The CGC Knowledge Center

To start with, here's a <b>key</b> for Docker semantics:

<b>IMAGE</b> <br \>

• e.g. rfranklin/rstatsdev <br \>
• a read-only Docker image that you can pull and run <br \>
• can have a TAG (e.g. rfranklin/rstatsdev:0.1) <br \>
• identified by <repository>/<name>[:<tag>]

<b>CONTAINER</b> <br \>

• a running instance of an image that you can read and write to <br \>
• each instance has a unique CONTAINER_ID

<b>COMMAND</b> <br \>

• a general command <br \>
• e.g. /bin/bash

<b>REGISTRY</b> <br \>

• a store for Docker images <br \>
• e.g. Docker Hub or the CGC Image Registry

How to modify a container, commit the changes to a new image, and push that image to DockerHub:

1. Boot up the Docker Quickstart Terminal, this will automatically fire up the docker daemon

2. docker login

   • put in your credentials or use stored credentials if they exist

3. docker images

   • get a list of all your Docker images

4. `docker run -ti <IMAGE>

   • e.g
     docker run -ti rfranklin/pythondev # access command line of container
   • the -i flag allows you to use your new container interactively
   • the -t flag allocates a pseudo-TTY or pseudoterminal (to execute commands)

5. docker ps -a

   • after exiting the container, you can retrieve its id with the above command

6. docker commit <NEW CONTAINER_ID> <IMAGE>

   • commit the container to a new image
   • e.g docker commit f692sd3s4e1b rfranklin/pythondev:v2
   • e.g docker commit f692sd3s4e1b cgc-images.sbgenomics.com/rfranklin/pythondev:v2

7. docker push <IMAGE:VERSION>

   • finally, push your new image

Run a local container interactively:

docker run -ti IMAGE

Look up info about container:

docker inspect IMAGE

Mount a local folder onto your container:

docker run -v /foo:/bar IMAGE

• take /foo in your local machine and make available as /bar in your container

Fix an error with the Docker Daemon:

docker-machine restart default && eval ($docker-machine env default)

Retag an image:

docker images

• get the IMAGE_ID for your image of interest
• docker tag <IMAGE_ID> <IMAGE>
• e.g. docker tag sha:4141532 cgc-images.sbgenomics.com/rfranklin/rstatsdev

Copy local file into container and then save files as a new image:

docker cp PATH/TO/FILE.EXT CONTAINER_ID:/FILE.EXT
docker commit CONTAINER_ID IMAGE:VERSION

Example:

docker cp test.py a4a27b7d1944:/test.py

• this will copy the files to that specific container

docker commit a4a27b7d1944 rfranklin/rstatsdev:v2

• now the changes you made (copying the files over) will persist to this new image. You can run this image to test your analyses.