Docker images: how to manage them between development and production - docker

Question

Assuming i have created a DJANGO project using docker-compose as per the example given in https://docs.docker.com/compose/django/
Below is a simple docker-compose.yml file for understanding sake
docker-compose.yml
version: '3'
services:
db:
image: postgres
web:
image: python:3
command: python3 manage.py runserver 0.0.0.0:8000
volumes:
- .:/code
ports:
- "8000:8000"
depends_on:
- db
Here i am using two images python:3 (called "web") and postgres (called "db") which are automatically found from hub.docker.com and build accordingly. We also want web container depends on db container. Just to recollect whats in the docker-compose.yml above
Once i set everything i do docker-compose up and we can see two containers are running and django project is running on my local machine.
Once i have worked with my django application now i want to deploy on the production server.
So how to copy the images to the development server so that i am working on the same docker images again there also.
Because i try to create a docker-compose.yml file at production server thn there will be chance that the db image and web image may change.
Like:
When I build the postgres image on my development computer say i have postgres version 9.5
But If i again build the postgres image on the production server then i may have postgres version 10.1 installed.
SO i will not be working on the same environment, may be on the same os but not the same version of packages.
So how to check this when i am shifting things to development
Partially Solved:
As per the answer of #Yogesh_D,
If i am using prebuilt images from Dokcer hub, we can easily get the same environment on the production server using the version number like postgres:9.5.1 or python:3.
Partially UnSolved:
But If i created an image on my own using my own Dockerfile and then tagged it while building. Now i want to use the same image in production how to do that. Since its not on the Docker Hub and also i may not intereseted to put it on Docker hub.
So will copying manually my image to the production server is a good idea or i just the copy the Dockerfile and again build the image there on the production server.

Answer 1

This should be fairly simple as the docker compose image directive allows you to even specify the tags for that image.
So in your case it would be something like this:
version: '3'
services:
db:
image: postgres:9.5.14
web:
image: python:3.6.6
command: python3 manage.py runserver 0.0.0.0:8000
volumes:
- .:/code
ports:
- "8000:8000"
depends_on:
- db
The above file ensures that you get the 9.5.14 version of postgres and 3.6.6 version of python.
And no matter where you deploy this is exactly what you get.
Look at https://hub.docker.com//python for all the tags/versions available in the python image and look at https://hub.docker.com//postgres to figure out all the versions/tags available for the postgres images.
Edit:
To solve the problem of custom images you have a few ways:
a. depending on where you deploy (your datacenter vs public cloud providers) you can start you own docker image registry. And there are lot of options here like this
b. If you are running in one of the popular cloud providers like aws, gcp, azure, most of them provide their own registries
c. Or you can use docker hub to setup private repositories.
And each of them support tags, so still use tags to ensure your own custom images are deployed just like the public images.