Understanding StatefulSets

Semi-Manual Installation

Create a MongoDB StatefulSet.

Apply the mongo-simple.yaml file to create the StatefulSet. Then apply the mongo-service.yaml file to create the headless service.

Check that the pods start in order.

Start a shell in an Ubuntu pod (kubectl run -i --rm --tty shell --image=ubuntu -- bash), install nslookup or dig, and attempt a DNS resolution of:

  • mongo
  • mongo-1.mongo

What do you observe?

To initialize MongoDB, use the following commands:

kubectl exec -it mongo-0 -- mongo
# Inside the container
rs.initiate({_id: "rs0", members:[{_id: 0, host: "mongo-0.mongo:27017"}]});
rs.add("mongo-1.mongo:27017");
rs.add("mongo-2.mongo:27017");

Try scaling the StatefulSet using kubectl scale. What happens?

Automated Installation

What are the areas for improvement for this StatefulSet?

Answers
  • Support for scaling
  • Automatic MongoDB cluster configuration
  • Implementation of persistent storage

First, delete the MongoDB StatefulSet.

Adding an Automated Configuration Script

Apply the mongo-configmap.yaml file.

Adding Persistent Storage

Using this example as inspiration, modify the mongo.yaml file to add a volumeClaimTemplates section.

Online Solution

https://github.com/k8s-school/k8s-school/blob/master/labs/1_kubernetes/.solution/15-14-mongo-pvc.yaml

Verify that the StatefulSet has started correctly, then list the PVs and PVCs.

Also, ensure that all nodes are subscribed to the cluster using the following command:

kubectl exec -it mongo-0 -- mongo --eval="printjson(rs.status())"
Note

The configuration script in the configmap is not highly robust and has some side effects. If one of the mongo pods is not present in the mongo replicaset, a simple solution is to delete the affected pod so that it is automatically recreated and the configuration script runs again.

Data Loading

curl -O https://raw.githubusercontent.com/mongodb/docs-assets/primer-dataset/primer-dataset.json

cat primer-dataset.json | kubectl exec -it mongo-0 -- mongoimport --db test --collection restaurants --drop

Finally, test data access:

kubectl exec -it mongo-0 -- mongo test --eval "db.restaurants.find()"

Congratulations! You have worked with all the fundamental components needed to deploy an application on a Kubernetes cluster 🚀!