brauner/rblas repository overview

⁠dockR

Docker images for R based on the official Ubuntu minimal build. The images have R set as their default entrypoint. Hence, they behave like R binaries.

⁠Some properties

• All images are available as automated builds from Docker Hub. You can just pull them with docker pull brauner/r-patched and docker pull brauner/r-devel.
• The generic R images which reside in the r-patched and r-devel folders are compiled without setting the march flag. This will make them run on any system. To see how to adapt the image to a specific architecture by setting the march flag take a look at the Dockerfiles which reside in the folders which have ivy appended to them. There you can also see how to enable 3D support and various other tweaks.
• Set up user so that the container does not need to run as root.
• Set up /home for user and a default repository in .Rprofile.
• Install all recommended R dependencies, pull R-patched or R-devel from SVN and compile R-patched or R-devel from source.
• The R binaries reside in /usr/local/bin/R.

⁠Workflow & R Library and Package Management

If you mainly run R as an ephemeral interactive container and install new packages you will need to commit the newly created layer to your image before you quit in order to have the packages you install available via library() when you start the container again. To circumvent this I suggest sharing volumes between your R containers in the following manner:

• Create an extremely tiny container from the busybox image exposing two folders R and R-dev (for devtools afficionados) with the right permissions which you share via the --volumes-from=DATACONTAINERNAME flag among you R containers. You will find the Dockerfile for this in the folder rlib and the image on Docker Hub. You can pull it with docker pull brauner/rlib.
• Run docker run --name=RDATA DATACONTAINERNAME true
• Run docker run --volumes-from = RDATA RCONTAINERNAME

Now you can install packages into RDATA without having to commit the container as data containers are handled differently by Docker. If you pull a new R image you can just keep running it with your libraries still intact. Should a new R version come out that requires upgrading all libraries you can just remove your RDATA container which will also remove all installed packages and start a new one.

• If you still want to install a package ephemerally which gets deleted when your R container exits. You can use devtools dev_mode() function to specify a new library path: library(devtools); dev_mode(, path = "~/your/path/to/new/library/here") or edit .libPaths() directly.

⁠Graphical Output from Docker Containers

There is a nice and semi-easy way of getting graphical output from a Docker container without having to run an sshd daemon inside of the container. Docker can provide bare metal performance when running a single process which in this case is supposed to be R. Running an sshd daemon will, marginal as it may be, introduce additional overhead. This is not made better by running the sshd daemon as a child process of the supervisor daemon. Both can be dispensed with when one makes good use of bind mounts. After building the image from which the container is supposed to be run we start an interactive container and bind mount the /tmp/.X11-unix folder into it. I will state the complete command and explain in detail what it does:

docker run -i -t --rm \
# -i sets up an interactive session; -t allocates a pseudo tty; --rm makes
# this container ephemeral
-e DISPLAY=$DISPLAY \
# sets the host display to the local machines display (which will usually
# be :0)
-u chbr \
# -u specify the process should be run by a user (here "chbr") and not by
# root. This step is important (v.i.)!
-v /tmp/.X11-unix:/tmp/.X11-unix \
# - v bind mounts the `X11` socket `/tmp/.X11-unix` into `/tmp/.X11-unix`
# in the container.
--name="rdev" ubuntu-r1 R
# --name="" specify the name of the container (here "rdev"); the image you
# want to run the container from (here "ubuntu-r"); the process you want
# to run in the container (here "R"). Note that the process for this image
(here "R") can be left unspecified as the program is the default
entrypoint of the image.

After issuing this command you should be looking at the beautiful R start output. If you were to try demo(graphics) to see if graphical output is already working you would note that it is not. That is because of the Xsecurity extension preventing you from accessing the socket. You could now type xhost + on your local machine and try demo(graphics) in your container again. You should now have graphical output. This method however, is strongly discouraged as you allow access to your xsocket to any remote host you're currently connected to. As long as you're only interacting with single-user systems this might be somehow justifiable but as soon as there are multiple users involved this will be absolutely unsafe! Hence, you should use a less dangerous method. A good way is to use the server interpreted xhost +si:localuser:username which can be used to specify a single local user (see man xhost). This means username should be the name of the user which runs the X11 server on your local machine and which runs the Docker container. This is also the reason why it is important that you specify a user when running your container. Last but not least there is always the more complex solution of using xauth and .Xauthority files to grant access to the X11 socket (see man xauth). This however will also involve a little more knowledge how X works.

⁠Entering a running container with docker exec

As of release 1.3. the recommended way of entering a running container is by using docker exec -it rdev bash (rdev is the name of the running container and bash the program which is supposed to be run in the container in this example.) which will spawn a new process in the running container.

⁠Entering a running container with nsenter

Should you need to enter a running container with a new tty you can also use nsenter. First find the PID of the (main) process running in the container by either issuing docker top containername or docker inspect --format {{.State.Pid}} containername. Then use nsenter which should usually be installed on your system. If not install it. It can be found util-linux (version must be at least 2.23). You can use the command nsenter --target PID-you-just-found-out --mount --ipc --net --pid or the short version nsenter -t PID-you-just-found-out -m -i -n -p.