Deploy Workloads on Far-Edge Devices

Far-Edge services follow the same flow as normal Kubernetes nodes services. Before deployment the Far-Edge service must be uploaded to a OCI registry.

Pushing embServe Service to OCI registry

To push an service to a registry:

1. Create a new folder:

mkdir service

2. Install the oras library:

cd service
python3 -m venv .venv
. .venv/bin/activate
python3 -m pip install oras

3. Create a config.json:

{
    "architecture": "arm",
    "os": "zephyr",
    "variant": "v7"
}

4. Copy the service manifest to the same folder. Folder should have this structure:

(.venv) user@LABWS:~/Workspace/embserve-sdk/sample$ ls -la
total 20
drwxrwxr-x 2 user user 4096 Sep 26 16:16 .
drwxrwxr-x 8 user user 4096 Sep 26 13:15 ..
-rw-rw-r-- 1 user user   71 Sep 26 16:16 config.json
-rw-rw-r-- 1 user user 2399 Sep 26 16:12 service.json

5. Upload the device to an OCI registry:

oras push <https://your-image-registry.example.com>/temperature_sensor:1.0.0 --config config.json:application/vnd.oci.image.config.v1+json service.json:application/vnd.embserve.v1+json

note

Don't forget to replace with your registry.

tip

If you want the same reference OCI registry reference to provide both an embServe workload and a Container image, check out the Multi-platform Artifacts documentation.

Creating a Kubernetes deployment for a Far-Edge device

Creating a deployment targeting a Far-Edge device follows the same logic as a normal Kubernetes deployment. Let's create a deployment for the service uploaded in the previous section.

1. Create a file pod.yaml with:

apiVersion: v1
kind: Pod
metadata:
  name: temperature
spec:
  containers:
  - name: temperature
    image: <your-image-registry.example.com>/temperature_sensor:1.0.0
    imagePullPolicy: IfNotPresent
  nodeSelector:
    extra.resources.fhp/embserve: "true"
    extra.resources.fhp/temperature_sensor: "true"
  tolerations:
    - key: "fita.fhp.pt/type"
      operator: "Equal"
      value: "far-edge"
      effect: "NoSchedule"

note

Far-Edge node Kubelet contains a taint which prevents scheduling of Pods. Ensure you add the toleration shown above and filter nodes by using extra.resources.fhp/embserve: "true" as the nodeSelector.

Other labels are available that allows to filter by device capability. Inspect a Far-Edge node to understand what labels are available.

2. Deploy the Pod:

microk8s kubectl apply -f pod.yaml -n fita

3. Observe that the Pod was deployed successfully:

kubectl get pods -n fita --output wide

Which should output:

NAME          READY   STATUS    RESTARTS   AGE   IP            NODE
temperature   1/1     Running   0          37s   10.1.42.190   labnuc05-b1-node1