Docker-Compose postgres upgrade initdb: error: directory "/var/lib/postgresql/data" exists but is not empty

Quoting @yosifkit from this issue

The volume needs to be empty or a valid already initialized postgres database with the file PG_VERSION in there so the init can be skipped.

... If there are any files or folders in there like lost+found it will probably fail to initialize. If there are files that you want to keep in the volume (or have no control over) you could adjust the PGDATA environment variable to point to a sub-directory in there like -e PGDATA=/var/lib/postgresql/data/db-files/.

So I added PGDATA to the environment section of the compose file to solve the issue (notice the some_name at the end):

services:
  postgres:
    image: postgres:12
    environment:
      PGDATA: /var/lib/postgresql/data/some_name/

I got this issue because the /var/lib/postgresql/data/postgresql.conf and the /var/lib/postgresql/data overlap in the docker container at /var/lib/postgresql/.

An example of a broken config is:

version: "3.8"
services:
  db:
    image: "postgres:10"
    ports:
      - "5432:5432"
    volumes:
      - ./postgresql.conf:/var/lib/postgresql/data/postgresql.conf
      - ./pg-data:/var/lib/postgresql/data

To avoid this I tell PostgreSQL to find it's config in /etc/postgresql.conf instead so the overlapping volumes don't occur like this:

version: "3.8"
services:
  db:
    image: "postgres:10"
    command: ["postgres", "-c", "config_file=/etc/postgresql.conf"]
    ports:
      - "5432:5432"
    volumes:
      - ./postgresql.conf:/etc/postgresql.conf
      - ./pg-data:/var/lib/postgresql/data

This is similar to what pmsoltani suggests, but I move the location of the postgresql.conf file instead of the data directory.

I solved it by providing a sub path

      volumeMounts:
        - mountPath: /var/lib/postgresql/data
          name: some-name
          subPath: data

I had the same issue today; I fixed it by removing the content of the volume db_data.

In your case:

docker volume ls
docker volume inspect db_data <-- will show you the mountpoint

I went to the directory (mountpoint) ex: /path/data

cp data data.backup
cd data
rm -R *

And start services:

docker-compose up -d

It pops up as the first search result without approved answer, so let me add my solution here.

I want to provide custom postgresql.conf and pg_hba.conf in /var/lib/postgresql/data where the files are stored by default in image postgres:latest (PostgreSQL 15 as of the time of writing).

There is a directory provided by the image where you can store .sql or .sh files to initialize the database: /docker-entrypoint-initdb.d/.

I COPY crafted postgresql.conf and pg_hba.conf files to /tmp and COPY a short Bash script that will move the configuration files into /var/lib/postgresql/data. It ensures proper order of invocation: first initdb, and then replacement of config files.

I had the exact same issue just a few minutes ago, and the reason (in my case) was that the ./postgres folder in my laptop wasn't empty.

I switched from one branch that was new to an older branch, and that was the reason the ./postgress folder contained different info than the branch I was working, so I just deleted the folder manually, ran again the container and it worked properly.

Here's my docker-compose config for the postgres service:

 postgres:
    image: postgres:16.4
    container_name: postgres-db
    restart: on-failure
    ports:
      - 5432:5432
    environment:
      - POSTGRES_USER=${POSTGRES_USER}
      - POSTGRES_PASSWORD=${POSTGRES_PASSWORD}
      - POSTGRES_DB=${POSTGRES_DB}
    volumes:
      - ./postgres:/var/lib/postgresql/data

I met this error on GCP (GKE) when I was deploying a postgres pod.

The reason was in my badly described PersistantVolumeClaim yaml file.

I just needed to add storageClassName:

apiVersion: v1
kind: PersistentVolumeClaim
metadata:
  name: postgres-pvc
  namespace: nestjs-microservices
spec:
  accessModes:
    - ReadWriteOnce
  resources:
    requests:
      storage: 1Gi
  storageClassName: "standard"

In my case the problem was when I ran in PowerShell from a directory with Russian and other specific symbols:

PS C:\Install\Конфигурация камунда + постгря\PostgreSQL>
docker-compose up -d

The problem disappeared when I ran the same from another directory with usual and latin symbols.

This works:

apiVersion: apps/v1
kind: Deployment
metadata:
  name: postgres
spec:
  replicas: 1
  selector:
    matchLabels:
      app: postgres
  template:
    metadata:
      labels:
        app: postgres
    spec:
      containers:
        - name: postgres
          image: postgres:17
          imagePullPolicy: "IfNotPresent"
          ports:
            - containerPort: 5432
          envFrom:
            - configMapRef:
                name: postgres-config
          env:
            - name: PGDATA
              value: /var/lib/postgresql/data/pgdata
          volumeMounts:
            - mountPath: "/var/lib/postgresql/data"
              name: postgredb
      volumes:
        - name: postgredb
          persistentVolumeClaim:
            claimName: postgres-pv-claim