Skip to main content

🐳 JobeInABox

Docker Stars Docker Pulls

The Moodle CodeRunner question type plugin requires a Jobe server on which to run student-submitted jobs. JobeInABox is a Docker image that provides a basic Jobe server that runs all the standard languages. For full information on Jobe servers, see the full Jobe documentation

Security

For security and performance reasons it is strongly recommended to run Jobe on a dedicated standalone server, even when running it in a container.

CodeRunner​

A question type that allows question authors to set programming questions in which the student answer is code in some programming language, which is graded by running it. More generally it can handle any question to which the answer is text that can be graded by a computer program written by the question author.

For production use you should build your own image using the local timezone. In this example we use Docker as follows:

Pull this repo from Github, cd into the jobeinabox directory and type a command of the form

docker build -t ruseler/jobeinabox .
  • jobebian
docker build -t ruseler/jobeinabox:latest .
docker build -t ruseler/jobeinabox:debian-slim .
  • jobuntu
docker build -t ruseler/jobeinabox:ubuntu .

You can then run your newly-built image with the command

sudo docker run -d -p 4000:80 --name jobe ruseler/jobeinabox:latest

This will give you a jobe server running on port 4000, which can then be tested locally and used by Moodle as explained in the section "Using jobeinabox" below.

Using the pre-built jobeinabox image on docker hub​

To run the pre-built Docker Hub image, just enter the command:

sudo docker run -d -p 4000:80 --name jobe ruseler/jobeinabox:latest

This will give you a jobe server running on port 4000, which can then be tested locally and used by Moodle as explained in the section "Using jobeinabox" below.

Setting the number of jobe users​

By default, Jobe will run up to 8 jobs simultaneously. This is usually a suitable value for 8-core systems but if you have more cores available you will probably want to raise this number. To do so, exec a shell in the container with a command of the form

docker exec -it jobeinabox bash

and then:

nano  /var/www/html/jobe/app/Config/Jobe.php

Find the line

public int $jobe_max_users = 8;

and change the value from 8 to the number of cores on your machine, or to a higher value if you expect to run mainly I/O-bound jobs.

Then re-install Jobe (within the container) with the commands:

cd /var/www/html/jobe
./install --purge

Checking performance​

You can check the performance of the container with the command

docker exec -it jobeinabox python3 /var/www/html/jobe/testsubmit.py --perf

Using API Keys​

You can provide API keys during the image build via the --secret option.

The keys can be stored in a separate file, following the format:

'c1425880-2289-4b76-ae29-bcea34997256' => 0,
'de7970e6-0466-4ce2-a6a4-c1a51363bd03' => 90

You can find more information on API keys in the Jobe documentation.

With the following command the keys stored in a file called api_keys in the jobeinabox directory are automatically added to /var/www/html/jobe/app/Config/Jobe.php and the option $require_api_keys is set to true.

If no API keys are provided the configuration remains untouched.

Warnings:​

  1. The image is over 1 GB, so may take a long time to start the first time, depending on your download bandwidth.

Using jobeinabox​

Having started a jobeinabox container by either of the above methods, you can check it's running OK by browsing to

and you should get a JSON-encoded list of the supported languages, namely

curl -k -X GET http://host.docker.internal:4000/jobe/index.php/restapi/languages

If you wish to run the test suite within the container, use the command

docker exec -t jobeinabox python3 /var/www/html/jobe/testsubmit.py
Supported languages:
    c: 14.2.0
    cpp: 14.2.0
    java: 21.0.11
    nodejs: 20.19.2
    octave: 9.4.0
    pascal: 3.2.2
    php: 8.4.21
    python3: 3.13.5
docker exec -it jobeinabox /container-test.sh
docker exec -t jobeinabox python3 /var/www/html/jobe/minimaltest.py

To set your Moodle/CodeRunner plugin to use this dockerised Jobe server, set the Jobe server field in the CodeRunner admin settings (Site Administration > Plugins > Question types > CodeRunner) to

[host_running_docker]:4000

Do not put http:// at the start.

To stop the running server, enter the command:

sudo docker stop jobeinabox

To remove the running server, enter the command:

sudo docker rm jobeinabox

To check if there is anything left, enter the command

sudo docker ps -a

Notes on security:​

  1. Note that while the container in which this Jobe runs should be secure, the container's network is currently just bridged across to the host's network. This means that Jobe can be accessed from anywhere that can access the host and can access any URI that the host can access. Firewalling of the host is essential for production use.

  2. Rebuild the container regularly to ensures that it is running with the latest jobe version and security updates.

Localhost use​

Under /admin/settings.php?section=qtypesettingcoderunner

Jobe server qtype_coderunner | jobe_host

Add host.docker.internal:4000

Under /admin/settings.php?section=httpsecurity

  • cURL blocked hosts list | curlsecurityblockedhosts Remove all internal IPs

  • Add port 4000 to cURL allowed ports list cURL allowed ports list | curlsecurityallowedport

Last updated on by Adriano Ruseler