Running Lessons With Docker
Correlation-One uses docker to run Jupyter Notebook training cases. This
allows us to eliminate portability issues between different operating systems.
This document describes how to use docker to run training cases.
Supported systems include macOS, Windows (7/8/10), and Linux.
Installing Docker
The installation process differs by system. We defer to the official docker
documentation for the installation instructions.
• Linux: docker-engine
• Windows 7/8 and Windows 10 Home (below version 2004): docker-toolbox
• macOS (before El Capitan 10.11): docker-toolbox
• Windows 10 Home (version 2004 or higher): docker-desktop
• Windows 10 Pro/Enterprise/Education (Build 16299 or later):
docker-desktop
• macOS (from El Capitan 10.11 on): docker-desktop
Note: If you are using Windows, you will need to enable virtualization on your
machine. This process is machine specific, we defer to this guide for further
guidance, but still you may need to search how to do it for your specific machine.
Running the Docker Image
Once docker is installed, please follow the below instructions on how to run the
relevant docker image for your operating system.
Windows 10 Home (version 2004 or higher)/Pro/Enterprise/Education
(Build 16299 or later)
Open the Command Prompt as an administrator. In order to do so, please find
CMD in the start menu, right-click it, and select “Run as Administrator”.
Enter the following command to pull and run the docker image:
docker run -p 8888:8888 -v c:\Users\<User>\path\to\your\notebooks:^
/home/jovyan/work jupyter/scipy-notebook:17aba6048f44
Case-writers: The path c:\Users\<User>\path\to\your\notebooks should be
replaced to the directory containing the starter-kit/ directory otherwise
tests will not be executed.
Replace the <User> field with your Windows system user-name. The path
c:\Users\<User>\path\to\your\notebooks should contain the case materials
for your lesson. This path is the only part of your local file system visible to the
1
docker container! Feel free to change the path to something more convenient if
you feel so inclined.
The prompt should pull the docker image and run a notebook server. Follow
the instructions in the accessing the notebook server section.
macOS (from El Capitan 10.11 on) and Linux
On macOS, launch the Terminal application. On Linux, open your preferred
terminal.
Enter the following command to pull and run the docker image:
docker run -p 8888:8888 -v "$PWD":/home/jovyan/work \
jupyter/scipy-notebook:17aba6048f44
Case-writers: Please run the command above from the directory containing
the starter-kit/ directory otherwise tests will not be executed.
Students: Ensure that the command is run from a directory containing the cases
used in your lesson.
The prompt should pull the docker image and run a notebook server. Follow
the instructions in the accessing the notebook server section.
Accessing the notebook server
After running the docker run command with the instructions according to your
operating system, you will be able to access the notebook server. You should
have been prompted with the following:
Copy/paste this URL into your browser when you connect for the first time,
to login with a token:
http://(b56ec2d7c00a or 127.0.0.1):8888/?token=<some-token>
Copy this URL and paste it into your browser (we’ve found MS Edge not to
work properlly), editing the (b56ec2d7c00a or 127.0.0.1) portion to be just
the 127.0.0.1 component.
Please note that the token field will be different each time you run the docker
run command. So you cannot reuse the same URL without first changing the
<some-token> value.
Windows 7/8/10 Home (below version 2004) and macOS (before El
Capitan 10.11)
Open the Docker Quickstart Terminal application.
Enter the following command to know the ip address on which your default
docker-machine is running:
docker-machine ip default
2
The prompt should display the address of your default docker-machine. You
will use this address to replace the <docker-machine ip address> in a later
step.
Then enter the following command to pull and run the docker image:
docker run -p 8888:8888 -v "$PWD":/home/jovyan/work \
jupyter/scipy-notebook:17aba6048f44
Case-writers: Please run the command above from the directory containing
the starter-kit/ directory otherwise tests will not be executed.
Students: Ensure that the command is run from a directory containing the cases
used in your lesson.
The prompt should pull the docker image and run a notebook server. You
should have been prompted with the following:
Copy/paste this URL into your browser when you connect for the first time,
to login with a token:
http://(b56ec2d7c00a or 127.0.0.1):8888/?token=<some-token>
This URL will not work on a Windows system using docker-toolbox . To make
it work, you need to replace the prompted IP address with the IP address of
your docker machine. Your URL should look something like this:
http://<docker-manchine ip address>:8888/?token=<some-token>
Where <docker-manchine ip address> is the ip address you got when running
docker-manchine ip default and <some-token> is the token outputted by
the docker run command. Copy the constructed URL into your favorite browser
(we’ve found MS Edge not to work properlly) to open the notebook.
Please note that the token field will be different each time you run the docker
run command. So you cannot reuse the same URL without first changing the
<some-token> value.
Environment Management
Case writers: Please ignore this section.
Once a docker process is running, you may follow the instructions listed in the
Managing Case Environments document.
Managing Docker Processes
Running docker images persist in the background once opened. Eventually, you
should remove these running processes.
List the running docker processes by using docker ps. Sample output:
3
docker ps
CONTAINER ... NAMES
aacda4d33aef elastic_zhukovsky
The output above shows a table of running processes. Most of the fields are
omitted for clarity. The NAMES field is used to reference the process. Names are
randomly generated.
To stop a process, use the docker stop command with the name of the process.
docker stop <name-of-process>
# e.g., docker stop elastic_zhukovsky
