Skip to Content
Technical Articles
Author's profile photo Bartosz Jarkowski

Your SAP on Azure – Part 6 – SAP HANA Express on Azure Kubernetes Cluster (AKS)

Using a docker image to install SAP HANA express edition can shorten the deployment time and ensure the consistency between environments. The easy way to use it is to build a Kubernetes cluster using Microsoft Azure Container Service and deploy containers in the cloud.

Did you know that you can like this post? It’s the easiest way to show your support! Just scroll up a bit and click on the big Like button.
Thanks!

A docker container is a package of libraries and system settings required to run an application. It allows to save the time needed to provide a working environment and you can focus on the target database configuration. It’s great especially in environments where you need to provide separated HANA instances for many developers.

Azure Container Service simplifies the creation and configuration of the Kubernetes cluster and management of the entire docker environment. The nodes of the cluster are managed by Azure while your responsibility is to maintain the running application.

CREATE THE KUBERNETES CLUSTER

Creation of Kubernetes cluster in Microsoft Azure is a relatively easy task. During the initial configuration, you will be asked to provide a service principal that will be used to manage the Azure resources. Log in to the portal, go to the Azure Active Directory and create new application registration:

Save the settings. It is not important what you type in the Sign-on URL. Generate the key in the application settings – copy it together with the application ID – you will be asked for those details in few minutes.

To deploy Kubernetes cluster you need to create an Azure Cluster Service (preview). In the first step, you are asked to choose a cluster name and select a resource group in which it will be created.

You need to provide the basic configuration on the second screen. In the Service Principal ID and Service Principal Client Secret enter the information generated during the app registration. Choose the number of nodes and their size – I chose two DS11_V2 servers which fulfill the SAP HANA database memory and CPU requirements:

To connect to the cluster you require the Azure CLI. You need also to install the AKS libraries.

az aks install-cli

Log in to your Azure account and connect with the Kubernetes cluster

az login 
az aks get-credentials --resource-group=<resource group name> --name=<cluster name>

Once we have established the connection we can display the Kubernetes cluster nodes:

kubectl get nodes

You can validate the information in the Azure portal:

The nodes of the clusters are standard virtual machines in a single Availability Set:

DEPLOY SAP HANA

Downloading an SAP HANA database image from the Docker website requires an authentication. Provide your username and password to create a secret:

kubectl create secret docker-registry docker-secret --docker-server=https://index.docker.io/v1/ --docker-username=<username> --docker-password=<password> --docker-email=<e-mail>

Copy the deployment script and save it to your local drive:

kind: ConfigMap
apiVersion: v1
metadata:
  creationTimestamp: 2018-01-18T19:14:38Z
  name: hxe-pass
data:
  password.json: |+
    {"master_password" : "HXEHana1"}
---
kind: PersistentVolume
apiVersion: v1
metadata:
  name: persistent-vol-hxe
  labels:
    type: local
spec:
  storageClassName: manual
  capacity:
    storage: 150Gi
  accessModes:
    - ReadWriteOnce
  hostPath:
    path: "/data/hxe_pv"
---
kind: PersistentVolumeClaim
apiVersion: v1
metadata:
  name: hxe-pvc
spec:
  storageClassName: manual
  accessModes:
    - ReadWriteOnce
  resources:
    requests:
      storage: 50Gi
---
apiVersion: v1
kind: Pod
metadata:
  name: hxe-pod
  labels:
    name: hxe-pod
spec:
  initContainers:
    - name: install
      image: busybox
      command: [ 'sh', '-c', 'chown 12000:79 /hana/mounts' ]
      volumeMounts:
        - name: hxe-data
          mountPath: /hana/mounts
  restartPolicy: OnFailure
  volumes:
    - name: hxe-data
      persistentVolumeClaim:
         claimName: hxe-pvc
    - name: hxe-config
      configMap:
         name: hxe-pass
  imagePullSecrets:
  - name: docker-secret
  containers:
  - name: hxe-container
    image: "store/saplabs/hanaexpress:2.00.022.00.20171211.1"
    ports:
      - containerPort: 39013
        name: port1
      - containerPort: 39015
        name: port2
      - containerPort: 39017
        name: port3
      - containerPort: 8090
        name: port4
      - containerPort: 39041
        name: port5
      - containerPort: 59013
        name: port6
    args: [ "--agree-to-sap-license", "--dont-check-system", "--passwords-url", "file:///hana/hxeconfig/password.json" ]
    volumeMounts:
      - name: hxe-data
        mountPath: /hana/mounts
      - name: hxe-config
        mountPath: /hana/hxeconfig

(source: sap.com)

Deploy the image using the command:

kubectl create -f hana.yaml

The deployment takes several minutes to finish and can be monitored using the below command. If you see the message Started Container it means the process is completed.

kubectl describe pod hana-pod

You can now log in to the container and verify that the instance is running:

kubectl exec -it hxe-pod bash
HDB info
hdbsql -i 90 -d HXE -u SYSTEM -p <password>

You can see on which node the pod is running by executing:

kubectl get pods -o wide

NODE SELECTION: DIRECT ASSIGNMENT

You can directly assign the node to which the container should be deployed by a Node Selector segment:

  containers:
  - name: hxe-container
    image: "store/saplabs/hanaexpress:2.00.022.00.20171211.1"
    ports:
      - containerPort: 39013
        name: port1
      - containerPort: 39015
        name: port2
      - containerPort: 39017
        name: port3
      - containerPort: 8090
        name: port4
      - containerPort: 39041
        name: port5
      - containerPort: 59013
        name: port6
    args: [ "--agree-to-sap-license", "--dont-check-system", "--passwords-url", "file:///hana/hxeconfig/password.json" ]
    volumeMounts:
      - name: hxe-data
        mountPath: /hana/mounts
      - name: hxe-config
        mountPath: /hana/hxeconfig
  nodeSelector:
    kubernetes.io/hostname: aks-agentpool-25335148-1

Deploy the cluster using the modified configuration file.

NODE SELECTION: MEMORY REQUIREMENTS

The other possibility to help the cluster to choose good node is to specify a minimum memory requirement. The virtual machine with SAP HANA express edition requires at least 8GB of memory, but as the docker should consume a smaller amount of RAM I have requested only 7GB.

  containers:
  - name: hxe-container
    image: "store/saplabs/hanaexpress:2.00.022.00.20171211.1"
    ports:
      - containerPort: 39013
        name: port1
      - containerPort: 39015
        name: port2
      - containerPort: 39017
        name: port3
      - containerPort: 8090
        name: port4
      - containerPort: 39041
        name: port5
      - containerPort: 59013
        name: port6
    args: [ "--agree-to-sap-license", "--dont-check-system", "--passwords-url", "file:///hana/hxeconfig/password.json" ]
    volumeMounts:
      - name: hxe-data
        mountPath: /hana/mounts
      - name: hxe-config
        mountPath: /hana/hxeconfig
    resources:
      requests:
        memory: "7Gi"

The current hardware utilization can be displayed using:

kubectl top nodes

The two previously deployed containers consume more than 10 GB of memory on node 1, therefore, the cluster creates the third HANA instance on node 0.

SCALE-OUT THE KUBERENETES CLUSTER

Let’s try to create one more instance:

As there is not enough available memory on any of the node, the container was not deployed and has status pending. In that case, you can scale-out the Kubernetes cluster and add the third node:

az aks scale --name <resource name> --resource-group <resource group> --node-count <nodes>

When we check the pod status again, we can see that the hxe-pod4 is assigned to the newly created node 2.

KUBERNETES CLUSTER DASHBOARD

Instead of using the command line interface some tasks can be executed from the Kubernetes Dashboard. The bellow command creates a proxy to the Kubernetes engine in Azure and allows you to contact the webpage through a localhost:

az aks browse --resource-group <resource group> --name <cluster name>

Open a browser and navigate to http://127.0.0.1:8001/ to display the dashboard.

REMOTE DATABASE ACCESS

If you wish to access the database from the Internet you can configure the load balancer. Execution of bellow command creates a new service and assigns the Public IP.

kubectl expose pod <pod name> --name=<service name> --type=LoadBalancer
kubectl get service <service name>

Assigned Tags

      6 Comments
      You must be Logged on to comment or reply to a post.
      Author's profile photo Krzysztof Spychalski
      Krzysztof Spychalski

      Nice blog Bartosz!

       

      Author's profile photo Frank Schuler
      Frank Schuler

      Absolutely brilliant blog again, Bartosz! I like it 😉

      With my Azure Free Trial license I had difficulties instantiating both nodes at once, but starting with one node and then extending to two nodes worked fine.

      Very best regards

      Frank

      Author's profile photo Bartosz Jarkowski
      Bartosz Jarkowski
      Blog Post Author

      Hello Frank,

      thank you!

      Thanks for sharing your solution! I hope you didn't encounter any other issues 🙂

      Best regards

      Bartosz

       

      Author's profile photo Chan Jin Park
      Chan Jin Park

       

      Thanks for your another great blog. I just tried with latest AKS and i am not able to get nodes name via CLI.

       

      PS C:\Users\i032156> kubectl get nodes
      Unable to connect to the server: dial tcp: lookup hxe-6ffa6e2a.hcp.southeastasia.azmk8s.io: no such host
      PS C:\Users\i032156> az aks browse --resource-group CSH --name csh
      Merged "csh" as current context in C:\Users\i032156\AppData\Local\Temp\tmp56xlwcyy
      Proxy running on http://127.0.0.1:8001/
      Press CTRL+C to close the tunnel...
      Forwarding from 127.0.0.1:8001 -> 9090

      However, i am able to launch a AKS dashboard and see the two nodes are there.

      and then i can see my cks has a different AP server address  of hxe-004b787c.hcp.southeastasia.azmk8s.io

       

      Any idea where i can fix this wrong server information ?

       

       

      Author's profile photo Chan Jin Park
      Chan Jin Park

      Hello Bartosz Jarkowski ,

       

      please disregard my positing. I found out i did similar HXE installation on GCP and the old configuration file was there in cache. I remove the cache discovery folder and restart PC, i am o.ke. 🙂

      Author's profile photo Sid Krishna
      Sid Krishna

      Hello,

       

      Can the auto-scaling be done based on Memory parameters and the nodes gets added by self ?Something like Azure Automation script or do we have a direct option in azure portal?

       

      Thanks

      Sid