charts/charts/stable/smarter-device-manager

smarter-device-manager

Manage hardware resource allocation without a need for privileged containers

This chart is not maintained by the upstream project and any issues with the chart should be raised here

Source Code

• https://gitlab.com/arm-research/smarter/smarter-device-manager/
• https://github.com/k8s-at-home/charts

Requirements

Kubernetes: >=1.16.0-0

Dependencies

Repository	Name	Version
https://library-charts.k8s-at-home.com	common	4.4.0

TL;DR

helm repo add k8s-at-home https://k8s-at-home.com/charts/
helm repo update
helm install smarter-device-manager k8s-at-home/smarter-device-manager

Installing the Chart

To install the chart with the release name smarter-device-manager

helm install smarter-device-manager k8s-at-home/smarter-device-manager

Uninstalling the Chart

To uninstall the smarter-device-manager deployment

helm uninstall smarter-device-manager

The command removes all the Kubernetes components associated with the chart including persistent volumes and deletes the release.

Configuration

Read through the values.yaml file. It has several commented out suggested values. Other values may be used from the values.yaml from the common library.

Specify each parameter using the --set key=value[,key=value] argument to helm install.

helm install smarter-device-manager \
  --set env.TZ="America/New York" \
    k8s-at-home/smarter-device-manager

Alternatively, a YAML file that specifies the values for the above parameters can be provided while installing the chart.

helm install smarter-device-manager k8s-at-home/smarter-device-manager -f values.yaml

Custom configuration

Please consider overriding the default configuration through values.yaml file. For example:

config: |
  - devicematch: ^snd$
    nummaxdevices: 20
  - devicematch: ^rtc0$
    nummaxdevices: 20
  - devicematch: ^video[0-9]*$
    nummaxdevices: 20
  - devicematch: ^ttyACM[0-9]*$
    nummaxdevices: 1  

The above configuration allows up to twenty pods to make use of /dev/snd, /dev/rtc0 and each of /dev/video[0-9]* devices. Only one pod at a time can use any of the discovered host /dev/ttyACM[0-9]* devices. This means that if a pod has reserved a dongle at /dev/ttyACM0, another pod requesting the same device will stay pending.

Once the devices are discovered, they will be added as allocatable resources to the nodes:

 $ kubectl describe nodes
...
Capacity:
  cpu:                                  4
  memory:                               16104560Ki
  pods:                                 110
  smarter-devices/gpiochip0:            0
  smarter-devices/i2c-0:                0
  smarter-devices/snd:                  20
  smarter-devices/ttyUSB-Z-Stick-Gen5:  1

The hardware is requested by pods through resources, e.g.:

  resources:
    limits:
      smarter-devices/ttyUSB-Conbee-2: 1
    requests:
      smarter-devices/ttyUSB-Conbee-2: 1

In this case host device /dev/ttyUSB-Conbee-2 will be given at the same path, and the pod will only be created when this device is available.

Values

Important: When deploying an application Helm chart you can add more values from our common library chart here

Key	Type	Default	Description
config	string	See values.yaml	Override default configuration See project repo for configuration examples.
dnsPolicy	string	nil	Defaults to "ClusterFirst" if hostNetwork is false and "ClusterFirstWithHostNet" if hostNetwork is true.
hostNetwork	bool	true	When using hostNetwork make sure you set dnsPolicy to ClusterFirstWithHostNet
image.pullPolicy	string	"IfNotPresent"	image pull policy
image.repository	string	"registry.gitlab.com/arm-research/smarter/smarter-device-manager"	image repository
image.tag	string	"v1.20.7"	image tag
persistence	object	See values.yaml	Configure persistence settings for the chart under this key.
priorityClassName	string	"system-node-critical"	Custom priority class for different treatment by the scheduler Setting this is not necessary, but it is recommended. [ref]
securityContext	object	See values.yaml	Configure the securityContext for this pod [ref]
termination.messagePath	string	"/var/log/termination-log"	Configure the path at which the file to which the main container's termination message will be written. Overrides the default of /dev/termination-log to allow read-only persistence.devfs at /dev. [ref]
termination.messagePolicy	string	"FallbackToLogsOnError"	Indicate how the main container's termination message should be populated. Valid options are File and FallbackToLogsOnError. smarter-device-manager does not support a termination-log, so use the container's log. [ref]

Changelog

Version 6.4.0

Added

N/A

Changed

• Upgraded common chart dependency to version 4.4.0

Fixed

N/A

Older versions

A historical overview of changes can be found on ArtifactHUB

Support

• See the Docs
• Open an issue
• Ask a question
• Join our Discord community

---

Autogenerated from chart metadata using helm-docs v0.1.1