a month ago
Account email: altafunnelltda@gmail.com
Project: energetic-gentleness (ID: 4fad3ed5-79ab-47ab-8cd1-3061f027c90d)
Environment: production (ID: 855afa4f-3e3f-4f98-a46a-cbccc3dd0ec7)
Service: Postgres (ID: 5760c4d3-fba9-4f30-ac0d-1fafa53d0ba6)
Volume: postgres-volume (ID: afce939e-2c90-41d4-b33c-bb345fdf8eb4)
We have lost all production data from our PostgreSQL database and need help
determining if it can be recovered.
Timeline:
-
Our application was writing to the database normally through at least
2026-08-29 (last confirmed good data).
-
8 days ago (~2026-08-23), the Postgres service had an automatic deployment
("initiated by an auto update", image ghcr.io/railwayapp-templates/postgres-ssl:18).
Railway automatically created a one-time snapshot at that time, labeled
"Pre-Security-Patch Backup" (98 MB), which is the only backup we have —
we are not on the Pro plan, so there is no recurring backup schedule.
-
As of today (2026-08-31), the database has NO TABLES at all (Database →
Data tab shows "You have no tables"), and the postgres-volume usage graph
is effectively at 0 B. Our application expects a single table
dadoswith one JSONB row (id='main'), which no longer exists.
Our app logs show no database connection errors around the time of loss —
the app appears to still be "Online" and connected, it's the underlying
data/table that disappeared.
Questions:
-
Is there any snapshot, WAL archive, or replica (even outside the normal
Backups UI, since we're not on Pro) that could let you recover data more
recent than 2026-08-23 — ideally as close as possible to 2026-08-29/30?
-
Can you tell us what happened to the volume/database between
2026-08-29 and now — was there an incident, a volume replacement, or
manual action tied to the auto-update deployment 8 days ago?
-
If nothing more recent is recoverable, we'd like confirmation that
restoring the "Pre-Security-Patch Backup" (2026-08-23) is our best
available option before we proceed with that restore ourselves.
This is a small business production system (scheduling/CRM data) and the
loss is affecting daily operations. Any help investigating quickly is
greatly appreciated.
1 Replies
a month ago
The security patch on 2026-08-22 created a pre-patch volume snapshot, which is visible in the Postgres service's Backups tab. That snapshot is your best available recovery point. The patch itself restarts the container against the same volume and does not remove data, so the tables were not lost by the auto-update. Our databases are unmanaged, so automatic snapshots, WAL recovery, and PITR aren't available to restore your data unless you've enabled them on the service, meaning no copy newer than that 2026-08-23 snapshot exists. Restoring the snapshot requires upgrading to Pro ($20/month), which unlocks the restore you run yourself from the Backups tab. Going forward, upgrading to Pro also enables native volume backups and point-in-time recovery on your Postgres service so you can self-restore if this happens again.
Status changed to Awaiting User Response Railway • about 1 month ago
a month ago
This thread has been marked as solved automatically due to a lack of recent activity. Please re-open this thread or create a new one if you require further assistance. Thank you!
Status changed to Solved Railway • 28 days ago