a month ago
We need a consistent offline export/snapshot of Railway volume 2955b41f-e218-41c4-b52a-33417b3dfb86 from project 0fdfc7e3-f5f4-49eb-9ccc-b61c5debc331.
The service is stopped/failed and railway volume files list/download cannot initialize the SFTP session and repeatedly times out.
We specifically need a consistent SQLite snapshot containing tradingos.db, tradingos.db-wal, and tradingos.db-shm from /var/lib/tradingos.
We do not want to restore a backup, start the application, or modify the volume. We need the files for offline read-only forensic analysis.
Could you either provide an offline export or restore read-only railway volume files download access for this stopped-service volume?
3 Replies
a month ago
This thread has been opened as a bounty so the community can help solve it.
Status changed to Open Railway • about 1 month ago
a month ago
You’ll need to either detatch the volume and attach it to a Filebrowser instance, or deploy the current service to access files thru the Console tab.
a month ago
Thanks. We tested the Filebrowser approach, but detaching and reattaching the volume automatically triggered deployments of the original service, so we cannot use that method safely.
We now have a dedicated one-shot forensic process that opens the existing SQLite database strictly read-only (mode=ro, PRAGMA query_only=ON) and performs no migrations, maintenance, or writes.
The remaining problem is simply how to execute that command against the existing volume of a stopped service without starting the service itself.
What we need is the Railway-supported equivalent of:
mount existing volume → run one custom command once → no normal start command → restart policy NEVER → exit
Specifically:
existing volume: 2955b41f-e218-41c4-b52a-33417b3dfb86
existing mount: /var/lib/tradingos
original service must remain stopped
no detach/reattach of the volume
no normal service deployment/startup
one command only
no automatic restart
Is there a supported way to run a one-off container/job against the volume already attached to a stopped service, or to atomically override the start command for exactly one deployment without ever executing the service's normal start command?
If not, is there a supported way to obtain an offline snapshot/export of the volume while the service is stopped?
We don't need database assistance or data recovery — only safe access to the existing volume without starting the application.
a month ago
Right after clicking “redeploy” when detaching the volume, you can navigate to the service to abort the build immediately to avoid starting the service.