Skip to content
Dockerfile rejected before build with no logs, but the same recipe passes Docker and CI
androvcr
OP

17 days ago

We are preparing an optional, isolated ClamAV scanner in an existing staging worker. Railway rejected the new Dockerfile during validation, before the image build or application startup, while the same recipe passed local Docker checks and the full GitHub Actions gate. The previous worker stayed active; we have now deployed the standard worker recipe successfully with attachment reception and scanning disabled. Could you provide the exact validation error and the instruction or rule that caused the rejection?

Resources and timing:

  • Project: 58e38b73-d350-4de5-acca-45fd7bbededf.
  • Environment: staging, 39ca9265-434b-42b1-ba2a-2d6881990eb5.
  • Service: worker, ed2db57b-fb7b-442c-877e-dc163f2b6908.
  • Failed deployment: 5a91cd11-438e-4700-93fd-2d3a21c924fe, shown as 2026-11-02 23:09 CST (UTC-06:00), failed after about 6 seconds.
  • Previously active deployment: 2cd6962f-4ba7-4d8d-8ac2-d2d7f054c3ff.
  • Current standard-worker deployment: 5e0095fe-6c24-4575-8cde-17133567ead1, ACTIVE with worker started in Deploy Logs, observed 2026-11-03 05:28 UTC.

The failed deployment used /docker/Dockerfile.worker-clamd-railway from commit e557afa2aa56ed9e8fc3135cf9dea237b91cc4d6:

https://github.com/AndroVcr/reqnex-proyectos/blob/e557afa2aa56ed9e8fc3135cf9dea237b91cc4d6/docker/Dockerfile.worker-clamd-railway

Railway Details shows:

The Dockerfile failed validation.

The Build Logs tab shows:

No build logs were found for this deployment.

Evidence and attempts:

  • The same Dockerfile recipe passed a local Docker build and docker build --check, including harmless CSV, JPEG and PDF checks.

  • Full CI for the exact commit passed, including this Dockerfile, native-reader isolation, Chromium PDF, real fictitious formats through ClamAV, and the container dependency policy:

    https://github.com/AndroVcr/reqnex-proyectos/actions/runs/41577404147

  • The failure was not retried. We changed only the Dockerfile path back to /docker/Dockerfile.worker and deployed the standard recipe successfully. Both recipes are present in the same reviewed commit.

  • No scanner started, no new email was received, and no customer documents or credentials are included here. Attachment reception and scanning remain disabled.

The AI assistant helping with this setup suggested contacting Railway because the dashboard provides only a generic validation error and no build logs. It could not establish a specific cause from the available checks, and it did not have an authenticated Railway CLI session. We are not claiming a platform bug or that Dockerfile syntax, source-image policy, or fetching has been ruled out.

Please inspect the validator result for the failed deployment and explain what must change. This is a diagnostic request only: please do not redeploy, modify or delete resources, execute SQL, edit the repository, or activate automatic AI diagnosis. We want to avoid blind rebuilds and additional full CI runs.

Prepared with Railway's support template (v1).

$20 Bounty

4 Replies

Status changed to Awaiting Railway Response Railway • 17 days ago


Railway
BOT

17 days ago

This thread has been opened as a bounty so the community can help solve it.

Status changed to Open Railway • 17 days ago


Could you share your Dockerfile?


androvcr
OP

17 days ago

Thanks for looking into this. The repository is private, so the GitHub links in the original post are not accessible publicly; I should have made that clear.

I have attached the exact Dockerfile from the failed deployment as a text file. It contains the deployment recipe and references to helper files, without including application source or credentials.

File: Dockerfile.worker-clamd.support.txt

SHA-256: f282299ba5d0398cdf81f3ad552c356a1852ad37b56696136a9fc50b57816cf9

Source commit: e557afa2aa56ed9e8fc3135cf9dea237b91cc4d6

The failed deployment is 5a91cd11-438e-4700-93fd-2d3a21c924fe. Railway rejected it during validation before producing build logs. The standard worker from the same commit is running, and the attachment/scanner gates remain off.

Could you check which validation rule rejects this recipe? We have not retried the failed build or changed its network/integrity safeguards.


Try removing the —-network=none from your RUN entries.


antonio313

15 days ago

There's a previous thread that matches the issue that you're having:

https://station.railway.com/questions/repeated-pre-docker-build-image-failure-2edd51a3

The TLDR is as 0x5b62656e5d said, drop the --network=none flag but only on the Railway variant and try again


Welcome!

Sign in to your Railway account to join the conversation.

Loading...