Signed-off-by: Mrunal Patel <mrunalp@gmail.com>
14 KiB
Persistent Installation of MySQL and WordPress on Kubernetes
This example describes how to run a persistent installation of WordPress and MySQL on Kubernetes. We'll use the mysql and wordpress official Docker images for this installation. (The WordPress image includes an Apache server).
Demonstrated Kubernetes Concepts:
- Persistent Volumes to define persistent disks (disk lifecycle not tied to the Pods).
- Services to enable Pods to locate one another.
- External Load Balancers to expose Services externally.
- Deployments to ensure Pods stay up and running.
- Secrets to store sensitive passwords.
Quickstart
Put your desired MySQL password in a file called password.txt
with
no trailing newline. The first tr
command will remove the newline if
your editor added one.
Note: if your cluster enforces selinux and you will be using Host Path for storage, then please follow this extra step.
tr --delete '\n' <password.txt >.strippedpassword.txt && mv .strippedpassword.txt password.txt
kubectl create -f https://raw.githubusercontent.com/kubernetes/kubernetes/master/examples/mysql-wordpress-pd/local-volumes.yaml
kubectl create secret generic mysql-pass --from-file=password.txt
kubectl create -f https://raw.githubusercontent.com/kubernetes/kubernetes/master/examples/mysql-wordpress-pd/mysql-deployment.yaml
kubectl create -f https://raw.githubusercontent.com/kubernetes/kubernetes/master/examples/mysql-wordpress-pd/wordpress-deployment.yaml
Table of Contents
Cluster Requirements
Kubernetes runs in a variety of environments and is inherently modular. Not all clusters are the same. These are the requirements for this example.
- Kubernetes version 1.2 is required due to using newer features, such
at PV Claims and Deployments. Run
kubectl version
to see your cluster version. - Cluster DNS will be used for service discovery.
- An external load balancer will be used to access WordPress.
- Persistent Volume Claims are used. You must create Persistent Volumes in your cluster to be claimed. This example demonstrates how to create two types of volumes, but any volume is sufficient.
Consult a Getting Started Guide to set up a cluster and the kubectl command-line client.
Decide where you will store your data
MySQL and WordPress will each use a Persistent Volume to store their data. We will use a Persistent Volume Claim to claim an available persistent volume. This example covers HostPath and GCEPersistentDisk volumes. Choose one of the two, or see Types of Persistent Volumes for more options.
Host Path
Host paths are volumes mapped to directories on the host. These should be used for testing or single-node clusters only. The data will not be moved between nodes if the pod is recreated on a new node. If the pod is deleted and recreated on a new node, data will be lost.
SELinux
On systems supporting selinux it is preferred to leave it enabled/enforcing.
However, docker containers mount the host path with the "svirt_sandbox_file_t"
label type, which is incompatible with the default label type for /tmp ("tmp_t"),
resulting in a permissions error when the mysql container attempts to chown
/var/lib/mysql.
Therefore, on selinx systems using host path, you should pre-create the host path
directory (/tmp/data/) and change it's selinux label type to "svirt_sandbox_file_t",
as follows:
## on every node:
mkdir -p /tmp/data
chmod a+rwt /tmp/data # match /tmp permissions
chcon -Rt svirt_sandbox_file_t /tmp/data
Continuing with host path, create the persistent volume objects in Kubernetes using local-volumes.yaml:
export KUBE_REPO=https://raw.githubusercontent.com/kubernetes/kubernetes/master
kubectl create -f $KUBE_REPO/examples/mysql-wordpress-pd/local-volumes.yaml
GCE Persistent Disk
This storage option is applicable if you are running on Google Compute Engine.
Create two persistent disks. You will need to create the disks in the
same GCE zone as the
Kubernetes cluster. The default setup script will create the cluster
in the us-central1-b
zone, as seen in the
config-default.sh file. Replace
<zone>
below with the appropriate zone. The names wordpress-1
and
wordpress-2
must match the pdName
fields we have specified in
gce-volumes.yaml.
gcloud compute disks create --size=20GB --zone=<zone> wordpress-1
gcloud compute disks create --size=20GB --zone=<zone> wordpress-2
Create the persistent volume objects in Kubernetes for those disks:
export KUBE_REPO=https://raw.githubusercontent.com/kubernetes/kubernetes/master
kubectl create -f $KUBE_REPO/examples/mysql-wordpress-pd/gce-volumes.yaml
Create the MySQL Password Secret
Use a Secret object
to store the MySQL password. First create a file (in the same directory
as the wordpress sample files) called
password.txt
and save your password in it. Make sure to not have a
trailing newline at the end of the password. The first tr
command
will remove the newline if your editor added one. Then, create the
Secret object.
tr --delete '\n' <password.txt >.strippedpassword.txt && mv .strippedpassword.txt password.txt
kubectl create secret generic mysql-pass --from-file=password.txt
This secret is referenced by the MySQL and WordPress pod configuration so that those pods will have access to it. The MySQL pod will set the database password, and the WordPress pod will use the password to access the database.
Deploy MySQL
Now that the persistent disks and secrets are defined, the Kubernetes pods can be launched. Start MySQL using mysql-deployment.yaml.
kubectl create -f $KUBE_REPO/examples/mysql-wordpress-pd/mysql-deployment.yaml
Take a look at mysql-deployment.yaml, and
note that we've defined a volume mount for /var/lib/mysql
, and then
created a Persistent Volume Claim that looks for a 20G volume. This
claim is satisfied by any volume that meets the requirements, in our
case one of the volumes we created above.
Also look at the env
section and see that we specified the password
by referencing the secret mysql-pass
that we created above. Secrets
can have multiple key:value pairs. Ours has only one key
password.txt
which was the name of the file we used to create the
secret. The MySQL image sets the
database password using the MYSQL_ROOT_PASSWORD
environment
variable.
It may take a short period before the new pod reaches the Running
state. List all pods to see the status of this new pod.
kubectl get pods
NAME READY STATUS RESTARTS AGE
wordpress-mysql-cqcf4-9q8lo 1/1 Running 0 1m
Kubernetes logs the stderr and stdout for each pod. Take a look at the
logs for a pod by using kubectl log
. Copy the pod name from the
get pods
command, and then:
kubectl logs <pod-name>
...
2016-02-19 16:58:05 1 [Note] InnoDB: 128 rollback segment(s) are active.
2016-02-19 16:58:05 1 [Note] InnoDB: Waiting for purge to start
2016-02-19 16:58:05 1 [Note] InnoDB: 5.6.29 started; log sequence number 1626007
2016-02-19 16:58:05 1 [Note] Server hostname (bind-address): '*'; port: 3306
2016-02-19 16:58:05 1 [Note] IPv6 is available.
2016-02-19 16:58:05 1 [Note] - '::' resolves to '::';
2016-02-19 16:58:05 1 [Note] Server socket created on IP: '::'.
2016-02-19 16:58:05 1 [Warning] 'proxies_priv' entry '@ root@wordpress-mysql-cqcf4-9q8lo' ignored in --skip-name-resolve mode.
2016-02-19 16:58:05 1 [Note] Event Scheduler: Loaded 0 events
2016-02-19 16:58:05 1 [Note] mysqld: ready for connections.
Version: '5.6.29' socket: '/var/run/mysqld/mysqld.sock' port: 3306 MySQL Community Server (GPL)
Also in mysql-deployment.yaml we created a
service to allow other pods to reach this mysql instance. The name is
wordpress-mysql
which resolves to the pod IP.
Up to this point one Deployment, one Pod, one PVC, one Service, one Endpoint, two PVs, and one Secret have been created, shown below:
kubectl get deployment,pod,svc,endpoints,pvc -l app=wordpress -o wide && \
kubectl get secret mysql-pass && \
kubectl get pv
NAME DESIRED CURRENT UP-TO-DATE AVAILABLE AGE
deploy/wordpress-mysql 1 1 1 1 3m
NAME READY STATUS RESTARTS AGE IP NODE
po/wordpress-mysql-3040864217-40soc 1/1 Running 0 3m 172.17.0.2 127.0.0.1
NAME CLUSTER-IP EXTERNAL-IP PORT(S) AGE SELECTOR
svc/wordpress-mysql None <none> 3306/TCP 3m app=wordpress,tier=mysql
NAME ENDPOINTS AGE
ep/wordpress-mysql 172.17.0.2:3306 3m
NAME STATUS VOLUME CAPACITY ACCESSMODES AGE
pvc/mysql-pv-claim Bound local-pv-2 20Gi RWO 3m
NAME TYPE DATA AGE
mysql-pass Opaque 1 3m
NAME CAPACITY ACCESSMODES STATUS CLAIM REASON AGE
local-pv-1 20Gi RWO Available 3m
local-pv-2 20Gi RWO Bound default/mysql-pv-claim 3m
Deploy WordPress
Next deploy WordPress using wordpress-deployment.yaml:
kubectl create -f $KUBE_REPO/examples/mysql-wordpress-pd/wordpress-deployment.yaml
Here we are using many of the same features, such as a volume claim for persistent storage and a secret for the password.
The WordPress image accepts the
database hostname through the environment variable
WORDPRESS_DB_HOST
. We set the env value to the name of the MySQL
service we created: wordpress-mysql
.
The WordPress service has the setting type: LoadBalancer
. This will
set up the wordpress service behind an external IP.
Find the external IP for your WordPress service. It may take a minute to have an external IP assigned to the service, depending on your cluster environment.
kubectl get services wordpress
NAME CLUSTER-IP EXTERNAL-IP PORT(S) AGE
wordpress 10.0.0.5 1.2.3.4 80/TCP 19h
Visit your new WordPress blog
Now, we can visit the running WordPress app. Use the external IP of the service that you obtained above.
http://<external-ip>
You should see the familiar WordPress init page.
Warning: Do not leave your WordPress installation on this page. If it is found by another user, they can set up a website on your instance and use it to serve potentially malicious content. You should either continue with the installation past the point at which you create your username and password, delete your instance, or set up a firewall to restrict access.
Take down and restart your blog
Set up your WordPress blog and play around with it a bit. Then, take down its pods and bring them back up again. Because you used persistent disks, your blog state will be preserved.
All of the resources are labeled with app=wordpress
, so you can
easily bring them down using a label selector:
kubectl delete deployment,service -l app=wordpress
kubectl delete secret mysql-pass
Later, re-creating the resources with the original commands will pick up the original disks with all your data intact. Because we did not delete the PV Claims, no other pods in the cluster could claim them after we deleted our pods. Keeping the PV Claims also ensured recreating the Pods did not cause the PD to switch Pods.
If you are ready to release your persistent volumes and the data on them, run:
kubectl delete pvc -l app=wordpress
And then delete the volume objects themselves:
kubectl delete pv local-pv-1 local-pv-2
or
kubectl delete pv wordpress-pv-1 wordpress-pv-2
Next Steps
- Introspection and Debugging
- Jobs may be useful to run SQL queries.
- Exec
- Port Forwarding