docs/content/tutorials/compute-to-data-docker-registry.md

title	description
Setting up private docker registry for Compute-to-Data environment	Learn how to setup own docker registry and push images for running algorithms in C2D environment.

Prerequisites

1. Running docker environment on the server.
2. Domain name is mapped to the server IP address.
3. SSL certificate

Step 1: Generate password file

Replace content in <> with appropriate content.

docker run \
  --entrypoint htpasswd \
  httpd:2 -Bbn <username> <password> > <path>/auth/htpasswd

Docker compose template file for registry

Copy the below yml content to docker-compose.yml file and replace content in <>.

version: '3'

services:
  registry:
    restart: always
    container_name: my-docker-registry
    image: registry:2
    ports:
      - 5050:5000
    environment:
      REGISTRY_HTTP_TLS_CERTIFICATE: /certs/domain.crt
      REGISTRY_HTTP_TLS_KEY: /certs/domain.key
      REGISTRY_AUTH: htpasswd
      REGISTRY_AUTH_HTPASSWD_PATH: /auth/htpasswd
      REGISTRY_AUTH_HTPASSWD_REALM: Registry Realm
      REGISTRY_HTTP_SECRET: <secret>
    volumes:
      - <path>/data:/var/lib/registry
      - <path>/auth:/auth
      - <path>/certs:/certs

Start the registry

docker-compose -f docker-compose.yml up 

List images in the registry

curl -X GET -u <username>:<password> https://example.com/v2/_catalog

Other useful commands

Login to registry

docker login example.com -u <username> -p <password>

Build and push image to registry

Use the commands below to build an image from a Dockerfile and push to your own private registry.

docker build . -t example.com/my-algo:latest

docker image tag example.com/my-algo:latest

Next step

You can publish an algorithm asset with the metadata containing registry url, image, and tag information to enable users to run C2D jobs.