Application Deployment to AKS Cluster Using Helm Plugin
- Karl Henselin
Objective
The goal of this tutorial is to deploy an Apache HTTP Server using Helm Chart on an Azure Kubernetes Service (AKS) cluster through FlexDeploy using the Helm Plugin. For this tutorial, we will be setting up an automated deployment of an Apache HTTP Server using the Helm Chart which is present in this chart repository (https://charts.bitnami.com/bitnami) . The steps will include:
Configuring properties e.g. config file path, cloud account, and CLI path
Deploying the Apache HTTP Server to an AKS cluster
Verifying the application is successfully created on the AKS cluster
We will walk through each of the FlexDeploy features that will be created/configured to accomplish this goal and have the Apache HTTP Server deployed on the AKS cluster in a very short amount of time.
Checklist
Checklist | Description |
Azure Tenant Id | Azure Tenant Id for the company |
Azure Subscription Id | Azure Subscription where the AKS cluster exists |
Azure Client ID | Client ID to connect to desired Azure Subscription |
Azure Client Key | Password for the Client ID |
Azure Client ID permission | The Client ID must have relevant access in AKS cluster to do the deployment |
Azure CLI installation | Azure CLI needs to be installed where the plugin operation will run (e.g. FlexDeploy server) |
Azure CLI in class path | Azure CLI should be added to class path on the FlexDeploy Server. Else the path can also be set under FlexDeploy environment level property |
AKS Setup | AKS cluster should be created in Azure and a kubeconfig file should be present with the context of the cluster |
Configure Cloud Account
To connect with AKS cluster, we required to configure Cloud account, with credentials details. Configure Azure Cloud Account under Integration. FlexDeploy will connect to the AKS cluster and deploy the Apache HTTP Server.
Navigate to the Integrations
Select Cloud from the left-hand pane
Create a new Cloud account with the “+” button. Create a new Cloud account of provider type “Azure”
It should have a Client ID, Client Key, Tenant Id, and Subscription Id configured in it. The Client ID must have relevant access to do the deployment on AKS cluster.
Client Key is a password field and hence needs to be kept hidden. To update the same click on the pencil icon as shown below
Update the Client Key value under Secret Text. This is to make sure no one else can retrieve the password
After configuration we can use the Cloud Account.
Pre-requisite
Configure Service Principal
To access the AKS cluster, we need a Service Principal( Client ID and Client Key). Created a Service Principal and corresponding secret (e.g. aksudpatesp)
You need to provide relevant Role to the Service Principal, so that it can do necessary deployment and other relevant activities. For our tutorial we have provided Azure Kubernetes Service RBAC Write Role and Azure Kubernetes Service Cluster User Role as shown below.
Any other K8s Role or custom Azure AD Role is also acceptable as long as the Service Principal has relevant access (e.g. deploy/run/delete etc) to execute kubectl commands in target Kubernetes cluster.
Configure kubeconfig File
A Kubeconfig is a YAML file with the details to connect to Kubernetes Cluster using certificate/secret tokens to authenticate the cluster. This is essential for Helm Plugin to connect to target AKS cluster. If you need to generate a new kubeconfig file, you can find the generated file path in the output when running the first command here: Quickstart: Deploy an Azure Kubernetes Service (AKS) cluster using the Azure portal - Azure Kubernetes Service.
Why Change in kubeconfig
Normally with earlier version of kubectl, the structure of config file was as given below.
Where access-token was regenerated every one hour based on refresh-token and kubectl command used the access-token to authenticate against the k8s cluster before executing any command.
However with current version of kubectl one would get below warning message:
WARNING: the azure auth plugin is deprecated in v1.22+, unavailable in v1.26+; use GitHub - Azure/kubelogin: A Kubernetes credential (exec) plugin implementing azure authentication instead.
This is because kubectl currently supports various vendor specific login.
The logic used by the Azure auth provider is embedded in the kubectl binary, which is why you can authenticate and connect to the cluster without having to install anything else.
Going forward this approach is going to change. It’s not going to support vendor specific authentication protocols.
The authentication plugins are being moved out of the kubectl binary into separate binaries, maintained by the Cloud providers and distributed independently.
Ref:
To overcome this issue, all Cloud providers have extended the client-go credential plugin to have custom Cloud specific authentication methods.
Kubelogin is a client-go credential plugin that implements Azure AD authentication. This is the direction for AKS.
How to change in kubeconfig
Execute below command to use azurecli as kubelogin option to retrieve access token. With this changes kubelogin will internally generate the token. No update required in kubeconfig file going forward.
kubelogin convert-kubeconfig -l azurecli
CLI Installation
Azure CLI should be installed in the m/c where the plugin is to be executed. Preferably add Azure CLI path in m/c classpath.
Install/copy kubelogin on the system . Add kubelogin in m/c path variable.
Deploy Workflows
Deploy Workflow
Navigate to Workflows and create a workflow using the button as highlighted below.
Below is a sample deploy workflow to deploy the application on an AKS cluster.
Step-i: deploy the application
This step will deploy the Apache HTTP Server on the AKS cluster, using the helm chart which is present at chart repository (https://charts.bitnami.com/bitnami) . Context must be provided to apply changes. The context should be present inside kubeconfig file.
In above configuration using following Inputs, for all input and their description please ref. deploy (Helm)
Input Name | Input Value | Description |
|---|---|---|
Chart Name | apache | Name of a chart to be installed. (Reference/Package/Directory/URL). If empty takes the chart name from the project configuration. |
K8s Context | k8s-test-cluster | K8s Context where the chart will be installed |
Release Name | test-helm-new | Release name. If empty takes the chart name from the project configuration. If unspecified, it will autogenerate one for you |
Repository Name | Name of a chart repository | |
Config file | /home/oracle/.kube/azure/config | K8s config file |
Project configuration
Navigate to the Project tab and create a Project with a logical name
Configure the Deploy workflow that has been created in previous steps as shown below.
Target Properties
Select Topology from the menu and then select Targets from menu. Select the target group and environment, provide the properties detail, according to the description.
Properties | Mandatory field | Description |
Cloud Account | Optional | Select the Cloud Account to connect cluster. |
CLI Path | Optional | Directory where Cloud CLI is installed. |
Kubeconfig File Path | Optional | Absolute path of kubeconfig file |
Below given are the environment-specific values which need to be updated.
Cloud Account
The Azure Cloud account needs to be set here from the drop-down. It will show all Cloud Accounts configured under Topology, which we have already mentioned earlier.
Kubeconfig file path
Absolute path of config file can be set as environment property, if path is not set then by default plugin look for config file in user directory.
CLI path
Azure CLI path can be set as environment property, if it’s not set then by default plugin use classpath.
Override properties at Project level
Let assume a scenario, where we want to change Cloud account for any specific project. Apart from setting at environment level, it can also be set at project properties by using Override Property. Please check below mentioned steps.
Navigate to the Project Configuration tab as shown above.
Next, select the PROPERTIES option from the left-hand pane.
Click on the OVERRIDE option.
Select the Cloud Account option from Property.
Select the Environment from the drop down list.
Select the Target Group from the drop down list.
Build Execution
For detailed steps on how to initiate deploy operation using Helm plugin, please refer to Deploy through FlexDeploy using Helm plugin
Post deployment
Post deployment success from plugin it can take a little while for the deployment and running.
Once deployment successfully completed, we can see the related details on AKS cluster by using Azure console.
Congratulations! You have successfully completed the Apache HTTP Server deployment.
Now that you have configured FlexDeploy for deployment operation using Helm plugin, it is extremely easy to replicate the same for other Cloud Provider. Simply use the Copy Project feature and a new project will be created with all of the configuration completed already. You just need to make the necessary configuration changes.
- style