a month ago
My Railway Postgres database is stuck in a crash loop because the disk is completely full with WAL files (pg_wal/xlogtemp.*: No space left on device). The database cannot complete recovery because it needs disk space to write WAL, but WAL cleanup (pgbackrest) can't run because Postgres isn't up — a catch-22.
Project: 1c724ce2-33d5-47b1-873a-4a450f76cc44
Service: Postgres (491f508d-dbf2-4634-9799-636e276bb022)
The disk filled up after running ~130,000 UPDATE statements in a migration job. I need either:
Manual WAL pruning from /var/lib/postgresql/data/pg_wal/ to allow recovery
Restore from the last clean backup (before ~11:15 UTC July 16 2026)
Please advise.
1 Replies
a month ago
Your Postgres volume is currently 500 MB and completely full with WAL from the migration, which is why recovery can't complete. You can break this cycle by resizing the volume from the volume settings page - your Hobby plan allows up to 5 GB, and we'd recommend at least 2 GB to give Postgres enough headroom to finish WAL replay and start normally. Regarding a backup restore, PITR was not enabled on this service before the incident so point-in-time recovery is not available; if you have any volume backups, you can restore from those via the Backups tab on your Postgres service.
Status changed to Awaiting User Response Railway • about 1 month ago
Status changed to Solved stavilamax3 • about 1 month ago