should add docs how to manage different type of data sources

[hoatle]

should add docs how to manage different type of data sources

[hoatle]

ref: https://coreos.com/blog/introducing-operators.html

[hoatle]

https://github.com/operator-framework use operator to automate it all

[hoatle]

expect: operators are used to manage all restful apps

[hoatle]

https://github.com/oracle/mysql-operator

[hoatle]

https://github.com/operator-framework/awesome-operators

[hoatle]

Rook Operator Deployment

• set the VM RAM to be 4GB is recommended

• from https://rook.io/docs/rook/v0.8/helm-operator.html:

• make sure to identify the kubelet_flexvolumes_plugins_dir, it is /var/lib/kubelet/volume-plugins by default on kubespray

• Set the agent.flexVolumeDirPath key to the right value:

$ helm install --namespace rook-ceph-system rook-beta/rook-ceph --set=agent.flexVolumeDirPath=/var/lib/kubelet/volume-plugins

Just need to note on the flexVolumeDirPath to set the right value, for other things, just follow the Rook docs.

[hoatle]

after rook-ceph-block storageclass is available, we can use it mysql-operator cluster, for example:

apiVersion: mysql.oracle.com/v1alpha1
kind: Cluster
metadata:
  name: my-app-db
  namespace: test
spec:
  members: 3
  volumeClaimTemplate:
    metadata:
      name: data
    spec:
      storageClassName: rook-ceph-block
      accessModes:
        - ReadWriteOnce
      resources:
        requests:
          storage: 1Gi

[hoatle]

for stateful apps, we need to share a ReadWriteMany (RWX) access mode volume, use this https://rook.io/docs/rook/master/nfs.html

[hoatle]

useful: https://commons.openshift.org/sig/operators.html

[hoatle]

related: https://github.com/storageos/cluster-operator