stage (I've redacted user and image):

failed to pull image: failed to resolve reference \"ghcr.io/....\": failed to authorize: failed to fetch oauth token: unexpected status from GET request to https://ghcr.io/token?scope=repository%3A.....%3Apull&service=ghcr.io: 403 Forbidden"

It could be also the token expired, malformed, etc.

why it more looks like "empty token" issue on a kubernetes/docker layer.

- when I pass incorrect token in the UI - railway says me that instantly.
- when I test this token locally - everything is fine.

I am attaching 2 screenshots - first one - malformed token (railway instantly fails), second one - valid token (deployment hangs).

✅ The ticket Deployment delays has been marked as completed.

We've tried reproducing this and are unable to. You're getting a 403 which means to the token you gave us that we pass through in builds doesn't have sufficient scope.

Is your token expired? Can you try regenerating the token?

I recreated the token and it worked.

Very wierd