Skip to content
Branch: master
Find file History
Permalink
Type Name Latest commit message Commit time
..
Failed to load latest commit information.
Dockerfile Dinathom/PSO Operator (#78) Apr 15, 2019
README.md 2.4.1 Release (#84) Apr 26, 2019
build.sh
install.sh 2.5.1 update - fixed ClusterRoleBinding & RoleBinding roleRef. (#95) May 22, 2019
update.sh Fix update script (#85) Apr 26, 2019
watches.yaml Dinathom/PSO Operator (#78) Apr 15, 2019

README.md

PSO Operator

Overview

PSO Operator is now the preferred install method for PSO on OpenShift 3.11 and higher versions. The PSO Operator packages and deploys the Pure Service Orchestrator (PSO) on OpenShift for dynamic provisioning of persistent volumes on FlashArrays and FlashBlades. The minimum supported version is OpenShift 3.11. This Operator is created as a Custom Resource Definition from the pure-k8s-plugin Helm chart using the Operator-SDK. This installation process does not require Helm installation.

Prerequisites

  • OpenShift (3.11+) or Kubernetes (1.11+) cluster installed and running
  • Access to a user account that has cluster-admin privileges

Installation

A single install script sets up the PSO-operator.
install.sh --image=<image> --namespace=<namespace> --orchestrator=<ochestrator> -f <values.yaml>

Parameter list:

  1. image is the PSO Operator image. If unspecified image resolves to the released version at quay.io/purestorage/pso-operator.
  2. namespace is the namespace/project in which the PSO-operator and its entities will be installed. If unspecified, the operator creates and installs in the pso-operator namespace. PSO Operator MUST be installed in a new project with no other pods. Otherwise an uninstall may delete pods that are not related to the PSO operator.
  3. orchestrator should be either k8s or openshift depending on which orchestrator is being used. If unspecified, openshift is assumed.
  4. values.yaml is the customized helm-chart configuration parameters. This is a required parameter and must contain the list of all backend FlashArrays and FlashBlades. All parameters that need a non-default value must be specified in this file. Refer to Configuration for values.yaml.

Install script steps:

The install script will do the following:

  1. Create New Project
    The script creates a new project (if it does not already exist) with the given namespace. If no namespace parameter is specified, the pso-operator namespace is used.
    OpenShift Note: In OpenShift 3.11, the default node-selector for a project does not allow PSO Operator to mount volumes on master and infra nodes. If you want to mount volumes on master and infra nodes OR run pods in the default namespace using volumes mounted by PSO, then modify the install script as follows.
Existing line:     $KUBECTL adm new-project ${NAMESPACE}

Change to allow volume mounts on master and infra nodes:     $KUBECTL adm new-project ${NAMESPACE} --node-selector=""
  1. Create a Custom Resource Definition (CRD) for the PSO Operator.
    The script waits for the CRD to be published in the cluster. If after 10 seconds the API server has not setup the CRD, the script times out. To wait longer, pass the parameter --timeout=<timeout_in_sec> to the install script.

  2. Create RBAC rules for the Operator.
    The PSO Operator needs the following Cluster-level Roles and RoleBindings.

Resource Permissions Notes
Namespace Get PSO Operator needs the ability to get created namespaces
Storageclass Create/Delete Create and cleanup storage classes to be used for Provisioning
ClusterRoleBinding Create/Delete PSO Operator needs to create and cleanup a ClusterRoleBinding called pure-provisioner-rights to the ClusterRole system:persistent-volume-provisioner for provisioning PVs
ClusterRoleBinding Get permission on ResourceName pure-provisioner-rights PSO-operator needs the ability to get the ClusterRoleBinding it has created for provisioning.

In addition, the operator needs access to multiple resources in the project/namespace that it is deployed in to function correctly. Hence it is recommended to install the PSO-operator in the non-default namespace.

  1. Creates a deployment for the Operator.
    Finally the script creates and deploys the operator using the customized parameters passed in the values.yaml file.

Upgrading

How to upgrade from helm install to PSO Operator

This upgrade will not impact the in-use volumes/filesystems from data path perspective. However, it will affect the in-flight volume/filesystem management operations. So, it is recommended to stop all the volume/filesystem management operations before doing this upgrade. Otherwise, these operations may need to be retried after the upgrade. Remove the helm-chart using instructions in https://helm.sh/docs/using_helm/#uninstall-a-release Once the helm chart has been uninstalled, follow the install instructions above.

Apply changes in values.yaml

The update.sh script is used to apply changes from values.yaml as follows.

./update.sh -f values.yaml

Uninstall

To uninstall the PSO Operator, run

oc delete all --all -n <pso-operator-installed-namespace>

where pso-operator-installed-namespace is the project/namespace in which the PSO operator is installed. It is strongly recommended to install the PSO Operator in a new project and not add any other pods to this project/namespace. Any pods in this project will be cleaned up on an uninstall.

You can’t perform that action at this time.