Using Docker with the CMS open data

Overview

Teaching: Self-guided min
Exercises: 40 min

Questions

• How do I use docker to effectively interface with the CMS open data?

• What container images are available for my work with the CMS open data?

Objectives

• Download the CMSSW open data docker image

• Open your own CMSSW open data container and check that graphical windows open

• Download the ROOT and python images and build your own container

• Restart an existing container

• Delete and rebuild containers

Overview

This exercise will walk you through setting up and familiarizing yourself with Docker, so that you can effectively use it to interface with the CMS open data. It is not meant to completely cover containers and everything you can do with Docker.

Three types of container images are provided: one with the CMS software (CMSSW) compatible with the released data, and two others with ROOT and python libraries needed in this workshop. The CMSSW container is mandatory if you want to access the CMS data in AOD and MiniAOD formats (you will learn about them later), as you will not be able to install CMSSW software on your own computer. The two others are provided to make setting up and using ROOT and/or python libraries easier for you for this tutorial, but if you wish, you can also install them on your computer.

All container images come with VNC for the graphical use interface. It opens directly in a browser window. Optionally, you can also connect to the VNC server of the container using a VNC viewer (VNC viewer (TigerVNC, RealVNC, TightVNC, OSX built-in VNC viewer, etc.) installed on your local machine, but only the browser option for which no additional tools are needed is described in these instructions. On native Linux, you can also use X11-forwarding.

For different CMSSW container images, some guidance can be found on the Open Data Portal introduction to Docker. In this tutorial, we will use the container image needed for the CMS open data from 2015. The use of graphical interfaces, such the graphics window from ROOT, depends on the operating system of your computer. Therefore, in the following, separate instructions are given for Windows WSL, Linux and MacOS.

Note that the container images are large (the compressed download size is 6.6GB for the CMSSW container, and of order of 1GB for the ROOT and python containers). Make sure that you make it in time to download them and work through the exercises before the workshop.

Download the docker image for CMSSW open data and start a container

The first time you start a container, a docker image file gets downloaded from an image registry. The CMSSW open data image is large (6.6GB) and it may take very long to download, depending on the speed of your internet connection. After the download, a container created from that image starts. The image download needs to be done only once. Afterwards, when starting a container, it will find the downloaded image on your computer, and it will be much faster.

The containers do not have modern editors and it is expected that you mount your working directory from the local computer into the container, and use your normal editor for editing the files. Note that all your compiling and executing still has to be done in the Docker container!

First, before you start up your container, create a local directory where you will be doing your code development. In the example below, it is called cms_open_data_work and it will live in the $HOME directory. You may choose a different location and a shorter directory name if you like. :)

Local machine

cd # This is to make sure I'm in my home directory
mkdir cms_open_data_work

Warning!

If you do not create the directory on your local computer before creating the container, the directory is created automatically but with the wrong user/group. When starting the container, you will get a message cannot make directory CMSSW_7_6_7 Permission denied. In that case, delete the directory with rm -rf cms_open_data_work/, and remove the failing container with docker rm <container-name> so that you can use the same name. In the following, we will use my_od as the container name. And then, remember to create the directory before creating the container!

Start the container following the instructions below depending on the operating system you are using.

docker run -it --name my_od --net=host --env="DISPLAY" -v $HOME/.Xauthority:/home/cmsusr/.Xauthority:rw  -v ${HOME}/cms_open_data_work:/code cmsopendata/cmssw_7_6_7-slc6_amd64_gcc493 /bin/bash

Setting up CMSSW_7_6_7
CMSSW should now be available.
This is a standalone image for CMSSW_7_6_7 slc6_amd64_gcc493.
(/code/CMSSW_7_6_7/src)

xhost local:root

docker run -it --name my_od -P -p 5901:5901 -p 6080:6080 -v ${HOME}/cms_open_data_work:/code cmsopendata/cmssw_7_6_7-slc6_amd64_gcc493 /bin/bash

Setting up CMSSW_7_6_7
CMSSW should now be available.
This is a standalone image for CMSSW_7_6_7 slc6_amd64_gcc493.
(/code/CMSSW_7_6_7/src)

 start_vnc

xauth:  file /home/cmsusr/.Xauthority does not exist

New 'myvnc:1' desktop is 1df549a6f098:1

Starting applications specified in /home/cmsusr/.vnc/xstartup
Log file is /home/cmsusr/.vnc/1df549a6f098:1.log

[1] 144
VNC connection points:
        VNC viewer address: 127.0.0.1:5901
        HTTP access: http://127.0.0.1:6080/vnc.html
To kill the vncserver enter 'vncserver -kill :1'

 stop_vnc
 exit

docker run -it --name my_od -P -p 5901:5901 -p 6080:6080 -v ${HOME}/cms_open_data_work:/code cmsopendata/cmssw_7_6_7-slc6_amd64_gcc493 /bin/bash

Setting up CMSSW_7_6_7
CMSSW should now be available.
This is a standalone image for CMSSW_7_6_7 slc6_amd64_gcc493.
(/code/CMSSW_7_6_7/src)

start_vnc

xauth:  file /home/cmsusr/.Xauthority does not exist

New 'myvnc:1' desktop is 1df549a6f098:1

Starting applications specified in /home/cmsusr/.vnc/xstartup
Log file is /home/cmsusr/.vnc/1df549a6f098:1.log

[1] 144
VNC connection points:
        VNC viewer address: 127.0.0.1:5901
        HTTP access: http://127.0.0.1:6080/vnc.html
To kill the vncserver enter 'vncserver -kill :1'

 stop_vnc
 exit

Download the docker images for ROOT and python tools and start container

Containers with ROOT and python libraries installed are provided for your convenience. These containers can be used in the C++, ROOT and python tools lesson and later on for your work with CMS open data.

ROOT container

ROOT is included in the CMSSW container, but it is an old version because it needs to be compatible with the environment needed to access CMS open data AOD and MiniAOD files. In this tutorial, and in your work with CMS open data, you will often work on data that have been derived from the AOD or MiniAOD files and are not tied to a specific ROOT version. Therefore, a container with more recent ROOT version is provided.

First, create a working directory on your local computer:

cd
mkdir cms_open_data_root

Then, download the ROOT container and start it with the docker run command.

If you are on native Linux and want to use X11-forwarding, use

docker run -it --name my_root --net=host --env="DISPLAY" -v $HOME/.Xauthority:/home/cmsusr/.Xauthority:rw -v ${HOME}/cms_open_data_root:/code gitlab-registry.cern.ch/cms-cloud/root-vnc:latest

On MacOS and Windows WSL2 (and on native Linux if you do not want to use X11-forwarding), use

docker run -it --name my_root -P -p 5901:5901 -p 6080:6080 -v ${HOME}/cms_open_data_root:/code gitlab-registry.cern.ch/cms-cloud/root-vnc:latest

This opens a bash shell where you can type your commands. Edit files in the cms_open_data_root directory on your local computer, but run the commands in the container.

For graphics, on native Linux, use X11-forwarding. On other systems, use VNC that is installed in the container and start the graphics windows with start_vnc. Open the browser window in the address given at the start message (http://127.0.0.1:6080/vnc.html) with the default VNC password is cms.cern. It shows an empty screen to start with and all graphics will pop up there.

Type exit to leave the container, and if you have started VNC, stop it first:

stop_vnc
exit

Python tools container

ROOT is not the only option for analysis of CMS open data. A container image is provided with all python libraries that will be needed in this tutorial.

First, create a working directory on your local computer:

cd
mkdir cms_open_data_python

Then, download the python container and start it with the docker run command.

If you are on native Linux and want to use X11-forwarding, use

docker run -it --name my_python -P -p 8888:8888 --net=host --env="DISPLAY" -v $HOME/.Xauthority:/home/cmsusr/.Xauthority:rw -v ${HOME}/cms_open_data_python:/code gitlab-registry.cern.ch/cms-cloud/python-vnc:latest

On MacOS and Windows WSL2 (and on native Linux if you do not want to use X11-forwarding), use

docker run -it --name my_python -P -p 5901:5901 -p 6080:6080 -p 8888:8888 -v ${HOME}/cms_open_data_python:/code gitlab-registry.cern.ch/cms-cloud/python-vnc:latest

This opens a bash shell where you can type your commands. Edit files in the cms_open_data_python directory on your local computer, but run the commands in the container.

You can run jupyter notebooks in this container by typing in the container prompt

jupyter-lab --ip=0.0.0.0 --no-browser

and opening the link in the message on your browser.

If you see Permission denied when you try to open a new notebook, you most likely forgot to create the local working directory before creating the container. In that case, the directory was created automatically but with the wrong user/group. Exit from the container with exit. Then remove the container, remove the working directory, create it again:

docker rm my_python
rm -rf cms_open_data_python
mkdir cms_open_data_python

and create the container again with the docker run ... command above.

For other graphics, on native Linux, use X11-forwarding. On other systems, use VNC that is installed in the container and start the graphics windows with start_vnc. Open the browser window in the address given at the start message (http://127.0.0.1:6080/vnc.html) with the default VNC password is cms.cern. It shows an empty screen to start with and all graphics will pop up there.

Type exit to leave the container, and if you have started VNC, stop it first:

stop_vnc
exit

Coming back to the same container

You can come back to the same container you’ve used earlier with the docker start ... command.

docker start -i my_od

Note that running the docker run ... command as before would create a new container from the image you’ve downloaded. This would be a new environment. However, as we are mounting the directory from the local computer into the container, you will see the files from your earlier container even in your new container. Most often, you do not want to create a new container but you would use the existing container to go to the same working area with all our files and code saved.

CHALLENGE! Test persistence

Go into the container and in the /code/CMSSW_7_6_7/src directory, create a test file using some simple shell commands. Type the following exactly as you see it. It will dump some text into a file and then print the contents of the file to the screen

echo "I am still here" > test.tmp
cat test.tmp

After you’ve done this, check if you see the file test.tmp in your local computer in the cms_open_data_work/CMSSW_7_6_7/src directory. If you did it correctly, you should be able to list the contents of the directory with ls -l and see your file from before! If not, check that you followed all the instructions above correctly or contact the facilitators.

Now, exit from the container and remove it with

docker rm my_od

Create a new container with the docker run command that you used in the first place. Check if you see the file that you created before.

Note that with the volume mount, your files will not disappear when you remove the container because they are stored in a directory on the local computer. If you really want to get rid of them, you will have to delete them either on the container or on your local computer.

You can make use of this, for example, when you have forgotten to stop VNC with stop_vnc when you exit the container. Probably the quickest way to clean is to remove and recreate the container. When you do it, the files that block the VNC from starting will be removed and as they are not located in the mounted directory, they will not be present when you create a new container. But the files in your working area (as test.tmp above) will be there again.

Stopping and removing containers

As you are learning how to use Docker, you may find yourself with multiple containers. Or maybe you started a container with your favourite name with some set of flags and now you want use that same name but with new flags. In that case, you will want to stop the container and remove it.

A container stops when you type the exit command in the container prompt. It may happen that you accidentally close the terminal where the container is running. In that case, the container will not stop and it will remain running. You can list the running containers with docker ps. You can either return to the container using its name (here “my_od”) with the start command on your local machine and then exit normally from the container prompt:

docker start -i my_od
exit

or stop the container with

docker stop my_od

To stop all running containers:

docker stop $(docker ps -q)

To remove the container “my_od”, you would type the following. Note that this will delete the container and all files, but the files in the CMSSW_7_6_7/src directory which is shared with your local computer will remain in your local computer’s directory.

docker rm my_od

To remove all containers:

docker rm $(docker ps -aq)

Don’t worry!

Note that these commands will not remove the actual Docker image that you downloaded and may have taken quite some time to download! Whew!

Key Points

• You have now set up a docker container as a working enviroment for CMS open data. You know how to open a graphical window in it and how to pass files between your own computer and the container.