Skip to content

010_Setup_Docker.md

Latest commit

 

History

History
119 lines (93 loc) · 5.25 KB

010_Setup_Docker.md

File metadata and controls

119 lines (93 loc) · 5.25 KB

Docker CE

You need Docker > 19.03.

If this version of Docker is not already installed from the standard Ubuntu repositories you need to uninstall docker first. Step 5 is mandantory also if you use the ubuntu version.

Then you can install the latest Docker version as described here: https://docs.docker.com/engine/install/ubuntu/

These are the most common steps:

  1. Install curl (if not already done)
$ sudo apt install curl
  1. Install the docker repository key:
$ curl -fsSL https://download.docker.com/linux/ubuntu/gpg | sudo apt-key add -
  1. Install the repository:
$ sudo add-apt-repository \
   "deb [arch=amd64] https://download.docker.com/linux/ubuntu \
   $(lsb_release -cs) stable"
  1. Update repositories and install docker_ce:
$ sudo apt update
$ sudo apt install docker-ce
  1. Add your user to the docker group:

In order to be able to run docker commands without being root, you should add your user to the docker group (new login required to take effect).

$ sudo adduser $(id -un) docker

NVIDIA Docker (2)

Then, you can proceed to install the NVIDIA Docker containers for 3d acceleration with access to (nvidia) graphics hardware.

Note: This requires to have the current NVIDIA driver installed, ubuntu is not updating major versions, check your package manager for the latest main package of the driver. When this document was written (30.11.2018) the latest version in the standard ubuntu package repositories was: nvidia-driver-390. You can look for more recent drivers (higher numbers) via apt search nvidia-driver-*.

The following commands can be used to install the nvidia docker images. Original repository and installation instructions can be found here: https://github.com/NVIDIA/nvidia-docker

# Add the package repositories
distribution=$(. /etc/os-release;echo $ID$VERSION_ID)
curl -s -L https://nvidia.github.io/nvidia-docker/gpgkey | sudo apt-key add -
curl -s -L https://nvidia.github.io/nvidia-docker/$distribution/nvidia-docker.list | sudo tee /etc/apt/sources.list.d/nvidia-docker.list

sudo apt update && sudo apt install nvidia-docker2
sudo systemctl restart docker

If successful, you should be able to reproduce this command and receive a similar output (also see here):

$ docker run --gpus all nvidia/cuda:10.0-base nvidia-smi
Mon Aug 24 19:28:37 2020       
+-----------------------------------------------------------------------------+
| NVIDIA-SMI 440.100      Driver Version: 440.100      CUDA Version: 10.2     |
|-------------------------------+----------------------+----------------------+
| GPU  Name        Persistence-M| Bus-Id        Disp.A | Volatile Uncorr. ECC |
| Fan  Temp  Perf  Pwr:Usage/Cap|         Memory-Usage | GPU-Util  Compute M. |
|===============================+======================+======================|
|   0  GeForce MX150       Off  | 00000000:01:00.0 Off |                  N/A |
| N/A   46C    P3    N/A /  N/A |    474MiB /  2002MiB |     27%      Default |
+-------------------------------+----------------------+----------------------+

+-----------------------------------------------------------------------------+
| Processes:                                                       GPU Memory |
|  GPU       PID   Type   Process name                             Usage      |
|=============================================================================|
+-----------------------------------------------------------------------------+

IMPORTANT: In order to enable 3d acceleration, existing containers have to be re-created.

  • Stop running containers: ./stop.bash
  • Clear container settings: rm -f has_gpu_support.txt *-container_id.txt
  • regenerate (and re-attach to) container: ./exec.bash

Restart and convenience

In order to be able to run docker commands without being root, you should add your user to the docker group (requires new login to take effect).

$ sudo adduser $(id -un) docker

In order to make sure that the kernel modules are properly loaded, you should restart your computer.

Now you are ready to use docker!

Using a registry

If you want to use the image generated by the docker build from this repository, you can also use the already build image from the registry server (docker hub).

If you want to pull images from a private registry, you have to consider the following three prerequisites.

  1. install the docker credential helper:
$ sudo apt install golang-docker-credential-helpers

(For ubuntu 16.04 and older this is not a package and has to be grabbed from https://github.com/docker/docker-credential-helpers/releases, install the bin in /usr/bin using the one containing "secretservice" in it's name.)

  1. Set up docker to use the credential helper: Open the file ~/.docker/config.json (or create it if it doesn't exist), and put the following content in it:
{
        "credsStore": "secretservice"
}
  1. Now login to the registry server:
$ docker login docker-reg.mydomain.de

Now you can pull images from the docker registry, e.g. by using docker pull.

The settings.bash file allows to set DOCKER_REGISTRY_AUTOPULL in order to automatically pull images, when ./exec.bash is run.