Setup

Provisioning

The provisioning consists of two parts:

Provision Cloud Resources
Deploying OpenShift

Cloud Resources

$ ./provision.sh

For easier explanation, further sections in the document assumes you have provisioned for one cloud say gcp

Connecting to OpenShift Node

The following commands shows how you can connect to the provisioned instance via ssh:

$ cd $PROJECT_HOME/out/gcp
$ ./connect.sh

The connect.sh script also holds information about the public IP, the ssh user and the private key to be used.

Deploy OpenShift

After successful provisioning of Provisioning, there should be one directory per cloud created under $PROJECT_HOME/out.

e.g. The following shows the directory tree for Azure and GCP

 out
   |-azr
   |---inventory (1)
   |-----host_vars (2)
   |- connect.sh (3)
   |- host_prepare.yaml (4)
   |- deploy.sh (5)
   |- docker-storage-setup (6)
   |- add_openshift_users.yaml (7)
   |- add-openshift-users.sh (8)
   |- openshift_users.yaml (9)
   |-gcp
   |---inventory
   |-----host_vars
   |- connect.sh
   |- host_prepare.yaml
   |- deploy.sh
   |- docker-storage-setup
   |- add_openshift_users.yaml (7)
   |- add-openshift-users.sh (8)
   |- openshift_users.yaml (9)

1	The cloud specific Ansible Inventory directory
2	host_vars, the Ansible host variables for the cloud provider
3	The SSH connect utility, this has the IP address of the OpenShift
4	The Cloud Host OpenShift Deployment preparation tasks
5	The OpenShift Deploy script
6	The Docker storage setup file
7	The Ansible playbook to add users who will have access to the OpenShift Web Console
8	The utility script to run the `add_openshift_users` play
9	openshift_users.yaml the users that need to be added/modified/deleted from OpenShift users file

e.g. Lets say you want to deploy OpenShift to your Google Cloud Platform(gcp), run the following commands:

$ cd $PROJECT_HOME/out/gcp
$ ./deploy.sh

Add Users to OpenShift

There are no users created by default with OpenShift installation, this section details on how to add new users.

The OpenShift installed is by default configured to use HTPasswd as the identity provider, with HTPasswd identity provider, the default htpasswd file is /etc/origin/master/htpasswd.

The following section details on how to add/update/remove users from the htpasswd file to allow users access to the OpenShift Web Console.

The out/<cloud>/openshift_users.yaml has two variables defined:

openshift_users - a list of dict/hash with keys username and an optional password, if password is omitted a random 8 letter password will be generated

e.g.

openshift_users:
    - {username: "developer",password: "supers3cret"}
    - {username: "demo"} (1)

1	in this case the password for the user `demo` in this case will be generated

openshift_delete_users - a list of usernames that needs to be removed or deleted from OpenShift users htpasswd file e.g.

openshift_delete_users:
    - developer (1)

1	the user `developer` will be deleted from the OpenShift users htpasswd file

After you have defined the users, run the following command:

$ cd $PROJECT_HOME/out/gcp
$ ./add-openshift-users.sh

Adding Admin User to OpenShift

Follow the steps defined above to add a new user called admin with the password of your choice, to provide the user admin with Cluster Admin Privileges you might need to login to the node and execute the following commands:

$ cd $PROJECT_HOME/out/gcp
$ ./connect.sh
$ sudo -i
$ oc login -u system:admin
$ oc adm policy add-cluster-role-to-user cluster-admin admin

DeProvisioning

The undeploying of Cloud Resources are controlled by three main variables that are defined in env/extravars

gcp_rollback: False (1)
azure_rollback: False (2)
aws_rollback: False (3)

1	Set to `True` to undeploy GCP resources
2	Set to `True` to undeploy Azure resources
3	Set to `True` to undeploy AWS resources

$ ./deprovision.sh

Sometime the Cloud resources might take time to get terminated or deleted, please verify via the respective cloud console to make sure the resources are deleted.