As more developers work within distributed environments, tools like Kubernetes have become central to keeping application components standardized across dynamic build and production environments. With the increasing complexity of application ecosystems and the growing popularity of Kuberbetes, tools that help manage resources within Kubernetes clusters have become essential.
Helm is an open-source package manager for Kubernetes that simplifies the process of deploying and upgrading applications on a Kubernetes cluster, while also providing a way to find and share ready-to-install applications that are packaged as Kubernetes Charts.
In this tutorial, we’ll use Helm for setting up WordPress on top of a Kubernetes cluster, in order to create a highly-available website. In addition to leveraging the intrinsic scalability and high availability aspects of Kubernetes, this setup will help keeping WordPress secure by providing simplified upgrade and rollback workflows via Helm.
We’ll be using an external MySQL server in order to abstract the database component, since it can be part of a separate cluster or managed service for extended availability. After completing the steps described in this tutorial, you will have a fully functional WordPress installation within a containerized cluster environment managed by Kubernetes.
If you’re looking for a managed Kubernetes hosting service, check out our simple, managed Kubernetes service built for growth.
In order to complete this guide, you will need the following available to you:
kubectl
command-line tool installed on your local machine or development server, configured to connect to your cluster. Please see the official Kubernetes documentation for instructions on how to set this up.Before moving on, make sure you’re able to log into your MySQL server, and that you have connectivity to your Kubernetes cluster. In case you have multiple clusters set up in your kubectl
config file, you should make sure that you’re connected to the correct cluster by running the following command from your local machine or development server:
- kubectl config get-contexts
This is an example output:
Output
CURRENT NAME CLUSTER AUTHINFO NAMESPACE
* do-sfo2-wordpress-cluster do-sfo2-wordpress-cluster do-sfo2-wordpress-cluster-admin
minikube minikube minikube
The asterisk sign (*) indicates which cluster is currently the default context. In case you need to change the current context, run:
- kubectl config use-context context-name
You should now be ready to follow the rest of the guide.
First, we’ll create a dedicated MySQL user and a database for WordPress, allowing connections from external hosts. This is necessary because our WordPress installation will live on a separate server inside the Kubernetes cluster. In case you already have a dedicated MySQL user and database set up for WordPress, you can skip to the next step.
From the MySQL server, log into MySQL with the following command:
- mysql -u root -p
You will be prompted to provide the password you set up for the root MySQL account when you first installed the software. After logging in, MySQL will give you a command prompt you can use to create the database and user we need for WordPress.
Note: For this tutorial, we’ll be creating a database named wordpress
and a user named wordpress_user
, identified by the password password
. Please note that these are insecure example values, and you should modify them accordingly throughout this guide.
To create the database, you can use the following statement:
- CREATE DATABASE wordpress;
Now, let’s create a dedicated MySQL user for this database:
- CREATE USER wordpress_user IDENTIFIED BY 'password';
The user wordpress_user
was created, but it doesn’t have any access permissions yet. The following command will give this user admin access (all privileges) to the wordpress database from both local and external networks:
- GRANT ALL PRIVILEGES ON wordpress.* TO wordpress_user@'%';
To update the internal MySQL tables that manage access permissions, use the following statement:
- FLUSH PRIVILEGES;
Now you can exit the MySQL client with:
- exit;
To test that the changes were successful, you can log into the MySQL command-line client again, this time using the new account wordpress_user
to authenticate:
- mysql -u wordpress_user -p
You should use the same password you provided when creating this MySQL user with the CREATE_USER
statement. To confirm your new user has access to the wordpress
database, you can use the following statement:
- show databases;
The following output is expected:
Output+--------------------+
| Database |
+--------------------+
| information_schema |
| wordpress |
+--------------------+
2 rows in set (0.03 sec)
After confirming the wordpress
database is included in the results, you can exit the MySQL command-line client with:
- exit;
You now have a dedicated MySQL database for WordPress, and valid access credentials to use within it. Because our WordPress installation will live on a separate server, we still need to edit our MySQL configuration to allow connections coming from external hosts.
While still on your MySQL server, open the file /etc/mysql/mysql.conf.d/mysqld.cnf
using your command-line editor of choice:
- sudo nano /etc/mysql/mysql.conf.d/mysqld.cnf
Locate the bind-address
setting within this file. By default, MySQL listens only on 127.0.0.1
(localhost). In order to accept connections from external hosts, we need to change this value to 0.0.0.0
. This is how your bind-address
configuration should look:
# Instead of skip-networking the default is now to listen only on
# localhost which is more compatible and is not less secure.
bind-address = 0.0.0.0
When you’re done making these changes, save and close the file. You’ll need to restart MySQL with the following command:
- sudo systemctl restart mysql
To test if you’re able to connect remotely, run the following command from your local machine or development server:
- mysql -h mysql_server_ip -u wordpress_user -p
Remember to change mysql_server_ip
to your MySQL server IP address or hostname. If you’re able to connect without errors, you are now ready to proceed to the next step.
Now that we have the necessary information to connect to the MySQL database, we can go ahead and install WordPress using Helm.
By default, the WordPress chart installs MariaDB on a separate pod inside the cluster and uses it as the WordPress database. We want to disable this behavior and configure WordPress to use an external MySQL database. This and other configuration options (such as the default WordPress admin user and password) can be set at installation time, either via command-line parameters or via a separate YAML configuration file.
In order to keep things organized and easily extendable, we are going to use a configuration file.
From your local machine or development server, create a new directory for your project settings and navigate into it:
- mkdir myblog-settings
- cd myblog-settings
Next, create a file named values.yaml
, using your text editor of choice:
- nano values.yaml
Within this file, we need to set up a few variables that will define how WordPress connects to the database, as well as some basic information about your site and the initial admin user for logging into WordPress when the installation is complete.
We’ll base our configuration on the default values.yaml
file from the WordPress Helm chart. The Blog/Site Info section contains general options for your WordPress blog, such as the name of the blog and the initial user credentials. The Database Settings section of this file contains the settings for connecting to the remote MySQL server. MariaDB is disabled in the final section.
Copy the following contents into your values.yaml
file, replacing the highlighted values with your custom values:
## Blog/Site Info
wordpressUsername: sammy
wordpressPassword: password
wordpressEmail: sammy@example.com
wordpressFirstName: Sammy
wordpressLastName: the Shark
wordpressBlogName: Sammy's Blog!
## Database Settings
externalDatabase:
host: mysql_server_ip
user: wordpress_user
password: password
database: wordpress
## Disabling MariaDB
mariadb:
enabled: false
We have just configured the following options:
When you’re done editing, save the file and exit the editor.
Now that we have all settings in place, it is time to execute helm
to install WordPress. The following command tells helm
to install the most recent stable release of the WordPress chart under the name myblog
, using values.yaml
as configuration file:
- helm install --name myblog -f values.yaml stable/wordpress
You should get output similar to the following:
Output
NAME: myblog
LAST DEPLOYED: Fri Jan 25 20:24:10 2019
NAMESPACE: default
STATUS: DEPLOYED
RESOURCES:
==> v1/Deployment
NAME READY UP-TO-DATE AVAILABLE AGE
myblog-wordpress 0/1 1 0 1s
==> v1/PersistentVolumeClaim
NAME STATUS VOLUME CAPACITY ACCESS MODES STORAGECLASS AGE
myblog-wordpress Pending do-block-storage 1s
==> v1/Pod(related)
NAME READY STATUS RESTARTS AGE
myblog-wordpress-5965f49485-8zfl7 0/1 Pending 0 1s
==> v1/Secret
NAME TYPE DATA AGE
myblog-externaldb Opaque 1 1s
myblog-wordpress Opaque 1 1s
==> v1/Service
NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGE
myblog-wordpress LoadBalancer 10.245.144.79 <pending> 80:31403/TCP,443:30879/TCP 1s
(...)
After the installation is finished, a service named myblog-wordpress is created within your Kubernetes cluster, but it may take a few minutes before the container is ready and the External-IP
information is available. To check the status of this service and retrieve its external IP address, run:
- kubectl get services
You should get output similar to the following:
Output
NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGE
kubernetes ClusterIP 10.245.0.1 <none> 443/TCP 20h
myblog-wordpress LoadBalancer 10.245.144.79 203.0.113.110 80:31403/TCP,443:30879/TCP 3m40s
This command gives you detailed information about services running on your cluster, including name and type of the service, as well as IP addresses used by these services. As you can see from the output, the WordPress installation is being served as myblog-wordpress
on the external IP address 203.0.113.110
.
]Note: In case you are using minikube
to test this setup, you’ll need to run minikube service myblog-wordpress
in order to expose the container web server so that you can access it from your browser.
Your WordPress installation is now operational. To access the admin interface, use the public IP address obtained from the output of kubectl get services
, followed by /wp-admin
in your web browser:
http://203.0.113.110/wp-admin
You should use the credentials defined in your values.yaml
file to log in and start configuring your WordPress site.
Because of its popularity, WordPress is often a target for malicious exploitation, so it’s important to keep it updated. We can upgrade Helm releases with the command helm upgrade
.
To list all of your current releases, run the following command from your local machine or development server:
- helm list
You should get output similar to this:
OutputNAME REVISION UPDATED STATUS CHART APP VERSION NAMESPACE
myblog 1 Fri Jan 25 20:24:10 2019 DEPLOYED wordpress-5.1.2 5.0.3 default
As you can see from the output, our current WordPress version is 5.0.3
(app version), while the chart version is 5.1.2
. If you want to upgrade a release to a newer version of a chart, first update your Helm repositories with:
- helm repo update
You can expect the following output:
OutputHang tight while we grab the latest from your chart repositories...
...Skip local chart repository
...Successfully got an update from the "stable" chart repository
Update Complete. ⎈ Happy Helming!⎈
Now you can check if there’s a newer version of the WordPress chart available with:
- helm inspect chart stable/wordpress
You should see output similar to this:
OutputapiVersion: v1
appVersion: 5.1.1
description: Web publishing platform for building blogs and websites.
engine: gotpl
home: http://www.wordpress.com/
icon: https://bitnami.com/assets/stacks/wordpress/img/wordpress-stack-220x234.png
keywords:
- wordpress
- cms
- blog
- http
- web
- application
- php
maintainers:
- email: containers@bitnami.com
name: Bitnami
name: wordpress
sources:
- https://github.com/bitnami/bitnami-docker-wordpress
version: 5.9.0
As you can see from the output, there’s a new chart available (version 5.9.0) with WordPress 5.1.1 (app version). Whenever you want to upgrade your WordPress release to the latest WordPress chart, you should run:
- helm upgrade -f values.yaml myblog stable/wordpress
This command will produce output very similar to the output produced by helm install
. It is important to provide the same configuration file we used when installing the WordPress chart for the first time, as it contains the custom database settings we defined for our setup.
Now, if you run helm list
again, you should see updated information about your release:
Output
NAME REVISION UPDATED STATUS CHART APP VERSION NAMESPACE
myblog 2 Fri May 3 14:51:20 2019 DEPLOYED wordpress-5.9.0 5.1.1 default
You have successfully upgraded your WordPress to the latest version of the WordPress chart.
Each time you upgrade a release, a new revision of that release is created by Helm. A revision sets a fixed checkpoint to where you can come back if things don’t work as expected. It is similar to a commit in Git, because it creates a history of changes that can be compared and reverted. If something goes wrong during the upgrade process, you can always rollback to a previous revision of a given Helm release with the helm rollback
command:
- helm rollback release-name revision-number
For instance, if we want to undo the upgrade and rollback our WordPress release to its first version, we would use:
- helm rollback myblog 1
This would rollback the WordPress installation to its first release. You should see the following output, indicating that the rollback was successful:
Output
Rollback was a success! Happy Helming!
Running helm list
again should now indicate that WordPress was downgraded back to 5.0.3, chart version 5.1.2:
Output
NAME REVISION UPDATED STATUS CHART APP VERSION NAMESPACE
myblog 3 Mon Jan 28 22:02:42 2019 DEPLOYED wordpress-5.1.2 5.0.3 default
Notice that rolling back a release will actually create a new revision, based on the target revision of the roll-back. Our WordPress release named myblog
now is at revision number three, which was based on revision number one.
In this guide, we installed WordPress with an external MySQL server on a Kubernetes cluster using the command-line tool Helm. We also learned how to upgrade a WordPress release to a new chart version, and how to rollback a release if something goes wrong throughout the upgrade process.
As additional steps, you might consider setting up Nginx Ingress with Cert-Manager in order to enable name-based virtual hosting and to configure an SSL certificate for your WordPress site. You should also check the recommended production settings for the WordPress chart we used in this guide.
If you want to learn more about Kubernetes and Helm, please check out the Kubernetes section of our community page.
Thanks for learning with the DigitalOcean Community. Check out our offerings for compute, storage, networking, and managed databases.
This textbox defaults to using Markdown to format your answer.
You can type !ref in this text area to quickly search our full set of tutorials, documentation & marketplace offerings and insert the link!
@erikaheidi thanks for the great article.
I’m heading to K8s, after following steps seem mysql server is not on a part of K8s and lost of high availability function?
Thanks for the tutorial. It seems a little outdated:
helm install --name myblog
should behelm install myblog
.stable/wordpress
image is deprecated in favor ofbitnami/wordpress
.Another problem with this tutorial (tested on DigitalOcean):
After deployment, visiting EXTERNAL-IP of the LoadBalancer service does not work. Doing
nmap -sT
on the address, I get both port 80 and port 443, buttelnet ... 80
immediately closes the connection.Presumably this means that the basic configuration is working, but that the tutorial does not cover any changes made to the image since its inception.
It would be nice for an updated tutorial, since otherwise I’m not sure what the immediate value is. I was looking for something to get me started with Helm, and this appeared very minimalist, except it doesn’t take you to the end.
Thanks for the tutorial!
When trying to install WordPress with: “helm install --name myblog -f values.yaml stable/wordpress” an error shows up:
Error: unknown flag: --name
Updating flag to “-name” other errors show up:
WARNING: This chart is deprecated
Error: create: failed to create: namespaces “ame” not found
Thanks for any help!
Thanks for the tutorial. I’m a newbie to k8s and was following this tutorial to get my new WP blog up. After I finished the installation, I got the External IP as “localhost”.
I can access the wp-admin as localhost/admin but the credentials used in values.yaml doesn’t work. I can see the wp tables got created in my Digital Ocean managed Database but the wp_users table was empty. I inserted an admin user with admin capabilities and tried using that credentials but still it doesn’t work.
Later I realized that there is a Maria DB pod running in my local but no clue on how to access it. Any help to know this better would be appreciated.
Thanks in advance!
For those reading this using the clusters, unfortunately,
stable/wordpress
does NOT support MySQL 8.x. Thestable/wordpress
helm chart uses the bitnami docker wordpress image – https://github.com/bitnami/bitnami-docker-wordpress/issues/214I’m wondering the same thing as @markojak. What is the benefit of setting up an external MySQL server, is it cheaper?
I would think that running MariaDB container as a pod would be preferable so you can upgrade/manage it more easily.
Thanks for this tutorial! Getting started with Kubernetes on DO now and it’s pretty cool.
Two questions:
Why are you using MySQL instead of the existing MariaDB that comes with the Bitnami Helm Chart release?
What is the thinking about using a Baremetal instance to run the Database (besides your brief mention of extended availability) - My understanding is that Kubernetes allows for scalability and running the Database on another instance seems to defeat this.
Thanks for great tutorial!