Presto Helm Charts¶

Use Helm to Deply a Presto Cluster¶

The Presto community also provides the helm chart to install Presto cluster on a Kubernetes cluster. Here is how you can leverage it to create a Presto cluster containing the same settings for this workshop.

Prerequisite¶

In order to use Presto helm chart, make sure the following tools are available

Helm
A Kubernetes cluster: use kind for the lab
kubectl

Optional: Create a kind Cluster¶

Use the following command to create a kind cluster named presto:

kind create cluster -n presto

Instructions¶

Add the Presto Helm Chart Repository:

Helm chart repository is a location where packaged charts can be stored and shared. Use the following command to add Presto charts repository to Helm client configuration:
```
helm repo add presto https://prestodb.github.io/presto-helm-charts
```
List the latest stable versions of available Helm charts with the command:
```
helm search repo presto
```
Sample outputs:
```
NAME            CHART VERSION   APP VERSION     DESCRIPTION
presto/presto   0.4.0           0.291           The official Helm chart for Presto
```
Create a namespace for the Presto cluster

Use a dedicate namespace to deploy the Presto cluster. In this case, the namespace is called presto
```
kubectl create ns presto
```

To avoid Docker hub rate limitation, create a default Docker image pull secret based on your docker config. Make sure the docker config file exist by running the following command:

docker login

Create the secret using the docker config file:

kubectl create secret generic regcred \
       --from-file=.dockerconfigjson=/path/to/your/.docker/config.json \
       --type=kubernetes.io/dockerconfigjson \
       --namespace presto

Make the secret as the default secret for pulling image:

kubectl patch serviceaccount default -p '{"imagePullSecrets": [{"name": "regcred"}]}' \
   --namespace presto

Deploy MongoDB and MySQL

kubectl apply -n presto -k helm/kustomize

Populate the Data into MongoDB and MySQL

kubectl exec -n presto -ti svc/presto-mongo mongosh < data/mongo.sql
kubectl exec -n presto -ti svc/presto-mysql -- mysql -u root -ppresto < data/mysql.sql

Install a Presto Chart

Install a Presto Helm chart and it will provision a Presto cluster using the settings from the helm/values.yaml

helm install presto presto/presto -f helm/values.yml -n presto

See the refernece of the values.yml here

Wait for a while for the Presto pods, including coordinator and workers, to be ready:

NAME                                  READY   STATUS    RESTARTS   AGE
presto-coordinator-68dff7c554-9dxbw   1/1     Running   0          20h
presto-mongo-846c4ff7c5-j5thm         1/1     Running   0          20h
presto-mysql-765d566f59-jtqtg         1/1     Running   0          20h
presto-worker-9f99c6d69-d4djp         1/1     Running   0          20h
presto-worker-9f99c6d69-hnxqx         1/1     Running   0          20h
presto-worker-9f99c6d69-thslm         1/1     Running   0          20h

The configuration would bring up 1 coordinator and 3 workers pods.

Access the Presto UI

Run the following command to port-forward the port 8080 of the presto service:
```
kubectl port-forward -n presto svc/presto --address 0.0.0.0 8080:08080
```
Then you can access the Presto UI at http://localhost:808