I am currently facing a problem with docker, docker-compose, and postgres that is driving me insane. I have updated my docker-compose with a new postgres password and I have updated my sqlalchemy create_all method with a new table model. But none of these changes are taking affect.
When I go to login to the database container it is still using the old password and the table columns have not been updated. I have run all the docker functions I can think of to no avail
docker-compose down --volumes
docker rmi $(docker images -a -q)
docker system prune -a
docker-compose build --no-cache
After running these commands I do verify that the docker image is gone. I have no images or containers living on my machine but the new postgres image still always is created using the previous password. Below is my docker-compose (I am aware that passwords in docker-compose files is a bad idea, this is a personal project and I intend to change it to pull a secret from KMS down the road)
services:
api:
# container_name: rebindme-api
build:
context: api/
restart: always
container_name: rebindme_api
environment:
- API_DEBUG=1
- PYTHONUNBUFFERED=1
- DATABASE_URL=postgresql://rebindme:password@db:5432/rebindme
# context: .
# dockerfile: api/Dockerfile
ports:
- "8443:8443"
volumes:
- "./api:/opt/rebindme/api:ro"
depends_on:
db:
condition: service_healthy
image: rebindme_api
networks:
web-app:
aliases:
- rebindme-api
db:
image: postgres
container_name: rebindme_db
# build:
# context: ./postgres
# dockerfile: db.Dockerfile
volumes:
- ./postgres-data:/var/lib/postgresql/data
# - ./sql/create_tables.sql:/docker-entrypoint-initdb.d/create_tables.sql
environment:
POSTGRES_USER: rebindme
POSTGRES_PASSWORD: password
POSTGRES_DB: rebindme
#03c72130-a807-491e-86aa-d4af52c2cdda
healthcheck:
test: ["CMD", "psql", "postgresql://rebindme:password@db:5432/rebindme"]
interval: 10s
timeout: 5s
retries: 5
restart: always
networks:
web-app:
aliases:
- postgres-network
client:
container_name: rebindme_client
build:
context: client/
volumes:
- "./client:/opt/rebindme/client"
# - nodemodules:/node_modules
# ports:
# - "80:80"
image: rebindme-client
networks:
web-app:
aliases:
- rebindme-client
nginx:
depends_on:
- client
- api
image: nginx
ports:
- "80:80"
volumes:
- "./nginx/default.conf:/etc/nginx/conf.d/default.conf"
- "./nginx/ssl:/etc/nginx/ssl"
networks:
web-app:
aliases:
- rebindme-proxy
# volumes:
# database_data:
# driver: local
# nodemodules:
# driver: local
networks:
web-app:
# name: db_network
# driver: bridge
The password commented out under POSTGRES_DB: rebindme
is the one that it is still using somehow. I can post more code or whatever else is needed, just let me know. Thanks in advance for your help!
2
Answers
The answer ended up being that the images were still existing. The below command did not actually remove all containers just unused ones:
I did go ahead and delete the postgres data as Pooya recommended though I am not sure that was necessary as I had already done that which I forgot to mention. The real solution for me was:
Then finally the new config for postgres took.
Because you mount volume manually (Host volumes) and when using
docker-compose down --volumes
actually docker doesn’t remove volume. If you don’t need to volume and you want to remove that you have to delete this folder (It depends on the operation system) and then run docker-composeCommand
docker-compose down -v
just remove below volumes type: