phoenixai-kubernetes-operator

Deploy Operator, PhoenixAI Cluster and Anywhere Console by kube-anywhere Chart

Helm is a package manager for Kubernetes. A Helm Chart is a Helm package and contains all of the resource definitions necessary to run an application on a Kubernetes cluster. This topic describes how to use Helm to automatically deploy a PhoenixAI operator and cluster on a Kubernetes cluster.

The kube-anywhere chart carries three subcharts:

operator — the PhoenixAI operator;
phoenixai — a PhoenixAI cluster (the PhoenixAICluster custom resource);
anywhere — the PhoenixAI Anywhere console, an operations & usage console for the deployed clusters. It is opt-in (anywhere.enabled=true) because it requires S3-compatible object storage (anywhere.dependencies.s3); see its own README for the values.

Before you begin

Install kube-anywhere Chart

Add the Helm Chart Repo for PhoenixAI. The Helm Chart contains the definitions of the PhoenixAI Operator and the custom resource PhoenixAICluster.

Add the Helm Chart Repo.

helm repo add phoenixai https://celerdata.github.io/phoenixai-kubernetes-operator

Update the Helm Chart Repo to the latest version.
```
helm repo update phoenixai
```

View the Helm Chart Repo that you added.

$ helm search repo phoenixai
NAME                          CHART VERSION    APP VERSION  DESCRIPTION
phoenixai/anywhere            2.0.0            v2.0.0       A Helm chart for PhoenixAI Anywhere — a read-only operations & usage console
phoenixai/kube-anywhere       2.0.0            4.1.5-ee     kube-anywhere includes three subcharts, operator, phoenixai and anywhere
phoenixai/operator            2.0.0            2.0.0        A Helm chart for PhoenixAI operator
phoenixai/phoenixai           2.0.0            4.1.5-ee     A Helm chart for PhoenixAI cluster
phoenixai/warehouse           2.0.0            4.1.5-ee     Warehouse is a feature of the PhoenixAI Enterprise Edition

kube-anywhere is the install path this guide follows: it carries the console as a subchart, enabled with anywhere.enabled=true. The separately listed phoenixai/anywhere is the same console packaged on its own, for installing it next to an operator that is managed separately — do not install both.

Use the default values.yaml of the Helm Chart to deploy the PhoenixAI Operator and PhoenixAI cluster, or create a YAML file to customize your deployment configurations.
1. Deployment with default configurations
  
  Run the following command to deploy the PhoenixAI Operator and the PhoenixAI cluster:
```
$ helm install kube-anywhere phoenixai/kube-anywhere
# If the following result is returned, the PhoenixAI Operator and PhoenixAI cluster are being deployed.
NAME: kube-anywhere
LAST DEPLOYED: Tue Aug 15 15:12:00 2023
NAMESPACE: phoenixai
STATUS: deployed
REVISION: 1
TEST SUITE: None
```
2. Deployment with custom configurations
  - Create a YAML file, for example, my-values.yaml, and customize the configurations for the PhoenixAI Operator and PhoenixAI cluster in the YAML file. For the supported parameters and descriptions, see the comments in the default values.yaml of the Helm Chart.
  - Run the following command to deploy the PhoenixAI Operator and PhoenixAI cluster with the custom configurations in my-values.yaml.
```
helm install -f my-values.yaml kube-anywhere phoenixai/kube-anywhere
```
Deployment takes a while. During this period, you can check the deployment status by using the prompt command in the returned result of the deployment command above. The default prompt command is as follows:
```
 $ kubectl --namespace default get phoenixaicluster -l "cluster=kube-anywhere"
 # If the following result is returned, the deployment has been successfully completed.
 NAME            FESTATUS   CNSTATUS
 kube-anywhere   running    running
```
You can also run kubectl get pods to check the deployment status. If all Pods are in the Running state and all containers within the Pods are READY, the deployment has been successfully completed.
```
 $ kubectl get pods
 NAME                                       READY   STATUS    RESTARTS   AGE
 kube-anywhere-cn-0                         1/1     Running   0          2m50s
 kube-anywhere-fe-0                         1/1     Running   0          4m31s
 kube-anywhere-operator-69c5c64595-pc7fv    1/1     Running   0          4m50s
```
Note: the kube-anywhere prefix on the resource names above comes from the operator and phoenixai subcharts’ nameOverride (which also sets the cluster name), not from the chart name — the two just happen to match. A release installed with an older chart version, whose default was kube-phoenixai, keeps that prefix on upgrade unless you change nameOverride, which would rename every resource and recreate the cluster.

Upgrade kube-anywhere Chart

If you need to upgrade the PhoenixAI Operator and PhoenixAI cluster, run the following command:

helm upgrade -f my-values.yaml kube-anywhere phoenixai/kube-anywhere

Uninstall kube-anywhere Chart

If you need to uninstall the PhoenixAI Operator and PhoenixAI cluster, run the following command:

helm uninstall kube-anywhere

Search Helm Chart maintained by PhoenixAI on Artifact Hub. See kube-anywhere.