Skip to content

Latest commit

 

History

History
168 lines (127 loc) · 4.73 KB

README.md

File metadata and controls

168 lines (127 loc) · 4.73 KB
Raw

Example of deploying Wordpress using MySQL

This example is based on the one from here updated to apply to the type of environment we use, and simplified and improved a little. It provides an example of:

  1. provisioning pods, services and routes
  2. provisioning persistent volumes and persistent volume claims using NFS

Openshift setup

Create a standard Openshift setup using oc cluster up or Ansible isntaller

Make sure you have provisioned NFS by adding a [nfs] section and adding the master to it. Inventories in our other examples should already have an NFS definition, as illustrated below:

[OSEv3:children]
...
nfs

[...]

[nfs]
MASTER_PRIVATE_FQDN

SSH to the master as the centos user.

NFS exports

Do these as root or using sudo.

mkdir -p /home/data/pv0001
mkdir -p /home/data/pv0002
chmod -R 777 /home/data/

Create /etc/exports.d/persitent-volumes.exports and make its contents as:

/home/data/pv0001 *(rw,root_squash)
/home/data/pv0002 *(rw,root_squash)

As root or using sudo, restart NFS to pick up the changes and check that the new exports are present.:

$ systemctl restart nfs-server
$ showmount -e localhost
Export list for localhost:
/home/data/pv0002       *
/home/data/pv0001       *
...

Clone repo

On the master (as a user that can create persistent volumes i.e. system:admin) clone this repo and move into the dir with these contents:

mkdir openrisknet
cd openrisknet/
git clone https://github.com/openrisknet/home
cd home/openshift/wordpress-mysql-example

Create Items

Project

Create a wordpress project and add our orn1 user to it as editor:

oc new-project wordpress
oc policy add-role-to-user edit orn1

Persistent Volumes (PV) and Claim (PVC) creation

If you are using a multi-node cluster rather than a single node or minishift you will need to modify the pv-NNNN.yaml files. Replace the nfs: server: value with the Private DNS value of the master. Something like ip-10-0-0-207.eu-west-1.compute.internal. Wtihout this change, pods started on the application node will not be able to connect to NFS, usually emitting the error requested NFS version or transport protocol is not supported

oc create -f pv-0001.yaml
oc create -f pv-0002.yaml
oc create -f pvc-mysql.yaml
oc create -f pvc-wp.yaml

Check that the claims have bound to the persistent volumes with the oc get pv command:

$ oc get pv
NAME      CAPACITY   ACCESSMODES   RECLAIMPOLICY   STATUS    CLAIM                   STORAGECLASS   REASON    AGE
pv0001    1Gi        RWO,RWX       Recycle         Bound     wordpress/claim-wp                               2m
pv0002    5Gi        RWO           Recycle         Bound     wordpress/claim-mysql                            2m

Pods

oc create -f pod-mysql.yaml
oc create -f pod-wp.yaml

Creating the pods may take a few moments as the Docker images need to be pulled from the appropriate registry. While you wait you can interrogate the pods using oc get po (oc get pods), waiting until the STATUS value becomes Running for both...

$ oc get po
NAME        READY     STATUS         RESTARTS   AGE
mysql       1/1       Running        0          11m
wordpress   1/1       Running        0          10m

Services

oc create -f svc-mysql.yaml
oc create -f svc-wp.yaml

Show that the Wordpress service is working by using oc get svc to find the wpfrontend and get its IP address and port.

$ oc get svc
NAME         CLUSTER-IP       EXTERNAL-IP                     PORT(S)          AGE
mysql        172.30.114.106   <none>                          3306/TCP         25s
wpfrontend   172.30.99.31     172.29.146.244,172.29.146.244   5055:31151/TCP   17s

Then check it is serving content with something like this: curl -L http://172.29.146.244:5055

Route

Edit the route-wp.yaml to define the correct hostname which will apply to the route.

oc create -f route-wp.yaml

Post install

Access Wordpress at something like: https://wordpress.example.com/

Look in the /home/data/pv0001 and /home/data/pv0002 directories to see the wordpress and mysql files.

Notes

  1. The route is not working correctly so Wordpress cannot be accessed. TODO - fix this.

Potential improvements

  1. A proper setup would use a replication controller to ensure the pods are present and to allow to scale them. TODO - include this.
  2. A proper setup would use templates for the pods, services and routes etc. to allow additional flexibility.