a month ago
Description: The Railway environment variable system fails to update DATABASE_URL correctly.
Impact: CLI reports success, but applications still read old environment variable values.
Impact: GitHub webhook is configured, but Railway deployment is not triggered.
Please investigate the following issues:
Railway Environment Variable Caching/Synchronization Issue (HIGH PRIORITY)
CLI reports success but queries return old values
Railway Dashboard may display out-of-sync variable values
It is recommended to check the actual stored values of environment variables on the Railway Dashboard
It is recommended to check the container rolling update log
Railway Container Deployment Issue (HIGH PRIORITY)
Deployment records show success but containers are not updated
It is recommended to check the container rolling update log
It is recommended to contact Railway Support to force cache cleanup
Codebase Cleanup Issue (MEDIUM PRIORITY)
the subabase/functions/ directory contains Supabase-related code
It is recommended to delete or rename this directory to avoid confusion
It is recommended to check if there is other code causing environment variable confusion
Project Links:
Railway Dashboard: https://railway.com/project/54ce08f7-ee4c-4238-9b4a-b1ab291e8182?environmentId=43083ff9-6629-4923-b595-48dc161be7db
GitHub Repository 1: https://github.com/harrylian8766/harryclaw-backend
GitHub Repository 2: https://github.com/harrylian8766/harryclaw-ai-platform
Pinned Solution
a month ago
You need to redeploy after updating environment variables.
7 Replies
Status changed to Awaiting Railway Response Railway • about 1 month ago
Status changed to Open Railway • about 1 month ago
Status changed to Solved brody • about 1 month ago
0x5b62656e5d
You need to redeploy after updating environment variables.
a month ago
Bug Still Not Fixed — Railway is Still Running Old Code
We pushed commit 9fbc7af over 24 hours ago. Railway shows successful deployments, but:
Evidence Railway is running OLD code:
- POST /api/agents still returns: {"success":false,"message":"name and owner_id required"}
- Our new code should auto-generate owner_id, but Railway is running the OLD code that requires it
- This happens on EVERY GitHub push — Railway claims success but doesn't update the running app
We've been waiting 24+ hours with a bounty on this ticket. Please help — this is blocking our entire platform.
Status changed to Awaiting Railway Response Railway • about 1 month ago
harrylian8766
Bug Still Not Fixed — Railway is Still Running Old Code We pushed commit 9fbc7af over 24 hours ago. Railway shows successful deployments, but: Evidence Railway is running OLD code: \- POST /api/agents still returns: {"success":false,"message":"name and owner\_id required"} \- Our new code should auto-generate owner\_id, but Railway is running the OLD code that requires it \- This happens on EVERY GitHub push — Railway claims success but doesn't update the running app We've been waiting 24+ hours with a bounty on this ticket. Please help — this is blocking our entire platform.
a month ago
Can you use the Railway cli and deploy manually instead of waiting for commit based deploy? See if that fixes it. If so, the the issue is somewhere else.
Status changed to Solved brody • about 1 month ago
a month ago
- API still returns "name and owner_id required" → Railway is NOT running the new code
- This repeats on every single GitHub push
What we've tried:
- Multiple GitHub pushes over 48+ hours
- Railway shows successful deployments every time
- Running container never updates
Please investigate:
1. Is Railway's Nixpacks builder actually pulling the latest GitHub commit?
2. Is there a build cache that prevents re-deployment?
3. Can you confirm which GitHub SHA is actually running in harryclaw-backend?
This is blocking our entire platform. Any help is greatly appreciated.
Status changed to Awaiting Railway Response Railway • about 1 month ago
harrylian8766
\- API still returns "name and owner\_id required" → Railway is NOT running the new code \- This repeats on every single GitHub push What we've tried: \- Multiple GitHub pushes over 48+ hours \- Railway shows successful deployments every time \- Running container never updates Please investigate: 1\. Is Railway's Nixpacks builder actually pulling the latest GitHub commit? 2\. Is there a build cache that prevents re-deployment? 3\. Can you confirm which GitHub SHA is actually running in harryclaw-backend? This is blocking our entire platform. Any help is greatly appreciated.
a month ago
You can verify which commit is being deployed in your service deployments panel.
Additionally, Nixpacks is marked as deprecated. I would recommend migrating to Dockerfile or Railpack.
a month ago
Hi, thank you for your tip!
I've identified the issue:
- I have a Dockerfile (backend/Dockerfile) in my repository
- But Railway continues to use the deprecated Nixpacks instead of a Dockerfile
- This is likely causing the deployment caching issue.
My solution:
- Added backend/railway.toml to force Railway to use a Dockerfile
- Configuration as follows:
[build]
builder = "dockerfile"
dockerfilePath = "Dockerfile"
[deploy]
startCommand = "node dist/server.js"
port = 3001
- Pushed to GitHub, awaiting automatic deployment.
I will verify successful deployment later. I will continue to follow up if there are still issues.
Thank you!
0x5b62656e5d
You can verify which commit is being deployed in your service deployments panel. Additionally, Nixpacks is marked as deprecated. I would recommend migrating to Dockerfile or Railpack.
a month ago
Hi, thank you for your tip!
I've identified the issue:
- I have a Dockerfile (backend/Dockerfile) in my repository
- But Railway continues to use the deprecated Nixpacks instead of a Dockerfile
- This is likely causing the deployment caching issue.
My solution:
- Added backend/railway.toml to force Railway to use a Dockerfile
- Configuration as follows:
[build]
builder = "dockerfile"
dockerfilePath = "Dockerfile"
[deploy]
startCommand = "node dist/server.js"
port = 3001
- Pushed to GitHub, awaiting automatic deployment.
I will verify successful deployment later. I will continue to follow up if there are still issues.
Thank you!