SLP: Docker Usage

Go up to the main SLP documents page (md)

Docker install

• You will have to enable virtualization, which is a BIOS option. If you have already run VirtualBox on this machine, then you've done that. Otherwise, there will be an option for "virtualization technology", "VT-X", "AMD-v", or similar.
• Windows can download and install from https://docs.docker.com/docker-for-windows/install/
  • That version requires Windows 10; previous versions of Windows can use the install at https://www.docker.com/products/docker-toolbox
  • Once installed, you run the Docker Quickstart Terminal
    • If you run into a problem such as the one described here, even if you are on a different version of Widnows, then it could be a driver issue: there is a bug with the NDIS6 driver, and you should select the NDIS5 driver during isntall (see here for details). Thus, you'll have to uninstall and reinstall docker, selecting the appropriate driver.
  • To connect to the mapped port, run docker-machine ip default in another docker quickstart terminal (while the other container is running). You will get an IP address such as 192.168.99.100. You can then view the mapped port at http://192.168.99.100:3000.
• Mac OS X can download and install from https://docs.docker.com/docker-for-mac/install/

Docker startup

• Download the docker image
  • Run docker pull aaronbloomfield/slp (it's a public image/repo)
    • You can run that last command in the future to upgrade to the latest version
  • The image name is now "aaronbloomfield/slp", which is long to type each time; you can rename it via docker tag aaronbloomfield/slp slp
    • That being said, the above commands only need to be run once (or if you are upgrading), and not each time you start up a container
  • Instead of the two docker commands above, you could build the image yourself from the included Dockerfile, but that will take much longer
• Choose a directory to save your work do
  • If you save it in the container, if the container is removed (as often happens), then your work is lost, hence why we want to save it locally
  • We'll call that directory /path/to/work
  • Note that the docker-run command, below, will likely save files into that directory as root, so you may have to use sudo to modify them
• Start up docker: docker run -it -p 3000:3000 -v /path/to/work:/root/work slp
  • This will provide you with a command prompt inside the container

To create and run a rails app:

• You likely will want to be in the /root/work directory, since that is the one that the docker-run command above mapped to your host file system (cd /root/work)
• Run rails new myapp (we aren't specifying a DB, so it defaults to sqlite3); change "myapp" to the name of your project
• Change the last line in your Gemfile from gem 'tzinfo-data', platforms: [:mingw, :mswin, :x64_mingw, :jruby] to just gem 'tzinfo-data'
• Remove the Gemfile.lock file in the project directory
• Run bundle install in the project directory
• You can now run rails server from the project directory
• You can view your app from http://localhost:3000
• NOTE: if and when you re-launch the container, you will likely have to run bundle install from your project directory again

To create and run a django app:

• You likely will want to be in the /root/work directory, since that is the one that the docker-run command above mapped to your host file system (cd /root/work)
• Run django-admin startproject mysite
  • Change "mysite" to the name of your project
• In the mysite/ directory, run python3 manage.py runserver 0.0.0.0:3000
  • Note that you have to specify the IP address and port!
• You can view your app from http://localhost:3000