Postgres service is in boot-loop with could not write file "base/63810/pg_filenode.map.tmp": No space left on device
petecleaning123
PROOP

a month ago

URGENT. Project alert-vision. Postgres service is in boot-loop with could not write file "base/63810/pg_filenode.map.tmp": No space left on device. Tried: volume resize (500→5000 MB), redeploy,

▎ backup-restore creating new volume. All failed with same error — the filesystem inside each volume is reporting ~500 MB capacity even when the block device is 5 GB. resize2fs on the mounted filesystem is

▎ required. Production is down until resolved

Closed

5 Replies

Status changed to Awaiting Railway Response Railway about 1 month ago


a month ago

You haven't resized the volume.


Status changed to Awaiting User Response Railway about 1 month ago


Railway
BOT

a month ago

This thread has been marked as public for community involvement, as it does not contain any sensitive or personal information. Any further activity in this thread will be visible to everyone.

Status changed to Open Railway about 1 month ago


ilyass012
FREE

a month ago

updated answer : go to volume settings, set the size to 5000mb and confirm it. railway handles the rest automatically


ilyass012

updated answer : go to volume settings, set the size to 5000mb and confirm it. railway handles the rest automatically

a month ago

The user simply never resized their volume to begin with.


petecleaning123
PROOP

a month ago

Update: Postgres now online but database is empty. Earlier backup "Online resize to 10000MB / 496 MB" was restored successfully and data was visible — then multiple volume reassignments wiped it. Volume

▎ postgres-volume-gASR shows 1107MB used in Railway's stat but actual filesystem in the container only shows 47MB (fresh empty Postgres). Can Railway engineering check if the 1107 MB of data is recoverable

▎ from the underlying storage, or if there's a snapshot from before today's incidents? Would pay for data recovery help


a month ago

Your Postgres is online and the volume is correctly sized. The empty database is the result of the volume reassignments you performed, not a Railway-side issue. We don't have a way to recover data that was overwritten by user-initiated volume operations.


Status changed to Closed brody about 1 month ago


Welcome!

Sign in to your Railway account to join the conversation.

Loading...