Skip to content

README.md

Latest commit

 

History

History
59 lines (45 loc) · 2.27 KB

README.md

File metadata and controls

59 lines (45 loc) · 2.27 KB

Running the edge-endpoint via k3s on a balena device

Setup

First, install the balena CLI. You can find installation instructions for your platform at https://github.com/balena-io/balena-cli/blob/master/INSTALL.md

Once installed, log in to your balena account:

balena login

Deploying the edge-endpoint to a balena-managed device

Deploying to a CPU-only device

Tested using an EC2 m6 instance with 64GB disk. Also known to work on a Raspberry Pi 5 with 8GB of RAM.

From the root of edge-endpoint, run:

./deploy_balena.sh <my-fleet> cpu

Deploying to an x86 device with an NVIDIA GPU

Tested using an HP Victus Laptop with Intel Core i5 processor and NVIDIA GeForce RTX 3050 Laptop GPU.

From the root of edge-endpoint, run:

./deploy_balena.sh <my-fleet> gpu [balena_os_version]

The optional balena_os_version argument must match the balenaOS version running on the device (defaults to 6.0.24%2Brev1). This is needed to compile NVIDIA kernel modules against the correct kernel headers.

Deploying to a Jetson Orin device

From the root of edge-endpoint, run:

./deploy_balena.sh <my-fleet> jetson-orin

What gets deployed

All flavors deploy two services:

  1. A server service running a k3s server as the cluster node
  2. A bastion service that installs the edge-endpoint via Helm and provides kubectl/helm access

For GPU builds, NVIDIA kernel modules are compiled as part of the server image (multi-stage Docker build) and loaded at runtime before k3s starts. No separate GPU container is needed.

Configuration

Configure the following variables via the <fleet>/Variables or <device>/Device Variables interfaces on the BalenaCloud dashboard:

GROUNDLIGHT_API_TOKEN - Required. The Groundlight API token for authorization.
RUN_EDGE_ENDPOINT    - Set to "1" to deploy the edge-endpoint pods via Helm.
EDGE_CONFIG          - Optional. YAML contents for edge-config.
GROUNDLIGHT_ENDPOINT - Optional. Override the upstream Groundlight API endpoint.
EDGE_ENDPOINT_VALUES - Optional. Comma-separated key=value pairs passed as helm --set flags.

The bastion automatically detects the inference flavor (cpu/gpu/jetson) from the server container and configures the Helm deployment accordingly.