Kubernetes Cluster Add-ons - OCI

OCI Add-ons

The following OCI specific resources should be added:

oci-file-service-storage

This is an optional add-on. Refer to the Introduction - OCI chapter for additional information.

Please note that persistent volume setup is an optional step. Skip this section if you are not intended to have persistent file storage.

The OCI File Storage service provides a durable, scalable, distributed, enterprise-grade network file system.

A persistent volume claim (PVC) is a request for persistent file storage. The OCI File Storage service file systems are mounted inside containers running on clusters created by Container Engine for Kubernetes using a CSI (Container Storage Interface) volume plugin deployed on the clusters.

To enable the CSI volume plugin to create and manage File Storage resources, appropriate IAM policies must be installed:

Policy to create and/or manage file systems, mount targets, and export paths:

ALLOW any-user to manage file-family in compartment <compartment-name> where request.principal.type = 'cluster'

Policy to use VNICs, private IPs, private DNS zones, and subnets:

ALLOW any-user to use virtual-network-family in compartment <compartment-name> where request.principal.type = 'cluster'

User can use the File Storage service to provision persistent volume claims (PVCs) in two ways:

Dynamic Provisioning (Deprecated)
Static Provisioning (preferred way)

Dynamic Provisioning

These steps describe how to create a dynamically provisioned volume using OCI Volume plugin.

Prepare a storageclass.yaml file with StorageClass manifest for OCI File Storage:

kind: StorageClass
apiVersion: storage.k8s.io/v1
metadata:
  name: fss-dyn-storage
provisioner: fss.csi.oraclecloud.com
parameters:
  availabilityDomain: <availability_Domain>
  mountTargetSubnetOcid: <mount_target_subnet_ocid from terraform output>
  kmsKeyOcid: <kms_key_ocid from terraform output, omit if terraform output is empty>

kmsKeyOcid property is optional and can be omitted if data is encrypted at rest using encryption keys managed by Oracle. Only specify if user managed encryption key is used, i.e., kms_key_ocid from terraform output isn't empty.

Deploy the storage class

kubectl apply -f storageclass.yaml

For more information, please refer to the dynamic provisioning documentation.

Static Provisioning

These steps describe how to create a PVC by creating a PV backed by the new file system and then create the PVC and binds the PVC to the PV backed by the File Storage service.

Prepare a pv.yaml file with PersistentVolume manifest for OCI File Storage:

apiVersion: v1
kind: PersistentVolume
metadata:
  name: fss-pv
spec:
  capacity:
    storage: 1Gi
  volumeMode: Filesystem
  accessModes:
    - ReadWriteMany
  persistentVolumeReclaimPolicy: Delete
  csi:
    driver: fss.csi.oraclecloud.com
    volumeHandle: <filesystem_ocid from terraform output>:<mount_target_IP_address from terraform output>:<filesystem_mount_path from terraform output>

Deploy the PersistentVolume

kubectl apply -f pv.yaml

Prepare a pvc.yaml file with PersistentVolumeClaim manifest for OCI File Storage

apiVersion: v1
kind: PersistentVolumeClaim
metadata:
  name: fss-pvc
spec:
  accessModes:
    - ReadWriteMany
  storageClassName: ""
  resources:
    requests:
      storage: 1Gi
  volumeName: fss-pv

Deploy the PersistentVolumeClaim

kubectl apply -f pvc.yaml -n uepe

Verify PVC is bound to the PV successfully

kubectl get pvc -n uepe

the output below shows persistent volume claim bound to persistent volume successfully

NAME     CAPACITY   ACCESS MODES   RECLAIM POLICY   STATUS      CLAIM   STORAGECLASS   VOLUMEATTRIBUTESCLASS   REASON   AGE
fss-pv   1Gi        RWX            Delete           Available                          <unset>                          9s

Pod cannot access file system due to insufficient permissions

When a pod attempts to access a persistent volume (PV) backed by a file system in the File Storage service, the attempt might fail with a "Permission Denied" message.

To resolve the "Permission Denied”, follow these steps:

Obtain the CSIDriver configuration file

kubectl get csiDriver fss.csi.oraclecloud.com -oyaml > fss_csi_driver.yaml

Edit the fss_csi_driver.yaml file and change the CSIDriver object's spec.fsGroupPolicy attribute from ReadWriteOnceWithFSType to File. For example,

kind: CSIDriver
metadata:
  creationTimestamp: "<timestamp>"
  name: fss.csi.oraclecloud.com
  resourceVersion: "<version>"
  uid: <identifier>
spec:
  attachRequired: false
  fsGroupPolicy: File
  podInfoOnMount: false
  requiresRepublish: false
  storageCapacity: false
  volumeLifecycleModes:
  - Persistent

Delete the existing CSIDriver object

kubectl delete csiDriver fss.csi.oraclecloud.com

Create the new CSIDriver object from fss_csi_driver.yaml

kubectl apply -f fss_csi_driver.yaml

For more information, please refer to the Troubleshooting File Storage Service Provisioning of PVCs

oci-native-ingress-controller

cert-manager needs to be installed prior to oci-native-ingress-controller installation as it refers to cert-manager internally.

The simplest way to install cert-manager is via the cluster add-ons. From the console, browse to Containers > Clusters > Cluster details, scroll down to vertical menu, select Resources > Add-ons, select Manage add-ons to install and enable cert-manager.

The OCI native ingress controller implements the rules and configuration options defined in a Kubernetes ingress resource to load balance and route incoming traffic to service pods running on worker nodes in a cluster. The OCI native ingress controller creates an OCI flexible load balancer to handle requests, and configures the OCI load balancer to route requests according to the rules defined in the ingress resource.

The OCI Native Ingress controller creates the following OCI load balancer resources:

A load balancer for each IngressClass resource where you have specified the OCI native ingress controller as the controller.
A load balancer backend set for each unique Kubernetes service name and port number combination that you include in routing rules in Ingress resources in the cluster.
A routing policy that reflect the rules defined in the ingress resource, that is used to route traffic to backend set.
A load balancer listener for each unique port that you include in routing rules in Ingress resources in the cluster

To install OCI Native Ingress Controller, follow these steps:

Create a config file named user-auth-config.yaml, containing credential information, in the following format:

auth:
  region: <region-identifier>
  user: <user-ocid>
  fingerprint: <fingerprint>
  tenancy: <tenancy-ocid>

Create a Kubernetes secret resource named oci-config in the cluster by entering:

kubectl create secret generic oci-config \
--from-file=config=user-auth-config.yaml \
--from-file=private-key=<private-key-file-path>.pem \
--namespace uepe

Grant permission to the OCI Native Ingress Controller to access resources created by other OCI services, such as the Load Balancer service and the Certificates service. Hence, these IAM policies must be installed.

Allow group <group-name> to manage load-balancers in compartment <compartment-name>
Allow group <group-name> to use virtual-network-family in compartment <compartment-name>
Allow group <group-name> to manage cabundles in compartment <compartment-name>
Allow group <group-name> to manage cabundle-associations in compartment <compartment-name>
Allow group <group-name> to manage leaf-certificates in compartment <compartment-name>
Allow group <group-name> to read leaf-certificate-bundles in compartment <compartment-name>
Allow group <group-name> to manage certificate-associations in compartment <compartment-name>
Allow group <group-name> to read certificate-authorities in compartment <compartment-name>
Allow group <group-name> to manage certificate-authority-associations in compartment <compartment-name>
Allow group <group-name> to read certificate-authority-bundles in compartment <compartment-name>
Allow group <group-name> to read cluster-family in compartment <compartment-name>

ALLOW any-user to manage network-security-groups in <compartment-name> Team-Stratus where request.principal.type = 'cluster'
ALLOW any-user to manage vcns in compartment <compartment-name> where request.principal.type = 'cluster'
ALLOW any-user to manage virtual-network-family in compartment <compartment-name> where request.principal.type = 'cluster'

Allow group <group-name> to inspect certificate-authority-family in compartment <compartment-name>
Allow group <group-name> to use certificate-authority-delegate in compartment <compartment-name>
Allow group <group-name> to manage leaf-certificate-family in compartment <compartment-name>
Allow group <group-name> to use leaf-certificate-family in compartment <compartment-name>
Allow group <group-name> to use certificate-authority-delegate in compartment <compartment-name>
Allow group <group-name> to manage certificate-associations in compartment <compartment-name>
Allow group <group-name> to inspect certificate-authority-associations in compartment <compartment-name>
Allow group <group-name> to manage cabundle-associations in compartment <compartment-name>

Clone the OCI native ingress controller repository from GitHub

git clone https://github.com/oracle/oci-native-ingress-controller

In the local Git repository, navigate to the oci-native-ingress-controller directory and create a config file named oci-native-ingress-controller-values.yaml with the following content:

compartment_id: <compartment_ocid from terraform output>
subnet_id: <loadbalancer_subnet_ocid from terraform output>
cluster_id: <cluster_ocid from terraform output>
authType: user
deploymentNamespace: uepe

Generate the manifest .yaml files for the required resources

helm template --include-crds oci-native-ingress-controller helm/oci-native-ingress-controller -f oci-native-ingress-controller-values.yaml --output-dir deploy/manifests

Comment out the Namespace resource from deploy/manifests/oci-native-ingress-controller/templates/deployment.yaml (otherwise it will try to create it, but it’s already created in a previous step).

#apiVersion: v1
#kind: Namespace
#metadata:
#  name: uepe

Deploy the required resources using the manifest .yaml files

kubectl apply -f deploy/manifests/oci-native-ingress-controller/crds

kubectl apply -f deploy/manifests/oci-native-ingress-controller/templates

Confirm that OCI native ingress controller has been installed successfully

kubectl logs <pod-names> -n uepe

The logs should look like this:

I0611 03:24:13.667434       1 leaderelection.go:258] successfully acquired lease uepe/oci-native-ingress-controller
I0611 03:24:13.667480       1 server.go:81] Controller loop...
I0611 03:24:13.672076       1 auth_service.go:94] secret is retrieved from kubernetes api: oci-config
I0611 03:24:13.672463       1 auth_service.go:42] Fetching auth config provider for type: user
I0611 03:24:14.819774       1 server.go:120] CNI Type of given cluster : OCI_VCN_IP_NATIVE
I0611 03:24:14.819999       1 backend.go:374] Starting Backend controller
I0611 03:24:14.819824       1 routingpolicy.go:282] Starting Routing Policy Controller
I0611 03:24:14.819827       1 ingress.go:685] Starting Ingress controller
I0611 03:24:14.819840       1 ingressclass.go:496] Starting Ingress Class controller

Having installed the OCI native ingress controller, these Kubernetes resources need to be created in order to start using it.

IngressClassParameters
IngressClass

IngressClassParameters resource

Use the custom IngressClassParameters resource to specify details of the OCI load balancer to create for the OCI native ingress controller.

Define the resource in a .yaml file named ingress-class-params.yaml

apiVersion: "ingress.oraclecloud.com/v1beta1"
kind: IngressClassParameters
metadata:
  name: native-ic-params
  namespace: uepe
spec:
  compartmentId: "<ocid of compartment>"
  subnetId: "<loadbalancer_subnet_ocid from terraform output>"
  loadBalancerName: "native-ic-lb-<your cluster name>"
  isPrivate: false
  maxBandwidthMbps: 400
  minBandwidthMbps: 100

To create the resource, execute

kubectl create -f ingress-class-params.yaml

IngressClass resource

Use the IngressClass resource to associate an Ingress resource with the OCI native ingress controller and the IngressClassParameters resource.

Define the resource in a .yaml file named ingress-class.yaml

apiVersion: networking.k8s.io/v1
kind: IngressClass
metadata:
  name: native-ic-ingress-class
  annotations:
    ingressclass.kubernetes.io/is-default-class: "true"
    oci-native-ingress.oraclecloud.com/id: <loadbalancer_ocid from terraform output>
spec:
  controller: oci.oraclecloud.com/native-ingress-controller
  parameters:
    scope: Namespace
    namespace: uepe
    apiGroup: ingress.oraclecloud.com
    kind: ingressclassparameters
    name: native-ic-params

To create the resource, execute

kubectl create -f ingress-class.yaml

Kubernetes Add-ons

The following general Kubernetes resources should be added:

external-dns

ExternalDNS is a Kubernetes add-on that configures public DNS servers with information about exposed Kubernetes services to make them discoverable.

To install ExternalDNS, follow these steps:

Create a Kubernetes secret containing the Oracle Cloud Infrastructure user authentication details for ExternalDNS to use when connecting to the Oracle Cloud Infrastructure API to insert and update DNS records in the DNS zone. Create a credentials file named oci.yaml and populate with the following content:

auth:
  region: <region-identifier>
  tenancy: <tenancy-ocid>
  user: <user-ocid>
  key: |
    -----BEGIN RSA PRIVATE KEY-----
   <private-key>
    -----END RSA PRIVATE KEY-----
  fingerprint: <fingerprint>
  # Omit if there is not a password for the key
  passphrase: <passphrase>
compartment: <compartment-ocid>

Create a Kubernetes secret named external-dns-config from the credentials file you just created.

kubectl create secret generic external-dns-config --from-file=oci.yaml -n uepe

Create a configuration file (for example, called external-dns-deployment.yaml) to create the ExternalDNS deployment, and specify the name of the Kubernetes secret you just created.

apiVersion: v1
kind: ServiceAccount
metadata:
  name: external-dns
---
apiVersion: rbac.authorization.k8s.io/v1
kind: ClusterRole
metadata:
  name: external-dns
rules:
- apiGroups: [""]
  resources: ["services","endpoints","pods"]
  verbs: ["get","watch","list"]
- apiGroups: ["extensions","networking.k8s.io"]
  resources: ["ingresses"]
  verbs: ["get","watch","list"]
- apiGroups: [""]
  resources: ["nodes"]
  verbs: ["list"]
---
apiVersion: rbac.authorization.k8s.io/v1
kind: ClusterRoleBinding
metadata:
  name: external-dns-viewer
roleRef:
  apiGroup: rbac.authorization.k8s.io
  kind: ClusterRole
  name: external-dns
subjects:
- kind: ServiceAccount
  name: external-dns
  namespace: uepe
---
apiVersion: apps/v1
kind: Deployment
metadata:
  name: external-dns
spec:
  strategy:
    type: Recreate
  selector:
    matchLabels:
      app: external-dns
  template:
    metadata:
      labels:
        app: external-dns
    spec:
      serviceAccountName: external-dns
      containers:
      - name: external-dns
        image: k8s.gcr.io/external-dns/external-dns:v0.13.4
        args:
        - --source=service
        - --source=ingress
        - --provider=oci
        - --txt-owner-id=<cluster_dns_zone_ocid from terraform output>
        volumeMounts:
          - name: config
            mountPath: /etc/kubernetes/
      volumes:
      - name: config
        secret:
          secretName: external-dns-config

Apply the configuration file to deploy ExternalDNS

kubectl apply -f external-dns-deployment.yaml -n uepe

Confirm that external-dns has been installed successfully

kubectl logs <pod-name> -n uepe

The logs should look like this:

time="2024-06-11T05:29:19Z" level=info msg="Instantiating new Kubernetes client"
time="2024-06-11T05:29:19Z" level=info msg="Using inCluster-config based on serviceaccount-token"
time="2024-06-11T05:29:19Z" level=info msg="Created Kubernetes client https://10.96.0.1:443"
time="2024-06-11T05:29:21Z" level=info msg="All records are already up to date"

ingress-nginx-controller

This is an optional add-on. Refer to the Introduction - OCI chapter for additional information.

The Ingress NGINX Controller is an ingress controller for Kubernetes using NGINX as a reverse proxy and load balancer.

To install the Ingress NGINX Controller, follow these steps:

Add the ingress-nginx helm repository:

helm repo add ingress-nginx https://kubernetes.github.io/ingress-nginx

Update the helm repository to get the latest software:
```
helm repo update
```

Create a file called ingress-nginx-values.yaml and populate it with the following helm values:

controller:
  scope:
    enabled: true
  admissionWebhooks:
    enabled: false
  metrics:
    enabled: false
    serviceMonitor:
      enabled: false
  ingressClassResource:
    name: nginx
    enabled: true
    default: false
    controllerValue: "k8s.io/ingress-nginx"
  watchIngressWithoutClass: false
  service:
    externalTrafficPolicy: "Local"
    targetPorts:
      http: 80
      https: 443
    type: NodePort
  extraArgs:
    v: 1
serviceAccount:
  create: false

Install the ingress-nginx-controller helm chart:
```
helm install ingress-nginx ingress-nginx/ingress-nginx --version <helm chart version> -f ingress-nginx-values.yaml -n uepe
```
Where <helm chart version> is a compatible version listed in the Compatibility Matrix (4.1).

Executing helm list -A should show all add-ons added in this section. Example:

NAME                        	NAMESPACE	REVISION	UPDATED                             	STATUS  	CHART                             	APP VERSION
ingress-nginx-controller        uepe     	1       	2024-02-22 11:44:54.18561 +0800 +08 	deployed	ingress-nginx-4.9.1               	1.9.6

This section is now complete. Now proceed to the Usage Engine Private Edition Preparations - OCI section.