Minikube Compute-to-Data Environment


  • functioning internet-accessable provider service
  • machine capable of running compute (e.g. we used a machine with 8 CPUs, 16 GB Ram, 100GB SSD and fast internet connection)
  • Ubuntu 20.04

Install Docker and Git

sudo apt update
sudo apt install git
sudo usermod -aG docker $USER && newgrp docker

Install Minikube

wget -q --show-progress
sudo dpkg -i minikube_1.22.0-0_amd64.deb

Start Minikube

First command is imporant, and solves a PersistentVolumeClaims problem.

minikube config set kubernetes-version v1.16.0
minikube start --cni=calico --driver=docker --container-runtime=docker

Depending on the number of available CPUs, RAM, and the required resources for running the job, consider adding options --cpu, --memory, and --disk-size to avoid runtime issues.

For other options to run minikube refer to this link

Install kubectl

curl -LO "$(curl -L -s"
curl -LO "$(curl -L -s"
echo "$(<kubectl.sha256) kubectl" | sha256sum --check

sudo install -o root -g root -m 0755 kubectl /usr/local/bin/kubectl

Wait untill all the defaults are running (1/1).

watch kubectl get pods --all-namespaces

Run IPFS host

export ipfs_staging=~/ipfs_staging
export ipfs_data=~/ipfs_data

docker run -d --name ipfs_host -v $ipfs_staging:/export -v $ipfs_data:/data/ipfs -p 4001:4001 -p 4001:4001/udp -p -p ipfs/go-ipfs:latest

sudo /bin/sh -c 'echo "    youripfsserver" >> /etc/hosts'

Storage class (Optional)

For minikube, you can use the default ‘standard’ class.

For AWS, please make sure that your class allocates volumes in the same region and zone in which you are running your pods.

We created our own ‘standard’ class in AWS:

kubectl get storageclass standard -o yaml
- matchLabelExpressions:
    - key:
          - us-east-1a
kind: StorageClass
    fsType: ext4
    type: gp2
reclaimPolicy: Delete
volumeBindingMode: Immediate

For more information, please visit

Download and Configure Operator Service

Open new terminal and run the command below.

git clone

Edit operator-service/kubernetes/postgres-configmap.yaml. Change POSTGRES_PASSWORD to nice long random password.

Edit operator-service/kubernetes/deployment.yaml. Optionally change:

  • add requests_cpu
  • add requests_memory
  • add limits_cpu
  • add limits_memory
      - env:
        - name: requests_cpu
          value: "4"
        - name: requests_memory
          value: "8Gi"
        - name: limits_cpu
          value: "8"
        - name: limits_memory
          value: "15Gi"
        - name: ALGO_POD_TIMEOUT
          value: "3600"

Download and Configure Operator Engine

git clone

Check the README section of operator engine to customize your deployment.

At a minimum you should add your IPFS URLs or AWS settings, and add (or remove) notification URLs.

Create namespaces

kubectl create ns ocean-operator
kubectl create ns ocean-compute

Deploy Operator Service

kubectl config set-context --current --namespace ocean-operator
kubectl create -f operator-service/kubernetes/postgres-configmap.yaml
kubectl create -f operator-service/kubernetes/postgres-storage.yaml
kubectl create -f operator-service/kubernetes/postgres-deployment.yaml
kubectl create -f operator-service/kubernetes/postgresql-service.yaml
kubectl apply  -f operator-service/kubernetes/deployment.yaml

Deploy Operator Engine

kubectl config set-context --current --namespace ocean-compute
kubectl apply  -f operator-engine/kubernetes/sa.yml
kubectl apply  -f operator-engine/kubernetes/binding.yml
kubectl apply  -f operator-engine/kubernetes/operator.yml
kubectl create -f operator-service/kubernetes/postgres-configmap.yaml

Optional: For production enviroments, it’s safer to block access to metadata. To do so run the below command:

kubectl -n ocean-compute apply -f /ocean/operator-engine/kubernetes/egress.yaml

Expose Operator Service

kubectl expose deployment operator-api --namespace=ocean-operator --port=8050

Run a port forward or create your ingress service and setup DNS and certificates (not covered here):

kubectl -n ocean-operator port-forward svc/operator-api 8050

Alternatively you could use another method to communicate between the C2D Environment and the provider, such as an SSH tunnel.

Initialize database

If your minikube is running on

curl -X POST "" -H  "accept: application/json"

Update Provider

Update your provider service by updating the operator_service.url value in config.ini

operator_service.url =

Restart your provider service.

Watch the explanatory video for more details