Docker Open Source GPU

​
Follow these steps to install HEAVY.AI as a Docker container on a machine running with supported NVIDIA GPU cards.
Preparation
Prepare your host by installing NVIDIA drivers, Docker, and NVIDIA runtime.
Install NVIDIA Drivers
To install NVIDIA drivers, open a terminal window on the host. Run apt update and apt upgrade to ensure that you are using the latest operating system software.
1
sudo apt update
2
sudo apt upgrade
Copied!
Use apt-get to install required utilities.
1
sudo apt-get install \
2
    apt-transport-https \
3
    ca-certificates \
4
    curl \
5
    gnupg-agent \
6
    software-properties-common
Copied!
Reboot your system to activate all of your changes.
1
sudo reboot
Copied!
Install CUDA
Verify that the gcc compiler is installed with the following command.
1
gcc --version
Copied!
If no version information returns, run the following command.
1
sudo apt install gcc
Copied!
To install the CUDA package.
1.Go to https://developer.nvidia.com/cuda-downloads.
2.Select the target platform by selecting the operating system (Linux), architecture (based on your environment), distribution (Ubuntu), version (based on your environment), and installer type (HEAVY.AI recommends deb (network)).
3.Install the CUDA package using the installation instructions on the NVIDIA web site.
4.Reboot your system to ensure that all changes are active:
1
sudo reboot
Copied!
Install Docker
Remove any existing Docker installs and the legacy NVIDIA docker runtime.
1
sudo docker volume ls -q -f driver=nvidia-docker | xargs -r -I{} -n1 docker ps -q -a -f volume={} | xargs -r docker rm -f
2
sudo apt-get purge nvidia-docker
Copied!
Remove Docker.
1
sudo apt-get remove docker docker-engine docker.io containerd runc
Copied!
Update with apt-get.
1
sudo apt-get update
Copied!
Use curl to download the latest Docker version.
1
sudo curl -fsSL https://download.docker.com/linux/ubuntu/gpg | sudo apt-key add -
Copied!
Add Docker to your Apt repository.
1
sudo add-apt-repository \
2
   "deb [arch=amd64] https://download.docker.com/linux/ubuntu \
3
   $(lsb_release -cs) \
4
   stable"
Copied!
Update your repository.
1
sudo apt-get update
Copied!
Install Docker, the command line interface, and the container runtime.
1
sudo apt-get install docker-ce docker-ce-cli containerd.io
Copied!
Optional: Run the following usermod command so that docker command execution does not require sudo privilege. Log out and log back in for the changes to take effect.
1
sudo usermod -aG docker $USER
Copied!
Checkpoint
Verify your Docker installation.
1
sudo docker run hello-world
Copied!
Install NVIDIA Docker runtime
Use curl to add a gpg key:
1
curl -s -L https://nvidia.github.io/nvidia-container-runtime/gpgkey | \
2
  sudo apt-key add -
3
  distribution=$(. /etc/os-release;echo $ID$VERSION_ID)
Copied!
Update your sources list:
1
curl -s -L https://nvidia.github.io/nvidia-container-runtime/$distribution/nvidia-container-runtime.list | \
2
  sudo tee /etc/apt/sources.list.d/nvidia-container-runtime.list
Copied!
Update apt-get and install nvidia-container-runtime:
1
sudo apt-get update
2
sudo apt-get install -y nvidia-container-runtime
Copied!
Edit /etc/docker/daemon.json to add the following, and save the changes:
1
{
2
 "default-runtime": "nvidia",
3
  "runtimes": {
4
     "nvidia": {
5
         "path": "/usr/bin/nvidia-container-runtime",
6
         "runtimeArgs": []
7
     }
8
 }
9
}
Copied!
Restart the Docker daemon:
1
sudo pkill -SIGHUP dockerd
Copied!
Checkpoint
Verify Docker and NVIDIA runtime work together.
1
sudo docker run --runtime=nvidia --rm nvidia/cuda:10.0-runtime-ubuntu18.04 nvidia-smi
Copied!
Standard NVIDIA-SMI output shows the GPUs in your instance.
Installation
Download HEAVY.AI from DockerHub and start HEAVY.AI in Docker.
1
sudo docker run --runtime=nvidia \
2
  -d --runtime=nvidia \
3
  -v /var/lib/heavyai:/heavyai-storage \
4
  -p 6273-6280:6273-6280 \
5
  heavyai/heavyai-ee-cuda:v5.4.2
Copied!
On startup, you receive the following error:
1
config file does not exist, ignoring --config / heavyai-storage/heavy.conf
Copied!
This is expected behavior, because HEAVY.AI does not ship with a default heavyai.conf file.

To use HEAVY.AI with the capabilities enabled by heavy.conf:
1.Stop HEAVY.AI.
2.Create the file as described in Configuration File and place it in /var/lib/heavyai.
3.Restart HEAVY.AI.
For more information on CUDA driver installation, see the CUDA Installation Guide.
See also the note regarding the CUDA JIT Cache in Optimizing Performance.
For more information on Docker installation, see the Docker Installation Guide.
Command-Line Access
You can access the command line in the Docker image to perform configuration and run HEAVY.AI utilities.
You need to know the container-id to access the command line. Use the command below to list the running containers.
1
sudo docker container ls
Copied!
You see output similar to the following.
1
CONTAINER ID        IMAGE               COMMAND                  CREATED             STATUS              PORTS                                            NAMES
2
9e01e520c30c        heavyai/heavyai-ee-cpu    "/bin/sh -c '/heavyai..."   3 days ago          Up 3 days           0.0.0.0:6273-6280->6273-6280/tcp                 confident_neumann
Copied!
Once you have your container ID, you can access the command line using the Docker exec command. For example, here is the command to start a Bash session in the Docker instance listed above. The -it switch makes the session interactive.
1
sudo docker exec -it 9e01e520c30c bash
Copied!
You can end the Bash session with the exit command.
Note
On Ubuntu, nvidia-docker has a dependency on nvidia-modprobe. If you receive the message: Error: Could not load UVM kernel module. Is nvidia-modprobe installed? you can install nvidia-modprobe using the following commands.
On Ubuntu, nvidia-docker has a dependency on nvidia-modprobe. If you receive the message: Error: Could not load UVM kernel module. Is nvidia-modprobe installed? you can install nvidia-modprobe using the following commands.
1
sudo apt-add-repository multiverse
2
sudo apt update
3
sudo apt install nvidia-modprobe
Copied!
Activation
To verify that all systems are go, load some sample data and perform an omnisql query.
HEAVY.AI ships with two sample datasets of airline flight information collected in 2008, and one dataset of New York City census information collected in 2015. To install the sample data, run the following command.
1
sudo docker exec -it <container-id> ./insert_sample_data
Copied!
Where <container-id> is the container in which HEAVY.AI is running.
When prompted, choose whether to insert dataset 1 (7,000,000 rows), dataset 2 (10,000 rows), or dataset 3 (683,000 rows). The examples below use dataset 2.
1
Enter dataset number to download, or 'q' to quit:
2
#     Dataset                   Rows    Table Name             File Name
3
1)    Flights (2008)            7M      flights_2008_7M        flights_2008_7M.tar.gz
4
2)    Flights (2008)            10k     flights_2008_10k       flights_2008_10k.tar.gz
5
3)    NYC Tree Census (2015)    683k    nyc_trees_2015_683k    nyc_trees_2015_683k.tar.gz
Copied!
Connect to HeavyDB by entering the following command (default password is HyperInteractive):
1
docker exec -it <container-id> /heavyai/bin/heavysql
2
password: ••••••••••••••••
Copied!
Enter a SQL query such as the following:
1
heavysql> SELECT origin_city AS "Origin", dest_city AS "Destination", AVG(airtime) AS
2
"Average Airtime" FROM flights_2008_10k WHERE distance < 175 GROUP BY origin_city,
3
dest_city;
Copied!
The results should be similar to the results below.
1
Origin|Destination|Average Airtime
2
Austin|Houston|33.055556
3
Norfolk|Baltimore|36.071429
4
Ft. Myers|Orlando|28.666667
5
Orlando|Ft. Myers|32.583333
6
Houston|Austin|29.611111
7
Baltimore|Norfolk|31.714286
Copied!

Docker Enterprise/Free Editions GPU

Getting Started on AWS

Last modified 2mo ago

Export as PDF

Copy link

Contents

Preparation

Install NVIDIA Drivers

Install CUDA

Install Docker

Checkpoint

Install NVIDIA Docker runtime

Installation

Command-Line Access

Activation