The new Postgres cluster initialized at /home/postgres/pgdata/data instead of the volume mount at /var/lib/postgresql/data. The volume (timescaledb-volume) previously held ~5GB of data. The lost+found directory at the old mount point is dated Feb 24. Can you check if the volume still has data or if there are infrastructure snapshots from before Apr 9?

Hey, you mentioned that it previously had around 5GB of data, did you check if the volume metrics (volume usage) has changed? I suppose you can SSH into your service and see if the data is still there using the Railway CLI.

Alternatively, since you are on the pro plan, check if you have any volume backups that you may restore to.

If you happen to still have the data, make sure to backup the volume and check that you have a service variable like PGDATA="/var/lib/postgresql/data/pgdata".

I SSH'd in. The current volume mounted at /var/lib/postgresql/data is /dev/zd12512 (434MB) and is nearly empty. this appears to be a NEW volume created during redeployment, not my original. My original DB was ~5GB. I can see many unmounted 4.7G block devices (zd0, zd96, zd160, zd176, zd224, etc.) via lsblk. My old data is likely on one of these unmounted disks. Can you identify which block device was my original volume and remount it? I don't have any backups to restore to.

We do not maintain infrastructure-level snapshots of volumes. The only recovery path is through the Backups feature, if you had manual or scheduled backups configured before April 9.

The unmounted block devices visible via lsblk belong to the underlying host infrastructure and are not your previous volume. We cannot identify or remount them. Your volume was not replaced with a new one during redeployment - the same volume mount was overwritten when the new TimescaleDB image ran its database initialization at the mount point. Without pre-existing backups, the overwritten data unfortunately cannot be recovered from our side.