Version information
This version is compatible with:
- Puppet Enterprise 2019.8.x, 2019.7.x, 2019.5.x, 2019.4.x, 2019.3.x, 2019.2.x, 2019.1.x, 2019.0.x, 2018.1.x, 2017.3.x
- Puppet >= 5.0.0 < 7.0.0
- Ubuntu
Start using this module
Documentation
nvidia_docker_runtime
This is a PDK based module https://puppet.com/pdk/latest/pdk_generating_modules.html .
Table of Contents
Description
Modifies Docker for use with NVIDIA GPU based containers, including in swarm mode.
Usage
class { 'nvidia_docker_runtime':
driver_version => '440.64.00-1',
nvidia_docker2_version => '2.2.2-1',
}
Limitations
Since GPU UUIDs are needed to specify node-generic-resources
, and these are not available through facts until
the NVIDIA driver has been installed, it takes two applies to fully setup. Until then swarm cannot use gpu resources.
Due to Docker limitations you can only specify gpus in compose files for compose format 2.3. As docker stack
requires format 3.0 this means you can't use gpus in stacks.
Reference
Table of Contents
Classes
nvidia_docker_runtime
: Allows the use ofdocker run --runtime=nvidia ...
Classes
nvidia_docker_runtime
Installs NVIDIA runtime for Docker and the required NVIDIA/CUDA drivers
Driver versions for CUDA versions: https://docs.nvidia.com/deploy/cuda-compatibility/index.html
Examples
include nvidia_docker_runtime
Parameters
The following parameters are available in the nvidia_docker_runtime
class.
driver_version
Data type: String
NVIDIA/CUDA driver version, for exmaple: 440.64.00-1
. Use to lock down to a specific version. Default: installed
Default value: installed
nvidia_docker2_version
Data type: String
nvidia_docker2 version, for example: 2.2.2-1
. Use to lock down to a specific version. Default: installed
Default value: installed
Dependencies
- puppetlabs/stdlib (>=4.24.0 <7.0.0)
- puppetlabs/augeas_core (>= 1.0.0 < 2.0.0)
- puppetlabs/apt (>= 6.3.0 < 8.0.0)
- puppetlabs/docker (>= 3.5.0 < 4.0.0)