/qiskitplayground

Primary LanguageGo

Qiskit playground operator

Prereqs

Build with the SDK operator. For installation, follow Installation guide

Development

Create project

operator-sdk init --domain ibm.com --repo github.io/blublinsky/qiskitplaygrounds

Install additional libraries

go get github.com/go-logr/logr@v0.3.0
go get github.com/onsi/ginkgo@v1.14.1
go get github.com/onsi/gomega@v1.10.2
go get k8s.io/api/core/v1@v0.19.2
go get github.com/openshift/api/route/v1

Create APIs

operator-sdk create api --group qiskit --version v1alpha1 --kind QiskitPlayground --resource --controller

Overall approach

Implementation is based on Jupyter images A simple DockerFile looks like follows:

# Start from JUpiter image
FROM jupyter/scipy-notebook
# Install Qiskit
RUN pip install qiskit[visualization]

With this in place build image with the following command

docker build -t qiskit .

The CRD takes the following parameters:

	Image string `json:"image,omitempty"`
	ImagePullPolicy apiv1.PullPolicy `json:"imagePullPolicy,omitempty"`
	PVC string `json:"pvc,omitempty"`
	Resources *apiv1.ResourceRequirements `json:"resources,omitempty"`

The only required parameter is image, for example:

apiVersion: "qiskit.ibm.com/v1alpha1"
kind: QiskitPlayground
metadata:
  name: test
  namespace: jupyter
spec:
  image: "jupyter/scipy-notebook:latest"

If not provided Image pull policy defaults to IfNotPresent PVC is a PVC which will be used for persistence of notebooks, for example:

apiVersion: "qiskit.ibm.com/v1alpha1"
kind: QiskitPlayground
metadata:
  name: test
  namespace: jupyter
spec:
  image: "jupyter/scipy-notebook:latest"
  pvc: "claim-admin"

If PVC is not defined, an internal POd disk is used

Finally resource allow to specify resources for the pod, for example:

apiVersion: "qiskit.ibm.com/v1alpha1"
kind: QiskitPlayground
metadata:
  name: test
  namespace: jupyter
spec:
  image: "jupyter/scipy-notebook:latest"
  pvc: "claim-admin"
  resources:
    requests:
      memory: "64Mi"
      cpu: "250m"
    limits:
      memory: "128Mi"
      cpu: "500m"

To make implementation more reliable, operator is using a deployment with 1 replica, which means that if the pod is deleted, it will be bring back by deployment.

When playground CR is deleted, it will delete all of the associated resources.

Operator works on both OpenShift and plain vanila clusters. If running on OpenShift, a route is created for Playgrond.