Use Azure Premium SSD v2 disks on Azure Kubernetes Service

Azure Premium SSD v2 disks offer IO-intense enterprise workloads, a consistent submillisecond disk latency, and high IOPS and throughput. The performance (capacity, throughput, and IOPS) of Premium SSD v2 disks can be independently configured at any time, making it easier for more scenarios to be cost efficient while meeting performance needs.

This article describes how to configure a new or existing AKS cluster to use Azure Premium SSD v2 disks.

Before you begin

Before creating or upgrading an AKS cluster that is able to use Azure Premium SSD v2 disks, you need to create an AKS cluster in the same region and availability zone that supports Premium Storage and attach the disks following the steps below.

For an existing AKS cluster, you can enable Premium SSD v2 disks by adding a new node pool to your cluster, and then attach the disks following the steps below.

Limitations

• Azure Premium SSD v2 disks have certain limitations that you need to be aware of. For a complete list, see Premium SSD v2 limitations.

Use Premium SSD v2 disks dynamically with a storage class

To use Premium SSD v2 disks in a deployment or stateful set, you can use a storage class for dynamic provisioning.

Create the storage class

A storage class is used to define how a unit of storage is dynamically created with a persistent volume. For more information on Kubernetes storage classes, see Kubernetes Storage Classes.

In this example, you create a storage class that references Premium SSD v2 disks. Create a file named azure-pv2-disk-sc.yaml, and copy in the following manifest.

apiVersion: storage.k8s.io/v1
kind: StorageClass
metadata:
 name: premium2-disk-sc
parameters:
 cachingMode: None
 skuName: PremiumV2_LRS
 DiskIOPSReadWrite: "4000"
 DiskMBpsReadWrite: "1000"
provisioner: disk.csi.azure.com
reclaimPolicy: Delete
volumeBindingMode: Immediate
allowVolumeExpansion: true

Create the storage class with the kubectl apply command and specify your azure-pv2-disk-sc.yaml file:

kubectl apply -f azure-pv2-disk-sc.yaml

The output from the command resembles the following example:

storageclass.storage.k8s.io/premium2-disk-sc created

Create a persistent volume claim

A persistent volume claim (PVC) is used to automatically provision storage based on a storage class. In this case, a PVC can use the previously created storage class to create an Ultra Disk.

Create a file named azure-pv2-disk-pvc.yaml, and copy in the following manifest. The claim requests a disk named premium2-disk that is 1000 GB in size with ReadWriteOnce access. The premium2-disk-sc storage class is specified as the storage class.

apiVersion: v1
kind: PersistentVolumeClaim
metadata:
 name: premium2-disk
spec:
 accessModes:
 - ReadWriteOnce
 storageClassName: premium2-disk-sc
 resources:
 requests:
 storage: 1000Gi

Create the persistent volume claim with the kubectl apply command and specify your azure-pv2-disk-pvc.yaml file:

kubectl apply -f azure-pv2-disk-pvc.yaml

The output from the command resembles the following example:

persistentvolumeclaim/premium2-disk created

Use the persistent volume

Once the persistent volume claim has been created and the disk successfully provisioned, a pod can be created with access to the disk. The following manifest creates a basic NGINX pod that uses the persistent volume claim named premium2-disk to mount the Azure disk at the path /mnt/azure.

Create a file named nginx-premium2.yaml, and copy in the following manifest.

kind: Pod
apiVersion: v1
metadata:
 name: nginx-premium2
spec:
 containers:
 - name: nginx-premium2
 image: mcr.microsoft.com/oss/nginx/nginx:1.15.5-alpine
 resources:
 requests:
 cpu: 100m
 memory: 128Mi
 limits:
 cpu: 250m
 memory: 256Mi
 volumeMounts:
 - mountPath: "/mnt/azure"
 name: volume
 volumes:
 - name: volume
 persistentVolumeClaim:
 claimName: premium2-disk

Create the pod with the kubectl apply command, as shown in the following example:

kubectl apply -f nginx-premium2.yaml

The output from the command resembles the following example:

pod/nginx-premium2 created

You now have a running pod with your Azure disk mounted in the /mnt/azure directory. This configuration can be seen when inspecting your pod via kubectl describe pod nginx-premium2, as shown in the following condensed example:

kubectl describe pod nginx-premium2

[...]
Volumes:
 volume:
 Type: PersistentVolumeClaim (a reference to a PersistentVolumeClaim in the same namespace)
 ClaimName: premium2-disk
 ReadOnly: false
 kube-api-access-sh59b:
 Type: Projected (a volume that contains injected data from multiple sources)
 TokenExpirationSeconds: 3607
 ConfigMapName: kube-root-ca.crt
 ConfigMapOptional: <nil>
 DownwardAPI: true
QoS Class: Burstable
Node-Selectors: <none>
Tolerations: node.kubernetes.io/memory-pressure:NoSchedule op=Exists
 node.kubernetes.io/not-ready:NoExecute op=Exists for 300s
 node.kubernetes.io/unreachable:NoExecute op=Exists for 300s
Events:
 Type Reason Age From Message
 ---- ------ ---- ---- -------
 Normal Scheduled 7m58s default-scheduler Successfully assigned default/nginx-premium2 to aks-agentpool-12254644-vmss000006
 Normal SuccessfulAttachVolume 7m46s attachdetach-controller AttachVolume.Attach succeeded for volume "pvc-ff39fb64-1189-4c52-9a24-e065b855b886"
 Normal Pulling 7m39s kubelet Pulling image "mcr.microsoft.com/oss/nginx/nginx:1.15.5-alpine"
 Normal Pulled 7m38s kubelet Successfully pulled image "mcr.microsoft.com/oss/nginx/nginx:1.15.5-alpine" in 1.192915667s
 Normal Created 7m38s kubelet Created container nginx-premium2
 Normal Started 7m38s kubelet Started container nginx-premium2
[...]

Set IOPS and throughput limits

Input/Output Operations Per Second (IOPS) and throughput limits for Azure Premium SSD v2 are currently not supported through AKS. To adjust performance, you can use the Azure CLI command az disk update and including the --disk-iops-read-write and --disk-mbps-read-write parameters.

The following example updates the disk IOPS read/write to 5000 and Mbps to 200. For --resource-group, the value must be the second resource group automatically created to store the AKS worker nodes with the naming convention MC_resourcegroupname_clustername_location. For more information, see Why are two resource groups created with AKS?.

The value for the --name parameter is the name of the volume created using the StorageClass, and it starts with pvc-. To identify the disk name, you can run kubectl get pvc or navigate to the secondary resource group in the portal to find it. See manage resources from the Azure portal to learn more.

az disk update --subscription subscriptionName --resource-group myResourceGroup --name diskName --disk-iops-read-write=5000 --disk-mbps-read-write=200 

Next steps

• For more about Premium SSD v2 disks, see Using Azure Premium SSD v2 disks.
• For more about storage best practices, see Best practices for storage and backups in Azure Kubernetes Service (AKS).