My TimescaleDB service (vol_jr384w0k76pc4qnu) is crash-looping and cannot start. After a redeploy, the container can no longer write to its

volume mount:

Mounting volume on: /var/lib/containers/railwayapp/bind-mounts/.../vol_jr384w0k76pc4qnu

mkdir: cannot create directory '/home/postgres/pgdata/data': Permission denied

The image is timescale/timescaledb-ha:pg17, which runs as uid 26 (postgres). The mount point /home/postgres/pgdata appears to be root-owned, so

the entrypoint cannot create or write to the data subdirectory.

I attempted a custom start command to chown the mount before the entrypoint runs, but chown also fails because the container is not running as

root:

chown: changing ownership of '/home/postgres/pgdata': Operation not permitted

chown: cannot read directory '/home/postgres/pgdata/lost+found': Permission denied

The service ran stable on this same image for 3 days prior. Only a redeploy was triggered — no image change, no volume resize, no config

changes to the service itself.

Project: dexfolio

Service: timescaledb

Volume ID: vol_jr384w0k76pc4qnu

Environment: production

What I need:

Chown the volume mount point to uid 26:gid 26 (postgres) on the platform side so the Timescale HA image's entrypoint can write to it

OR remount the volume with the correct ownership applied at mount time

I cannot shell into the container remotely, and no in-container workaround succeeds because the process is never root.