Kubernetes on AzureStackHub with AKS-Engine and external AD/DNS integration

0hlov3
6 min readNov 10, 2021

You can use the AKS engine command-line tool to deploy and manage a Kubernetes cluster on Azure and Azure Stack Hub. Use the AKS engine to create, upgrade, and scale Azure Resource Manager native clusters. You can use the engine to deploy a cluster in both connected and disconnected environments. This article provides an overview of the AKS engine, supported scenarios for using the engine with Azure Stack Hub, and an introduction to operations such as deploy, upgrade, and scale.

We can use AKS Engine to deploy a Kubernetes-Cluster on Azure Stack Hub. So i started to create a Standard Kubernetes-Cluster on Azure Stack Hub, but today we are talking about the ability to use a custom Windows-DNS and Windows-ADS with aks-engine.
First of all, we need to create a Virtual Network, two subnets and a ressource group.

What is AKS Engine

AKS Engine provides convenient tooling to quickly bootstrap Kubernetes clusters on Azure. By leveraging ARM (Azure Resource Manager), AKS Engine helps you create, destroy and maintain clusters provisioned with basic IaaS resources in Azure.

Check your Quotas

You need to check if you have enough space left to deploy your servers, if you choose to deploy 3 Master and 3 Worker Nodes with a VM Size of Standard_DS2_v2, you should have space for 12 CPUs.
If your subscription doesn’t contain enough quotas, your deployment will fail.

Create a Service Priciple (service account)

You need to create an Azure service principal for the deployment. I used one with Password-based authentication.

az ad sp create-for-rbac --name $AKS-SP

please store the service principal name, id and password at a save place. You can create your own keyvault for that. Cause we need the ID and password later.
The service principal needs enough privileges to customize the virtual networks and the subnets as well the privileges on the ressource group to add ressources and make changes.

Create a Virtual Network

If you don’t already have a virtual network for your AKS installation, you need to create one. You can decide where to create your virtual network, maybe you already have a network resource group. If you already habe a vnet, you can skip this step.

az network vnet create --address-prefixes 10.222.0.0/16 --name $k8sVnetName --resource-group $RG-NETWORK --subnet-name $KubernetesMasterSubnet --subnet-prefixes 10.222.1.0/25

Create a Subnet

Here you have to decide on the network size and think about the names of the subnet. The following are really only examples of what it might look like.

az network vnet subnet create -g $RG-NETWORK --vnet-name $k8sVnetName -n $KubernetesMasterSubnet --address-prefixes 10.222.1.0/25
az network vnet subnet create -g $RG-NETWORK --vnet-name $k8sVnetName -n $KubernetesWorkerSubnet --address-prefixes 10.222.1.128/25

Create a Ressource Group

We have to create a Ressource Group for our deployment of AKS if we don’t have already a RG.

az group create -l $LOCATION -g $RG-NAME-k8s

Create the deployment file

If we have taken the above steps and our service principle has the right permissions, e.g. at subscription level, we can start to prepare the deployment.
For that, we need to install an aks-engine version, after that we have to create a deployment-file kubernetes-azurestack.json.

Deploymentfile without a custom DNS-Server

{
"apiVersion": "vlabs",
"location": "$LOCATION",
"properties": {
"orchestratorProfile": {
"orchestratorType": "Kubernetes",
"orchestratorRelease": "1.17",
"orchestratorVersion": "1.17.5",
"kubernetesConfig": {
"cloudProviderBackoff": true,
"cloudProviderBackoffRetries": 1,
"cloudProviderBackoffDuration": 30,
"cloudProviderRateLimit": true,
"cloudProviderRateLimitQPS": 3,
"cloudProviderRateLimitBucket": 10,
"cloudProviderRateLimitQPSWrite": 3,
"cloudProviderRateLimitBucketWrite": 10,
"kubernetesImageBase": "mcr.microsoft.com/k8s/azurestack/core/",
"useInstanceMetadata": false,
"networkPlugin": "kubenet",
"kubeletConfig": {
"--node-status-update-frequency": "1m",
"--kube-reserved": "memory=500Mi",
"--system-reserved": "memory=500Mi",
"--eviction-hard": "memory.available<500Mi"
},
"controllerManagerConfig": {
"--node-monitor-grace-period": "5m",
"--pod-eviction-timeout": "5m",
"--route-reconciliation-period": "1m"
},
"privateCluster": {
"enabled": true
},
"etcdDiskSizeGB": "32"
}
},
"customCloudProfile": {
"portalURL": "https://$AzureStackHubURL"
},
"featureFlags": {
"enableTelemetry": true
},
"masterProfile": {
"dnsPrefix": "master",
"distro": "aks-ubuntu-16.04",
"count": 3,
"vmSize": "Standard_DS2_v2",
"vnetSubnetId": "/subscriptions/$SUBSCRIPTIONID/resourceGroups/$RG-NETWORK/providers/Microsoft.Network/virtualNetworks/$k8sVnetName/subnets/$KubernetesMasterSubnet",
"firstConsecutiveStaticIP": "..REDACTED...",
"customVMTags": {
"Application": "k8s Test"
}
},
"agentPoolProfiles": [
{
"name": "worker",
"count": 3,
"vmSize": "Standard_DS2_v2",
"distro": "aks-ubuntu-16.04",
"availabilityProfile": "AvailabilitySet",
"AcceleratedNetworkingEnabled": false,
"vnetSubnetId": "/subscriptions/$SUBSCRIPTIONID/resourceGroups/$RG-NETWORK/providers/Microsoft.Network/virtualNetworks/$k8sVnetName/subnets/$KubernetesWorkerSubnet",
"customVMTags": {
"Application": "k8s Test"
}
}
],
"linuxProfile": {
"adminUsername": "$ROOTUserName",
"ssh": {
"publicKeys": [
{
"keyData": "$ROOTUserSSHKey"
}
]
}
},
"servicePrincipalProfile": {
"clientId": "$SPID",
"secret": "$SPPassword"
}
}
}

Deploymentfile with custom DNS-Server

{
"apiVersion": "vlabs",
"location": "$LOCATION",
"properties": {
"orchestratorProfile": {
"orchestratorType": "Kubernetes",
"orchestratorRelease": "1.17",
"orchestratorVersion": "1.17.5",
"kubernetesConfig": {
"cloudProviderBackoff": true,
"cloudProviderBackoffRetries": 1,
"cloudProviderBackoffDuration": 30,
"cloudProviderRateLimit": true,
"cloudProviderRateLimitQPS": 3,
"cloudProviderRateLimitBucket": 10,
"cloudProviderRateLimitQPSWrite": 3,
"cloudProviderRateLimitBucketWrite": 10,
"kubernetesImageBase": "mcr.microsoft.com/k8s/azurestack/core/",
"useInstanceMetadata": false,
"networkPlugin": "kubenet",
"kubeletConfig": {
"--node-status-update-frequency": "1m",
"--kube-reserved": "memory=500Mi",
"--system-reserved": "memory=500Mi",
"--eviction-hard": "memory.available<500Mi"
},
"controllerManagerConfig": {
"--node-monitor-grace-period": "5m",
"--pod-eviction-timeout": "5m",
"--route-reconciliation-period": "1m"
},
"privateCluster": {
"enabled": true
},
"etcdDiskSizeGB": "32"
}
},
"customCloudProfile": {
"portalURL": "https://$AzureStackHubURL"
},
"featureFlags": {
"enableTelemetry": true
},
"masterProfile": {
"dnsPrefix": "master",
"distro": "aks-ubuntu-16.04",
"count": 3,
"vmSize": "Standard_DS2_v2",
"vnetSubnetId": "/subscriptions/$SUBSCRIPTIONID/resourceGroups/$RG-NETWORK/providers/Microsoft.Network/virtualNetworks/$k8sVnetName/subnets/$KubernetesMasterSubnet",
"firstConsecutiveStaticIP": "..REDACTED...",
"customVMTags": {
"Application": "k8s Test"
}
},
"agentPoolProfiles": [
{
"name": "worker",
"count": 3,
"vmSize": "Standard_DS2_v2",
"distro": "aks-ubuntu-16.04",
"availabilityProfile": "AvailabilitySet",
"AcceleratedNetworkingEnabled": false,
"vnetSubnetId": "/subscriptions/$SUBSCRIPTIONID/resourceGroups/$RG-NETWORK/providers/Microsoft.Network/virtualNetworks/$k8sVnetName/subnets/$KubernetesWorkerSubnet",
"customVMTags": {
"Application": "k8s Test"
}
}
],
"linuxProfile": {
"adminUsername": "$ROOTUserName",
"ssh": {
"publicKeys": [
{
"keyData": "$ROOTUserSSHKey"
}
]
},
"customSearchDomain": {
"name": "$ADDOMAINNAME",
"realmUser": "$ADSERVICEACCOUNT",
"realmPassword": "$ADSERVICEACCOUNTPASSWORD"
},
"customNodesDNS": {
"dnsServer": "$CustomDNSIP"
}
},
"servicePrincipalProfile": {
"clientId": "$SPID",
"secret": "$SPPassword"
}
}
}

Deployment

aks-engine deploy -f --azure-env AzureStackCloud --api-model kubernetes-azurestack.json --location $LOCATION --resource-group $RG-NAME-k8s --client-id $SPID --client-secret $SPPassword --subscription-id $SUBSCRIPTIONID --output-directory $RG-NAME-k8s

Hotfix

with the above deploymentfile you can start a deployment with a custom WindowsDNS, but since we have backwords compatibility for Windows 2000 and below, the Active-Directory computer section is limited to 15 chars, so we needed a ugly hotfix.
We needed to overwrite the fallowing file Node by node, so if we checked the log in /var/log/azure/cluster-provision.sh the script ist waiting for /opt/azure/containers/setup-custom-search-domains.sh to get ready, cause the #EOF isn’t included in the script at the moment.
So we overwrote the script node by node, but only, when the current server is successfully joined the AD.

cat <<EOF > /opt/azure/containers/setup-custom-search-domains.sh
#!/bin/bash
set -x
source /opt/azure/containers/provision_source.shecho " dns-search $ADDOMAINNAME" | tee -a /etc/network/interfaces.d/50-cloud-init.cfg
systemctl_restart 20 5 10 networking
wait_for_apt_locks
retrycmd 10 5 120 apt-get update
wait_for_apt_locks
retrycmd 10 5 120 apt-get -y install realmd sssd sssd-tools samba-common samba samba-common python2.7 samba-libs packagekit
wait_for_apt_locks
echo "$ADSERVICEACCOUNTPASSWORD" | realm join -U $ADSERVICEACCOUNT@$(echo "$ADDOMAINNAME" | tr /a-z/ /A-Z/) $(echo "$ADDOMAINNAME" | tr /a-z/ /A-Z/)
#EOF
EOF

You need to do it on all master and worker Nodes. Step by Step.

If your deployment was successful

If the deployment is successful, all files from the output directory $RG-NAME-k8s must be placed in a secure location, using a keyvault and a storage account with blob storage is a nice place.
Cause some files containing secrets, they should be stored safely. Microsoft recommends an extra virtual instance.

These files from the folder are needed for later changes, upgrades, scaling and other changes of the cluster.

Problem or Feature i’m not sure

I don;t know if backwards compatibility is a feature, but i think its a bug, that an older AD doesn’t work with aks-engine, so opened an issue for my problem, cause I don’t know how to solve it, concerning the backwards compatibility of the AD.

Suggestions welcome

I am trying to do my best, if you have a suggestion for improvement, feel free to write me a message.
I need good and negative criticism to improve myself, maybe you have a solution for the problem or need help yourself.

don’t trust me…

I do not know what I am doing, so I cannot take responsibility for the accuracy of the information provided here

--

--

Passionate DevOps engineer with a preference for Infrastructure as Code, Automation, Kubernetes, Go, Azure- and HetznerCloud.