Deploy WordPress on GKE with Persistent Disk and Cloud SQL

This tutorial shows you how to set up a single-replica WordPress deployment on Google Kubernetes Engine (GKE) using a MySQL database. Instead of installing MySQL, you use Cloud SQL, which provides a managed version of MySQL. WordPress uses PersistentVolumes (PV) and PersistentVolumeClaims (PVC) to store data.

A PV is a representation of storage volume in the cluster that is provisioned by an admin, or dynamically provisioned by Kubernetes, to fulfill a request made in a PVC. A PVC is a request for storage of a certain storage class by a user that can be fulfilled by a PV. PVs and PVCs are independent from Pod lifecycles and preserve data through restarting, rescheduling, and even deleting Pods. WordPress uses Persistent Disk as storage to back the PVs.

Background

WordPress is a blogging tool that uses a relational database to store the blog articles and their related objects and metadata, and the local file system to store assets, such as pictures in a blog post. This tutorial uses the official WordPress Docker image from Docker Hub.

In general, a container's root file system isn't suitable to store persistent data. The containers you run on GKE are typically disposable entities, and the cluster manager can delete, evict, or reschedule any containers that become unavailable due to node failures or other causes. When nodes fail, all data saved to a container's root file system is lost.

Using PVs backed by Persistent Disk let you store your WordPress platform data outside the containers. This way, even if the containers are deleted, their data persists. With the default storage class, your Persistent Disk (and hence your data) doesn't move with your Pod if the Pod is rescheduled to another node. There are different ways to handle moving the data, but that's outside the scope of this tutorial. For more information, see Persistent volumes with Persistent Disk.

WordPress requires a PV to store data. For this tutorial, you use the default storage class to dynamically create a Persistent Disk and create a PVC for the deployment.

Setting up your environment

1. In Cloud Shell, set the default region for the Google Cloud CLI:

   gcloud config set compute/region region
   

   Replace the following:

   • region: Choose a region that's closest to you. For more information, see Regions and Zones.

2. Set the PROJECT_ID environment variable to your Google Cloud project ID (project-id).

   export PROJECT_ID=project-id
   

3. Download the app manifest files from the GitHub repository:

   git clone https://github.com/GoogleCloudPlatform/kubernetes-engine-samples
   

4. Change to the directory with the wordpress-persistent-disks file:

   cd kubernetes-engine-samples/quickstarts/wordpress-persistent-disks
   

5. Set the WORKING_DIR environment variable:

   WORKING_DIR=$(pwd)
   

   For this tutorial, you create Kubernetes objects using manifest files in YAML format.

Creating a GKE cluster

You create a GKE cluster to host your WordPress app container.

• In Cloud Shell, create an GKE cluster named persistent-disk-tutorial:

  CLUSTER_NAME=persistent-disk-tutorial
  gcloud container clusters create-auto $CLUSTER_NAME
  

  Once created, connect to your new cluster:

  gcloud container clusters get-credentials $CLUSTER_NAME --location REGION
  

Creating a PV and a PVC backed by Persistent Disk

Create a PVC as the storage required for WordPress. GKE has a default StorageClass resource installed that lets you dynamically provision PVs backed by Persistent Disk. You use the wordpress-volumeclaim.yaml file to create the PVCs required for the deployment.

This manifest file describes a PVC that requests 200 GB of storage. A StorageClass resource hasn't been defined in the file, so this PVC uses the default StorageClass resource to provision a PV backed by Persistent Disk.

1. In Cloud Shell, deploy the manifest file:

   kubectl apply -f $WORKING_DIR/wordpress-volumeclaim.yaml
   

   It can take up to ten seconds to provision the PV backed by Persistent Disk and to bind it to your PVC. You can check the status with the following command:

   kubectl get persistentvolumeclaim
   

   The output shows a PersistentVolumeClaim with a status of Pending, similar to the following:

   NAME                    STATUS    VOLUME   CAPACITY   ACCESS MODES   STORAGECLASS   AGE
   wordpress-volumeclaim   Pending                                      standard-rwo   5s
   

   This PersistentVolumeClaim remains in the Pending state until you use it later in this tutorial.

Creating a Cloud SQL for MySQL instance

1. In Cloud Shell, create an instance named mysql-wordpress-instance:

   INSTANCE_NAME=mysql-wordpress-instance
   gcloud sql instances create $INSTANCE_NAME
   

2. Add the instance connection name as an environment variable:

   export INSTANCE_CONNECTION_NAME=$(gcloud sql instances describe $INSTANCE_NAME \
       --format='value(connectionName)')
   

3. Create a database for WordPress to store its data:

   gcloud sql databases create wordpress --instance $INSTANCE_NAME
   

4. Create a database user called wordpress and a password for WordPress to authenticate to the instance:

   CLOUD_SQL_PASSWORD=$(openssl rand -base64 18)
   gcloud sql users create wordpress --host=% --instance $INSTANCE_NAME \
       --password $CLOUD_SQL_PASSWORD
   

   If you close your Cloud Shell session, you lose the password. Make a note of the password because you need it later in the tutorial.

You have completed setting up the database for your new WordPress blog.

Deploying WordPress

Before you can deploy WordPress, you must create a service account. You create a Kubernetes secret to hold the service account credentials and another secret to hold the database credentials.

Configure a service account and create secrets

1. To let your WordPress app access the MySQL instance through a Cloud SQL proxy, create a service account:

   SA_NAME=cloudsql-proxy
   gcloud iam service-accounts create $SA_NAME --display-name $SA_NAME
   

2. Add the service account email address as an environment variable:

   SA_EMAIL=$(gcloud iam service-accounts list \
       --filter=displayName:$SA_NAME \
       --format='value(email)')
   

3. Add the cloudsql.client role to your service account:

   gcloud projects add-iam-policy-binding $PROJECT_ID \
       --role roles/cloudsql.client \
       --member serviceAccount:$SA_EMAIL
   

4. Create a key for the service account:

   gcloud iam service-accounts keys create $WORKING_DIR/key.json \
       --iam-account $SA_EMAIL
   

   This command downloads a copy of the key.json file.

5. Create a Kubernetes secret for the MySQL credentials:

   kubectl create secret generic cloudsql-db-credentials \
       --from-literal username=wordpress \
       --from-literal password=$CLOUD_SQL_PASSWORD
   

6. Create a Kubernetes secret for the service account credentials:

   kubectl create secret generic cloudsql-instance-credentials \
       --from-file $WORKING_DIR/key.json
   

Deploy WordPress

The next step is to deploy your WordPress container in the GKE cluster.

The wordpress_cloudsql.yaml manifest file describes a Deployment that creates a single Pod running a container with a WordPress instance. This container reads the WORDPRESS_DB_PASSWORD environment variable that contains the cloudsql-db-credentials secret you created.

This manifest file also configures the WordPress container to communicate with MySQL through the Cloud SQL proxy running in the sidecar container. The host address value is set on the WORDPRESS_DB_HOST environment variable.

1. Prepare the file by replacing the INSTANCE_CONNECTION_NAME environment variable:

   cat $WORKING_DIR/wordpress_cloudsql.yaml.template | envsubst > \
       $WORKING_DIR/wordpress_cloudsql.yaml
   

2. Deploy the wordpress_cloudsql.yaml manifest file:

   kubectl create -f $WORKING_DIR/wordpress_cloudsql.yaml
   

   It takes a few minutes to deploy this manifest file while a Persistent Disk is attached to the compute node.

3. Watch the deployment to see the status change to running:

   kubectl get pod -l app=wordpress --watch
   

   When the output shows a status of Running, you can move on to the next step.

   NAME                     READY     STATUS    RESTARTS   AGE
   wordpress-387015-02xxb   2/2       Running   0          2m47s
   

Expose the WordPress service

In the previous step, you deployed a WordPress container, but it's currently not accessible from outside your cluster because it doesn't have an external IP address. You can expose your WordPress app to traffic from the internet by creating and configuring a Kubernetes Service with an attached external load balancer. To learn more about exposing apps using Services in GKE, see the how-to guide.

1. Create a Service of type:LoadBalancer:

   kubectl create -f $WORKING_DIR/wordpress-service.yaml
   

   It takes a few minutes to create a load balancer.

2. Watch the deployment and wait for the service to have an external IP address assigned:

   kubectl get svc -l app=wordpress --watch
   

3. When the output shows an external IP address, you can proceed to the next step. Note that your external IP is different from the following example.

   NAME        CLUSTER-IP      EXTERNAL-IP    PORT(S)        AGE
   wordpress   10.51.243.233   203.0.113.3    80:32418/TCP   1m
   

4. Make a note of the EXTERNAL_IP address field to use later.

Warning: Don't leave your WordPress installation with a load balancer that exposes the cluster on the installation page. Other visitors can set up a website on your cluster and use it to serve potentially malicious content. Either continue setting up your WordPress blog or clean up this deployment.

Setting up your WordPress blog

In this section, you set up your WordPress blog.

1. In your browser, go to the following URL, replacing external-ip-address with the EXTERNAL_IP address of the service that exposes your WordPress instance:

   http://external-ip-address
   

2. On the WordPress installation page, select a language, and then click Continue.

3. Complete the Information needed page, and then click Install WordPress.

4. Click Log In.

5. Enter the username and password that you previously created.

6. You now have a blog site. To visit your blog, in your browser, go to the following URL:

   http://external-ip-address