3 Using Containers

Let’s actually dive into to using containers. You can follow along with this activity with either Docker or Podman. We recommend using Docker but if for admin reasons you are unable to use Docker, Podman’s commands work nearly identically to Docker.

3.0.1 A note about “window juggling”

When learning another software skill like containers, it sometimes means you’ll have to keep track of more windows. We call this “Window Juggling”. It’s an overlooked skill but something you’ll become more comfortable with.

IMPORTANT: Use the checkboxes to follow along with each step in the activities! This is to help you keep track because steps have to be done and in order!

3.1 Activity Instructions

3.1.1 Docker

3.1.1.1 Set up

Go here to install Docker, following the instructions for your particular operating system. If you don’t have a Docker account create an account when prompted, or go here. After you install Docker, start up Docker desktop by double clicking on the app. It may take some time to start up.

3.1.1.2 Get workshop files

Download the files for this activity clicking here: https://github.com/fhdsl/containers-for-scientists-sandbox/archive/refs/heads/main.zip Put this file on your desktop so it is easily findable. Double click the zip file (or right click and choose “unzip” or “decompress” to unzip the file.

3.1.1.3 Activity 1: Pull and run a Docker image

This activity is built so that you will encounter errors that this guide will work you through! Most common container stumbling block is understanding the idea that containers are isolated from your computer

Which means: If your computer has files, software, or anything else. Your container by default does NOT have it unless you do something to get those files and software on there.

We will encounter both an error of when we don’t have a file we need AND an error of when we don’t have a software package we need. We will guide you through what to do when these expected errors occur.

3.1.1.4 Step 1: Docker pull

We need to get the Docker image we want to use. We do this by “pulling” it. Pulling is a term used in GitHub terminology too. It just means we’re taking something from the internet and getting a copy locally (on our computer).

Run this command in your Terminal or Command Prompt window:

docker pull cansav09/practice-image:1

3.1.1.4.1 Step 2: Check what images you have

To see what images we have we can run docker image ls. We should see cansav09/practice-image:1 show up in the output.

3.1.1.4.2 Step 3: Run the docker image to start a container!

Remember images can be shared easily but to actually use the stuff we need to turn our image into a container! To do that we will do docker run

To launch the image for use, we can use docker run to turn it into a container.

docker run cansav09/practice-image:1

To see what containers we have running we have we can run:

docker ps

Pat yourself on the back! You have a running container!

What can we do with running containers?

3.1.1.4.3 Step 4: Run a single command

To run stuff interactively from the command line we can do:

docker exec -it <PUT_CONTAINER_ID_HERE> bash

If you run this you should see your terminal/command prompt now have the initial part change to something like:

root@e7b19cfb4461:/

Where you’ll noticed the CONTAINER ID is now shown there.

You are now using command line inside of your container. Trying running any bash command like ls to see what files are here for example.

Your screen will end up looking like this (but this printout is from a Mac, windows will look a little different in some parts):

(base) firstnamelastname@FirstNames-MacBook-Pro ~ % docker exec -it e7b19cfb4461 bash

root@e7b19cfb4461:/# ls
bin   etc   lib    libexec  mnt   rocker_scripts  sbin  tmp
boot  home  lib32  libx32   opt   root            srv   usr
dev   init  lib64  media    proc  run             sys   var
root@e7b19cfb4461:/#

When you are ready to leave your container type exit and press Enter.

3.1.1.4.4 Step 5: Try calling a script

To run a script using the docker container we could just add reference to a script at the end:

docker exec -it <PUT_CONTAINER_ID_HERE> bash run_analysis.sh

BUT! You will find that this command won’t work yet though, why?

bash: run_analysis.sh: No such file or directory

Question: Does our container have all of the same files that our computer has?

This won’t work because the file run_analysis.sh is not a file that our container has. Docker containers do not have all the files that our computer does; they only have the files we add to it.

3.1.2 Podman

3.1.2.1 Set up Podman

Go here to install Podman, following the instructions for your particular operating system. Open up your command line. Start up Podman by running the following:

podman machine init
podman machine start

3.1.2.2 Get workshop files

Which means: If your computer has files, software, or anything else. Your container by default does NOT have it unless you do something to get those files and software on there.

3.1.2.2.1 Step 1: Podman pull

We need to get the image we want to use. We do this by “pulling” it. Pulling is a term used in GitHub terminology too. It just means we’re taking something from the internet and getting a copy locally (on our computer).

Run this command in your Terminal or Command Prompt window:

podman pull cansav09/practice-image:1

If this command is running properly you should see some output that looks like this:

3.1.2.2.2 Step 2: Check what images you have

To see what images we have we can run podman image ls. You should see cansav09/practice-image:1 show up in the output and it will look a bit like this below:

docker.io/cansav09/practice-image  1   91619ca583b8  1 month ago  2.39 GB

3.1.2.2.3 Step 3: Run the image to start a container!

Remember images can be shared easily but to actually use the stuff we need to turn our image into a container! To do that we will do podman run

To launch the image for use, we can use podman run to turn it into a container.

podman run cansav09/practice-image:1

If this command has run properly it will return something like this:

Don’t worry too much about all this output, every image is configured differently and this output will change depending on that.

To see what containers we have running we have we can run:

podman ps

Your output will look a little like this. NOTE though that the CONTAINER ID and NAMES are randomly generated for each container and will change, so that part won’t be the same!

CONTAINER ID  IMAGE                                COMMAND     CREATED        STATUS        PORTS       NAMES
e7b19cfb4461  docker.io/cansav09/practice-image:1  /init       2 minutes ago  Up 2 minutes  8787/tcp    hopeful_proskuriakova

Pat yourself on the back! You have a running container!

What can we do with running containers?

3.1.2.2.4 Step 4: Run the container interactively

To run stuff interactively from the command line we can do:

But you’ll need to replace the <PUT_CONTAINER_ID_HERE> portion of this command with the CONTAINER ID reported when you ran podman ps.

podman exec -it <PUT_CONTAINER_ID_HERE> bash

If you run this you should see your terminal/command prompt now have the initial part change to something like:

root@e7b19cfb4461:/

Where you’ll noticed the CONTAINER ID is now shown there.

You are now using command line inside of your container. Trying running any bash command like ls to see what files are here for example.

Your screen will end up looking like this:

(base) firstnamelastname@FirstNames-MacBook-Pro ~ % podman exec -it e7b19cfb4461 bash

root@e7b19cfb4461:/# ls
bin   etc   lib    libexec  mnt   rocker_scripts  sbin  tmp
boot  home  lib32  libx32   opt   root            srv   usr
dev   init  lib64  media    proc  run             sys   var
root@e7b19cfb4461:/#

When you are ready to leave your container type exit and press Enter.

3.1.2.2.5 Step 5: Try calling a script

To run a script using the container we could just add reference to a script at the end:

podman exec -it <PUT_CONTAINER_ID_HERE> bash run_analysis.sh

BUT! You will find that this command won’t work yet though, why?

This won’t work because the file run_analysis.sh is not a file that our container has. Docker containers do not have all the files that our computer does; they only have the files we add to it.

bash: run_analysis.sh: No such file or directory

Question: Does our container have all of the same files that our computer has?

This won’t work because the file run_analysis.sh is not a file that our container has. Docker containers do not have all the files that our computer does; they only have the files we add to it.