ipedrazas / mysql-operator

MySQL operator creates/configures/manages MySQL databases in containers, in AWS, GKE and Azure.

MySQL operator

MySQL operator creates/configures/manages MySQL databases in containers, in AWS, GKE and Azure.

How does it work?

First you run the mysql-operator which consists in 2 containers:

• KWatch: that watches the API Server for changes in the MySQL thirdpartyobject
• Klunch: that interfaces with the cloud provider specified in the 3rd party object.

First, you create a MySQL 3rd party object (`mysql-3rdpart.yaml):

    apiVersion: extensions/v1beta1
    kind: ThirdPartyResource
    description:  "A specification of a MySQL Database."
    metadata:
      name: "mysql.infra.k8s.uk"
      labels:
          resource: database
          object: mysql
    versions:
      - name: v1

Once we have the third party object created we can use the API to create objects of that type. If we list the available APIs we can see that there's a new entry

    curl -k -l -E ~/.minikube/certificate.pfx:test https://192.168.99.100:8443
    {
      "paths": [
        "/api",
        "/api/v1",
        "/apis",
        "/apis/apps",
        "/apis/apps/v1alpha1",
        "/apis/authentication.k8s.io",
        "/apis/authentication.k8s.io/v1beta1",
        "/apis/authorization.k8s.io",
        "/apis/authorization.k8s.io/v1beta1",
        "/apis/autoscaling",
        "/apis/autoscaling/v1",
        "/apis/batch",
        "/apis/batch/v1",
        "/apis/batch/v2alpha1",
        "/apis/certificates.k8s.io",
        "/apis/certificates.k8s.io/v1alpha1",
        "/apis/extensions",
        "/apis/extensions/v1beta1",
        "/apis/infra.k8s.uk",
        "/apis/infra.k8s.uk/v1",
        "/apis/infra.sohohouse.com",
        "/apis/infra.sohohouse.com/v1",
        "/apis/policy",
        "/apis/policy/v1alpha1",
        "/apis/rbac.authorization.k8s.io",
        "/apis/rbac.authorization.k8s.io/v1alpha1",
        "/apis/storage.k8s.io",
        "/apis/storage.k8s.io/v1beta1",
        "/healthz",
        "/healthz/ping",
        "/logs",
        "/metrics",
        "/swaggerapi/",
        "/ui/",
        "/version"
      ]
    }

we can see that we have 2 new APIs:

        /apis/infra.k8s.uk
        /apis/infra.k8s.uk/v1

Now, if we create a Mysql object we will be able to query the API. There are a few examples in the kubernetes folder. This is the simpler case:

    apiVersion: "infra.k8s.uk/v1"
    kind: "Mysql"
    description: "Object to define a database"
    metadata:
      name: "wordpress"
    spec:
      name: "wordpress"
      provider: "kubernetes"
      username: "wpsecretadmin"
      password: "S3cr3t_P4ssw0rd"
      service: "mywordpress"

Note that the apiVersion is now infra.k8s.uk/v1 that matches our 3rd party object name and version attribute.

What happens once I create the 3rd party object?

Nothing, once you create the 3rd party object what you're doing behind the scenes is to define the new API endpoints. Nothing happens until you create a Mysql object.

How do I create a Mysql object?

The same way that you create all the other kubernetes resources, using the API and sending the manifest with your desired state.

    kubectl create -f kubernetes/wordpress-db.yaml

This will create a mysql object in the cluster:

    -> % kubectl get mysql
    NAME           LABELS    DATA
    wordpres-gke   <none>    {"apiVersion":"infra.k8s.uk/v1","description":"Obj...
    wordpres-rds   <none>    {"apiVersion":"infra.k8s.uk/v1","description":"Obj...
    wordpress      <none>    {"apiVersion":"infra.k8s.uk/v1","description":"Obj...

That's it, from the user point of view, the only thing you have to do is to define a yaml file with a minimal configuration and submit it to the cluster, once the object is created, the heavy lifting happens behind the scenes and it's completely transparent to the user.

Great, but how is my database created?

The Operator has 2 processes: one that watches the api for changes and one that submits kubernetes jobs to the api server.

We split functionality in 2 services because it's more extesible this way. From the conceptual point of view there are two functions also: watching and sending work loads.