Scaling a WordPress application using Rook on Azure Kubernetes Service

WordPress is a popular free and open source content management system based on PHP and MySQL. One of the major benefits of using cloud computing is having the ability to better support your application by scaling based on user traffic. There are currently many ways to support WordPress scaling, including reverse proxy caching, persistent object caching, database distribution, using CDN, and more.

However, there are challenges with elastic (horizontal) scaling related to stateful configurations. A WordPress application is comprised of plugins, themes, content images, and many PHP files. All of these must be centrally shared for the same application and are not hosted on the MySQL DB tier. The quick tutorial below walks through how we can address this challenge, using Rook and AKS.

Azure Kubernetes Service (AKS)

Azure Kubernetes Service supports several storage options. All of them provide a central persistent storage layer with different performance capabilities. Using Azure Managed Disk provided adequate performance but by nature, a disk is bound to a single Pod, hence not supporting Pod scaling. Also, a virutal machine (a Kubernetes node, specifically) can support only up to a certain number of attached disks. Azure Files and Blob Fuse can serve multiple Kubernetes pods simultaneously. However, we opted for even better performance levels while leveraging Kubernetes’s extensive stack and plugins (in Kubernetes storage solutions).

Rook

Rook is an open source “incubating level” CNCF project that runs as a native Kubernetes service. Automating Ceph (open source software-defined storage), Rook offers storage for your Kubernetes applications through persistent volumes, providing optimal integration with applications in need of storage. Rook handles the heavy lifting behind the scenes, such as provisioning and management. We’ll use Rook as our persistent storage layer for the WordPress frontend.

Prerequisites

If you don’t already have an Azure account, create a free Azure account with Azure Kubernetes Service (AKS).

The Process

Installing Rook is quite easy and consists of four steps. The official installation guide can be found here.

Here’s a tailored guide for installing Rook on AKS:

1. Install FlexVolume for AKS:

kubectl apply -f https://raw.githubusercontent.com/Azure/kubernetes-keyvault-flexvol/master/deployment/kv-flexvol-installer.yaml

2. Install the Rook Operator:

kubectl apply -f https://raw.githubusercontent.com/shayshahak/Rook-for-AKS/master/operator.yaml

# verify the rook-ceph-operator, rook-ceph-agent, and rook-discover pods are in the `Running` state before proceeding (should take about two minutes):

kubectl -n rook-ceph-system get pod

3. Install the Rook Cluster:

kubectl apply -f https://raw.githubusercontent.com/shayshahak/Rook-for-AKS/master/cluster.yaml

# Use kubectl to list pods in the rook namespace. You should be able to see the following pods once they are all running:

kubectl -n rook-ceph get pod

NAME                                                   READY     STATUS      RESTARTS
rook-ceph-mgr-a-5977cffcf8-vsmqz                       1/1       Running     0
rook-ceph-mon-a-74dc9d4bcd-62g8s                       1/1       Running     0
rook-ceph-mon-b-558cf6f9d4-7sh24                       1/1       Running     0
rook-ceph-osd-0-585bf769cc-jp5sw                       1/1       Running     0
rook-ceph-osd-1-76576b697f-v9zwz                       1/1       Running     0
rook-ceph-osd-prepare-aks-agentpool-19295063-0-btkbb   0/2       Completed   0
rook-ceph-osd-prepare-aks-agentpool-19295063-1-2fsfz   0/2       Completed   0

4. Install the Shared File System:

kubectl apply -f https://raw.githubusercontent.com/shayshahak/Rook-for-AKS/master/filesystem.yaml

# To confirm the file system is configured, wait for the mds pods to start:

kubectl -n rook-ceph get pod -l app=rook-ceph-mds

# You should be able to see the following pods once they are all running:

NAME                                    READY  STATUS    AGE  Node
rook-ceph-mds-myfs-a-5b7cbd4977-6tgv2   1/1    Running   1m   aks-agentpool-33026748-0
rook-ceph-mds-myfs-b-85dc94778c-8qfkb   1/1    Running   1m   aks-agentpool-33026748-1

That’s it! The Rook cluster and the shared file system are ready. Let’s deploy a simple WordPress application using what we just created:

kubectl apply -f https://raw.githubusercontent.com/shayshahak/Rook-for-AKS/master/deploy.yaml

kubectl apply -f https://raw.githubusercontent.com/shayshahak/Rook-for-AKS/master/service.yaml

kubectl get svc wordpress-service

Next, connect to the WordPress application and configure its initial MySQL configuration. Notice that /var/www/html is hosted on the shared file system. Change the ‘replicas’ under the deployment YAML and verify that in fact all of your Kubernetes pods are using the same persistent storage.

Summary

By using Rook on AKS, we achieved:

• Scaling WordPress Frontend with a centralized persistent storage
• Gained HA for the Frontend without data being bound to a single Pod or Node
• Great performance in terms of end-to-end site response time
• Used a well known Kubernetes integrated plugin without using an additional external storage layer for the Kubernetes cluster

In this example we used Rook for a shared file system, but there are additional things you can achieve with Rook, including object and block storage for your containerized applications running on Kubernetes. Rook also provides helpful operations tools, such as its Ceph Dashboard and Monitoring.

Questions or feedback? Let us know in the comments.

Note: these configurations examples are for testing and learning purposes. Do not use these configurations examples for production. For production consider changing the cluster properties, filesystem properties as well as the namespaces in which your cluster and application resides.