b2136db9-ae75-4237-ac77-4e079154cc94

please add more context

Dockerfile:

FROM pgvector/pgvector:pg16
RUN apt-get update && apt-get install -y postgis postgresql-16-postgis-3

Docker Compose:

version: '3.1'

services:
  db:
    image: postgis_and_pg_vector:latest
    environment:
      POSTGRES_DB: ${POSTGRES_DB}
      POSTGRES_USER: ${POSTGRES_USER}
      POSTGRES_PASSWORD: ${POSTGRES_PASSWORD}
      RAG_USER_PASSWORD: ${RAG_USER_PASSWORD}
      PGDATA: ${PGDATA}
    ports:
      - "5432:5432"
    volumes:
      - ./init:/docker-entrypoint-initdb.d
      - pgdata:/var/lib/postgresql/data
volumes:
  pgdata:

Volume is mounted at /var/lib/postgresql/data

PGDATA is currently set at /var/lib/postgresql/data/pgdata

railway services can not have two volumes

Gotcha

What approach would you recommend here?

your dockerfile would need to copy the docker-entrypoint-initdb.d file into the applicable location

Okay makes sense, thanks!

but even so, its a init script and will not be ran on every redeploy, it will only be ran if there is an empty data volume

Yeah that's fine. I'm assuming adding a volume is the only way to persist data in a Railway service? Or will it do it by default?

you are correct, volumes are the only way to persist files on disk

can you add some more context though? what does docker-entrypoint-initdb.d do?

Just creates some tables and sets up permissions. I only need it on the original creation of the DB

then yeah, copy it into the correct location, and then wipe the volume to allow it to run again