Docker Image for Mapper

This Docker image provides the Ubuntu 16.04 environment with X Windows for multiphysics coupling.

This image runs the lightweight LXDE Windows Manager, and has FEniCS 17.1.0, Python 3.5.2 (with NumPy, SciPy, Pandas and Spyder), Jupyter Notebook, and g++-5.4 preinstalled. The X Windows will display in your web browser in full-screen mode. You can use this Docker image on 64-bit Linux, Mac or Windows.

Preparation

Before you start, you need to first install Python and Docker on your computer by following the steps below.

Installing Python

If you use Linux or Mac, Python is most likely already installed on your computer, so you can skip this step.

If you use Windows, you need to install Python if you have not yet done so. The easiest way is to install Miniconda, which you can download at https://repo.continuum.io/miniconda/Miniconda3-latest-Windows-x86_64.exe. During installation, make sure you check the option to make Miniconda the system's default Python.

Installing Docker

Download the Docker Community Edition for free at https://www.docker.com/community-edition#/download and then run the installer. Note that you need administrator's privilege to install Docker. After installation, make sure you launch Docker before proceeding to the next step.

Notes for Windows Users

1. Docker only supports 64-bit Windows 10 Pro or higher. If you have Windows 8 or Windows 10 Home, you need to upgrade your Windows operating system before installing Docker. Stony Brook students can get Windows 10 Education free of charge at https://stonybrook.onthehub.com. Note that the older Docker Toolbox supports older versions of Windows, but it should not be used.
2. After installing Docker, you may need to restart your computer to enable virtualization and Microsoft Hyper-V. Note that you may also need to change the BIOS of your computer to enable hardware virtualization if it was not yet enabled.
3. For security reasons, do not use Docker in the Administrator account, even if you are the sole user on the computer. For Docker version 17.06 or later, your Windows account must be a member of the local group “docker-users” for you to run Docker. To add your account to the group, you need to log into an Administrator account and run Computer Management. Open up Local Users and Groups, select Groups, right click on docker-users in the list, and then click on Add to Group... to add your username to the group.
4. If you previously installed VMWare or VirtualBox on your Windows computer, they might conflict with Docker. You may get an error message stating that virtualization must be enabled when starting Docker, even though virtualization was already enabled. To resolve the issue, go to Windows Features in the Control Panel, disable Hyper-V and then re-enable it.
5. When you use Docker for the first time, you must change its settings to make the C drive shared. To do this, right-click the Docker icon in the system tray, and then click on Settings.... Go to Shared Drives tab and check the C drive.
6. Docker for Windows saves the images and data volumes in a shared public folder C:\Users\Public\Documents\Hyper-V\Virtual Hard Disks\MobyLinuxVM.vhdx. This is a major security risk because all your images and data can be accessed and modified by other Docker users on the same computer. If you are using a shared Windows computer, make sure you create a private folder such as C:\Users\YourUserName\Documents\Hyper-V\Virtual Hard Disks and then go to Advanced tab in Docker Settings, and change the Image and Volume VHD Location to this folder.

Notes for Mac Users

1. Docker Version 17.06.0-ce has a bug that causes Docker to restart when your computer wakes up from sleep and when your network settings change. As a workaround, change the Proxies setting in Docker's preferences to "No Proxy".
2. On Mac, the clock in Docker would lag when your computer goes to sleep and then wakes up. You can resolve this issue either by restarting Docker after waking up or by installing sync-docker-time.

Notes for Linux Users

• Most Linux distributions have a docker package. You can use the package installer for your system to install docker. Note that on some system (e.g., OpenSUSE), you may need to run the following commands to start up docker after installing docker:

sudo systemctl enable docker
sudo systemctl start docker

• After you install Docker, make sure you add yourself to the docker group. On Ubuntu, this is done by running the command:

sudo adduser $USER docker

On other systems, try the following command.

sudo usermod -G docker -a $USER 

After adding yourself to the docker group, you need to log out and log back in before you can use Docker.

Running the Docker Image as Jupyter-Notebook Server

To run the Docker image, first download the script mapper_jupyter.py and save it to the working directory where you will store your notebooks, codes and data. You can download the script using command line: On Windows, start Windows PowerShell, use the cd command to change to the working directory where you will store your codes and data, and then run the following command:

curl https://raw.githubusercontent.com/unifem/mapper-desktop/master/mapper_jupyer.py -outfile mapper_jupyer.py

On Linux or Mac, start a terminal, use the cd command to change to the working directory, and then run the following command:

curl -s -O https://raw.githubusercontent.com/unifem/mapper-desktop/master/mapper_jupyer.py

After downloading the script, you can start the Docker image using the command

python mapper_jupyer.py -p

This will download and run the Docker image and then launch your default web browser to show the Jupyter Notebook page. The -p option is optional, and it instructs the Python script to pull and update the image to the latest version.

For additional command-line options, use the command

python mapper_jupyter.py -h

Running the Docker Image with Desktop Environment

Besides using the Docker Image as a Jupyter Notebook server, you can also start it in an X-Windows desktop environment. Simply replace mapper_jupyter.py with mapper_desktop.py in the preceding commands. That is, on Windows use the following commands in the PowerShell:

curl https://raw.githubusercontent.com/unifem/mapper-desktop/master/mapper_desktop.py -outfile mapper_desktop.py
python mapper_desktop.py -p

and on Linux and Mac use the following commands in a terminal:

curl -s -O https://raw.githubusercontent.com/unifem/mapper-desktop/master/mapper_desktop.py
python mapper_desktop.py -p

It will automatically launch your default web browser to show the desktop. For a better experience, use VNC Viewer to connect to the Docker image.

Running the Docker Image Offline

After you have download the Docker image using the curl and python commands above, you can run the image offline without internet connection using the following command:

python mapper_jupyter.py

or

python mapper_desktop.py

in the directory where you ran the curl command above.

Stopping the Docker Image

To stop the Docker image, press Ctrl-C twice in the terminal (or Windows PowerShell on Windows) on your host computer where you started the Docker image, and close the tab for the desktop in your web browser.

Entering Full-Screen Mode for Desktop Environment

For the best experience, use VNC Viewer to connect to Docker image with the port and password displayed in the terminal output, which supports the full-screen mode. If you don't have the VNC viewer, you can also use the full-screen mode in a web browser.

When using a web browser, we recommend Google Chrome or Chromium browser, which have the same user interface. On Windows or Linux, you can enter full-screen mode by selecting the menu View --> "Full Screen" Alternatively, open the Chrome menu (the three vertical dots at the top right) and select the square to the far right of the Zoom buttons (the "+" and "-" buttons). To exit the full-screen mode, press the F11 key. On Mac, it behaves similarly except that the menu item is named Enter Full Screen instead of Full Screen, and the keyboard shortcut is Ctrl-Cmd-f instead of F11. You can also click on the green circle at the top-left corner of Google Chrome to enter and exit the full-screen mode. Note that in the full-screen mode, you need to disable Always Show Toolbar in Full Screen under the View menu of Google Chrome, and you can reveal the menu and the toolbar by sliding your mouse to the top of the display.

Alternatively, you can also use the "native" browsers on different platforms.

• On Windows, you can use the native browser Microsoft Edge. Toggle on and off the full-screen mode by pressing Win+Shift+Enter (hold down the Windows and Shift keys and press Enter).
• On Mac, you can use the native browser Safari, for which you can toggle the full-screen mode by clicking on the green circle at the top-left corner of Safari or selecting the View --> "Enter Full Screen" menu. To exit the full-screen mode, press Ctrl-Cmd-f, or slide your mouse to the top of the display to enable the menus.
• On Linux, the default browser Firefox does not hide its address bar in its native full-screen mode. You are recommended to use Google Chrome or Chromium browser instead. However, you can use Firefox for a full-screen viewing mode by clicking on the Fullscreen button in the left sidebar of Docker desktop environment. However, this is not recommended for day-to-day use, because Firefox would exit this full-screen mode whenever you press Esc, which may happen quite often.

If your Docker desktop environment started automatically in a non-recommended browser, you can copy and paste the URL into a recommended browser.

Tips and Tricks

1. By default, Docker uses two CPU cores and 2GB of memory on Mac and Windows. If you want to run large jobs, you can increase the amount of memory or the number of cores dedicated to Docker. Just click on the Docker icon in the system tray, select Settings (or Preferences for Mac) and then select the Advanced tab to adjust the settings.
2. When using the Docker desktop, the files under $HOME/.config, $HOME/.ssh, , $HOME/unifem, $HOME/shared and any other directory that you might have mounted explicitly are persistent. Any change to files in other directories will be lost when the Docker container stops. Use $HOME/.config to store the configuration files of the desktop environment. $HOME/shared maps to the working directory on the host, and you are recommended to use it or a mounted project directory to store codes and data.
3. The $HOME/.ssh directory in the Docker container maps to the .ssh directory on your host computer. This is particularly convenient for you to use your ssh-keys for authentications with git repositories (such as github or bitbucket). To use your ssh keys, run the ssh-add in a terminal to add your keys to the ssh-agent.
4. You can copy and paste between the host and the Docker desktop through the Clipboard box in the left toolbar, which is synced automatically with the clipboard of the Docker desktop. To copy from the Docker desktop to the host, first, select the text in the Docker desktop, and then go to the Clipboard box to copy. To copy from host to the Docker desktop, first, paste the text into the Clipboard box, and then paste the text in the Docker desktop.
5. To stop the Docker container, do not just close the browser window, because the Docker container would still be running in the background. Instead, you can stop the container using one of the following approaches:

• Use the logout button in the lower-left corner of the Docker desktop,
• Press Ctrl-C twice in the terminal where you started the python script, or
• Run the command docker stop <Container ID> in a terminal on the host, and you can find the Container ID using the docker ps -a command.