Update: PostgreSQL still not deploying. n8n crashing with "connect ECONNREFUSED ::1:5432".

Service variables not expanding correctly. PostgreSQL deployment fails silently.

Need urgent support to restore connectivity or clean up infrastructure.

Project: 7156a8f4-83ff-4f7b-9ec8-182954c75647

Update: the volume with my actual data was deleted — need urgent help checking for a recoverable snapshot

Hi Railway team,

Following up on this thread with a more critical issue than the original disk-space problem.

What happened:

My original PostgreSQL volume (service b5fad79e-a072-4da4-b475-972c1f5ce6e4, volume named postgres-volume) ran out of disk space (a 341MB file in base/16384/16805).

I resized the volume to 5GB, but PostgreSQL kept crashing with PANIC: could not write to file "pg_logical/replorigin_checkpoint.tmp": No space left on device even after the resize appeared to apply — the expansion never actually took effect.

Believing (incorrectly) that my n8n workflows/credentials lived on a separate volume from PostgreSQL, I had the corrupted postgres-volume deleted to recreate it clean. The railway agent chat told me htat there was no issue of deleting the data. Claude says otherwise. you can check the log history

That was wrong — in an n8n + Postgres deployment, workflows, credentials, and execution history live inside the Postgres database itself, not on the n8n service's own volume. So deleting postgres-volume deleted my actual n8n data (~25 workflows, credentials, users).

After that, several fresh Postgres instances/volumes were created and removed while troubleshooting connectivity (postgres-data, volume d38254fa-31cb-48d4-b658-82864e7fe908, postgres-volume-8mPC), ending at the current empty, undeployed service Postgres-Gl0F with volume postgres-volume-AL8O — none of these have my real data, they're all fresh/empty.

What I need:

Is there any way to recover the original postgres-volume (attached to service b5fad79e-a072-4da4-b475-972c1f5ce6e4) from a backend snapshot, soft-delete retention window, or disk image on Railway's infrastructure side? I don't have an external backup of this database, so this is my only path to recovering my workflows. I understand Hobby plan doesn't expose backups in the UI, but I wanted to ask directly in case something is still retrievable server-side before it's purged for good.

Project details:

Project ID: 7156a8f4-83ff-4f7b-9ec8-182954c75647

Environment: production (3b905289-22d3-47f8-9fc6-15e1c0d79917)

n8n service: 24ed0daf-1697-4c1b-a0c1-a229f9d711b7

Original PostgreSQL service: b5fad79e-a072-4da4-b475-972c1f5ce6e4 (volume postgres-volume, deleted)

Domain: n8n-production-121a.up.railway.app

Any urgency you can give this is much appreciated — this is production data I can't reconstruct from scratch.

Thanks.

You can restore deleted volume within 48 hours after deletion, after that, deletion become permanent and can't be restored

You can restore it by using the restoration link sent to your email

and about the original problem? any lead= its seems a common problem of railway

It is probably error because the old postgres is deleted, since you have old volume restored you can mount it to your new Postgres service and edit your n8n variable to use the new DATABASE_URL from fresh Postgres service

from the agenta chat because i cant make it work:

Update - Critical Issue Persists:

After attempting multiple fixes, my n8n and PostgreSQL services are still not working. Here's the current status:

Problem:

PostgreSQL volume filled up completely (341MB file in base/16805)

Created new PostgreSQL service from template with 5GB volume

n8n cannot connect to the new PostgreSQL service

Services are not deploying correctly

Current Errors:

n8n: connect ECONNREFUSED ::1:5432 (trying to connect to localhost instead of PostgreSQL service)

PostgreSQL: No active deployment - won't deploy

Service reference variables like ${{Postgres-Gl0F.PGHOST}} are not expanding

What I've tried:

Removed all corrupted PostgreSQL services

Created fresh PostgreSQL from template

Updated n8n environment variables to reference new PostgreSQL

Triggered multiple deployments - all failed

Variable expansion is not working at infrastructure level

Service IDs:

n8n: 24ed0daf-1697-4c1b-a0c1-a229f9d711b7 (CRASHED)

PostgreSQL: b5fad79e-a072-4da4-b475-972c1f5ce6e4 (Not deploying)

This appears to be a Railway platform issue with service deployment and variable resolution. I need urgent support to restore connectivity or clean up the corrupted infrastructure state.

Are there any Postgres errors?

claude through opus correct it. the proboems were 2. change of version from 16 to 18 in postgres, and also 500 mb file size postgres that colapsed. both corrected. Also the chat onf railway was wrong: it says that i could delete postgres without losing my data.. something that wasnt true and claude showed to me