a month ago
Existing PostgreSQL 18 volume is full at 5 GB on Hobby. Startup fails while writing pg_wal with “No space left on device”. No volume backups are available, and the service has no running instance. Please advise a supported way to inspect/recover the existing volume without wiping data, manually deleting WAL/database files, or upgrading the plan. Is volume-file inspection available while the service is crashed?
3 Replies
a month ago
Your Postgres volume is at about 4,985 MB of its 5,000 MB size, and 5,000 MB is the largest volume Hobby allows, so it can't be grown in place on the current plan. Your data is still on the volume.
Volume-file inspection through the CLI (railway volume files) needs the attached service to have an active deployment, so it isn't available while Postgres is crashed.
A plan with a higher volume ceiling would let the volume be resized so Postgres has room to start with the existing data intact.
Status changed to Awaiting User Response Railway • about 1 month ago
Railway
Your Postgres volume is at about 4,985 MB of its 5,000 MB size, and 5,000 MB is the largest volume Hobby allows, so it can't be grown in place on the current plan. Your data is still on the volume. Volume-file inspection through the CLI (`railway volume files`) needs the attached service to have an active deployment, so it isn't available while Postgres is crashed. A plan with a higher volume ceiling would let the volume be resized so Postgres has room to start with the existing data intact.
a month ago
Please have a human review this recovery request. A higher plan is not an option. Your official PostgreSQL troubleshooting documentation describes running the existing container with a temporary sleep infinity start command before inspecting it. Could that allow read-only inspection/download of this existing full volume without starting Postgres or upgrading? Please confirm a supported recovery path; I do not want to delete WAL/database files or reset WAL.
Status changed to Awaiting Railway Response Railway • about 1 month ago
a month ago
Yes, that works for inspection without starting Postgres. Set the Postgres service's Custom Start Command to sleep infinity and deploy. The new deployment keeps the container running with your existing volume mounted, but Postgres never starts, so nothing new gets written to the full disk.
With that deployment running, you can connect to the container with railway ssh and look around the data directory. Because the service now has an active deployment, the CLI's railway volume files commands also work, so you can list and download files from the volume. The volume and its data stay as they are. Don't run pg_resetwal or any of the later WAL-reset steps from that troubleshooting guide, since you've said you want the files left alone.
When you're done, remove sleep infinity from the Custom Start Command and deploy again to go back to the normal start. Postgres will hit the same disk-full error at startup while the volume is still at the 5 GB Hobby maximum.
Status changed to Awaiting User Response chandrika • 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 • about 1 month ago