MAS DevOps Ansible Collection Ansible CLI
Home Roles: OCP Mgmt ocp_provision Role Variables Role Variables - GPU Node Support Role Variables - ROKS Role Variables - ROSA Role Variables - FYRE Role Variables - IPI Role Variables - AWS Role Variables - GCP Example Playbook License

ocp_provision¤

Provision OCP cluster on IBM Cloud ROKS, ROSA, or DevIT Fyre.

Fyre clusters will be automatically reconfigured to enable NFS storage. By default this is made available via the nfs-client storage class and supports both ReadWriteOnce and ReadWriteMany access modes. The image-registry-storage PVC used by the OpenShift image registry component will also be reconfigured to use this storage class.

Role Variables¤

cluster_type¤

Infrastructure provider type for cluster provisioning.

  • Required
  • Environment Variable: CLUSTER_TYPE
  • Default: None

Purpose: Specifies which infrastructure provider to use for provisioning the OpenShift cluster. Determines provisioning method and required variables.

When to use: - Always required for cluster provisioning - Each type requires different provider-specific variables - Determines available features (e.g., GPU support for ROKS)

Valid values: fyre, roks, rosa, ipi - fyre: IBM DevIT Fyre clusters (internal development) - roks: IBM Cloud Red Hat OpenShift Kubernetes Service - rosa: AWS Red Hat OpenShift Service on AWS - ipi: Installer-Provisioned Infrastructure (bare metal/on-premises)

Impact: Determines provisioning workflow and which provider-specific variables are required. Each type has different capabilities and configuration options.

Related variables: - cluster_name: Name for the new cluster - ocp_version: OpenShift version to install - Provider-specific variables (ibmcloud_apikey, rosa_token, fyre_apikey, etc.)

Note: Fyre clusters automatically configure NFS storage. ROKS requires version format like 4.19_openshift.

cluster_name¤

Name for the new cluster.

  • Required
  • Environment Variable: CLUSTER_NAME
  • Default: None

Purpose: Specifies the name for the OpenShift cluster to be provisioned. Used as the cluster identifier in the provider's system.

When to use: - Always required for cluster provisioning - Must be unique within the provider's account/region - Used for cluster identification and resource naming

Valid values: String following provider naming conventions (typically lowercase alphanumeric with hyphens)

Impact: Determines the cluster name in the provider's system. Used for DNS, resource naming, and cluster identification.

Related variables: - cluster_type: Provider where cluster will be created - ocp_version: OpenShift version for the cluster

Note: Name must follow provider-specific naming rules. Some providers have length limits or character restrictions.

ocp_version¤

OpenShift version to install.

  • Required
  • Environment Variable: OCP_VERSION
  • Default: None

Purpose: Specifies which version of OpenShift Container Platform to install on the provisioned cluster.

When to use: - Always required for cluster provisioning - Use specific version for production (e.g., 4.19.14) - Use default for latest MAS-supported version - Use rotate for testing (version changes by day of week)

Valid values: - Specific version: 4.19, 4.19.14 - Alias: default (newest MAS-supported version) - Alias: rotate (predetermined version by day, for testing) - ROKS format: Must append _openshift (e.g., 4.19_openshift, 4.19.14_openshift)

Impact: Determines OpenShift version installed. Version must be compatible with MAS and available from the provider.

Related variables: - cluster_type: ROKS requires _openshift suffix

Note: IMPORTANT - For ROKS (cluster_type=roks), version MUST include _openshift suffix. The default alias selects the newest MAS-supported version. The rotate alias is for testing only.

ocp_storage_provider¤

Storage provider configuration for Fyre clusters.

  • Optional
  • Environment Variable: OCP_STORAGE_PROVIDER
  • Default: None

Purpose: Configures NFS storage for Fyre clusters, creating an nfs-client storage class and reconfiguring the image registry.

When to use: - Set to nfs for Fyre clusters to enable NFS storage - Only applies when cluster_type=fyre - Leave unset for other cluster types (ROKS, ROSA, IPI)

Valid values: nfs (for Fyre clusters only)

Impact: - nfs: Creates nfs-client storage class connected to infrastructure node, reconfigures image registry PVC to use NFS - Unset: No storage configuration changes

Related variables: - cluster_type: Must be fyre for this to have effect

Note: Only functional for Fyre clusters. When enabled, the existing image registry PVC is deleted and recreated with NFS storage. NFS storage class supports both ReadWriteOnce (RWO) and ReadWriteMany (RWX) access modes.

Role Variables - GPU Node Support¤

ocp_provision_gpu¤

Enable GPU worker nodes during provisioning.

  • Optional
  • Environment Variable: OCP_PROVISION_GPU
  • Default: false

Purpose: Controls whether GPU-enabled worker nodes are provisioned with the cluster. Required for GPU-intensive applications like MAS Visual Inspection (MVI).

When to use: - Set to true for MAS Visual Inspection deployments - Set to true for other GPU-intensive workloads - Leave as false (default) for standard deployments - Currently only supported for ROKS clusters

Valid values: true, false

Impact: - true: Provisions GPU worker pool with specified number of GPU nodes - false: No GPU nodes provisioned (standard cluster)

Related variables: - gpu_workerpool_name: Name of GPU worker pool - gpu_workers: Number of GPU nodes to provision - cluster_type: Must be roks for GPU support

Note: GPU support is currently only available for ROKS clusters. GPU nodes use mg4c.32x384.2xp100-GPU flavor with P100 GPUs.

gpu_workerpool_name¤

Name for GPU worker pool.

  • Optional
  • Environment Variable: GPU_WORKERPOOL_NAME
  • Default: gpu

Purpose: Specifies the name for the GPU worker pool to be created or modified in the cluster.

When to use: - Use default (gpu) for new GPU deployments - Set to existing pool name to modify rather than create new - Only applies when ocp_provision_gpu=true

Valid values: String following worker pool naming conventions

Impact: Determines GPU worker pool name. Using an existing name modifies that pool; using a new name creates a new pool.

Related variables: - ocp_provision_gpu: Must be true for this to apply - gpu_workers: Number of nodes in this pool - cluster_type: Must be roks

Note: If a GPU worker pool already exists with this name, it will be modified rather than creating a duplicate. Use the existing name to avoid multiple GPU pools.

gpu_workers¤

The number of GPU worker nodes that will be deploy in the cluster. The node created will have mg4c.32x384.2xp100-GPU flavor. This variable depends on ocp_provision_gpu and is currently supported on ROKS clusters only.

  • Optional
  • Environment Variable: GPU_WORKERS
  • Default Value: 1

compute_node_count¤

The number of compute nodes (i.e. worker nodes) allocate to the OCP cluster.

  • Optional
  • Environment Variable: COMPUTE_NODE_COUNT
  • Default Value: 3

controlplane_node_count¤

The number of control plane nodes (i.e. master nodes) allocate to the OCP cluster.

  • Optional
  • Environment Variable: CONTROLPLANE_NODE_COUNT
  • Default Value: 3

gpu_workerpool_name¤

The name of the gpu worker pool to added to or modify in the cluster. If already existing, use the existing name to avoid recreating another gpu worker pool unless that is the goal.

  • Environment Variable: GPU_WORKERPOOL_NAME
  • Default Value: gpu

Role Variables - ROKS¤

The following variables are only used when cluster_type = roks.

ibmcloud_apikey¤

The APIKey to be used by ibmcloud login comand.

  • Required if cluster_type = roks
  • Environment Variable: IBMCLOUD_APIKEY
  • Default Value: None

ibmcloud_endpoint¤

Override the default IBMCloud API endpoint.

  • Optional
  • Environment Variable: IBMCLOUD_ENDPOINT
  • Default Value: https://cloud.ibm.com

ibmcloud_resourcegroup¤

The resource group to create the cluster inside.

  • Optional
  • Environment Variable: IBMCLOUD_RESOURCEGROUP
  • Default Value: Default

roks_zone¤

IBM Cloud zone where the cluster should be provisioned.

  • Optional
  • Environment Variable: ROKS_ZONE
  • Default Value: dal10

roks_flavor¤

Worker node flavor

  • Optional
  • Environment Variable: ROKS_FLAVOR
  • Default Value: b3c.16x64.300gb

roks_workers¤

Number of worker nodes for the roks cluster

  • Optional
  • Environment Variable: ROKS_WORKERS
  • Default Value: 3

roks_flags¤

Can be used to specify additional parameters for the cluster creation

  • Optional
  • Environment Variable: ROKS_FLAGS
  • Default Value: None

Role Variables - ROSA¤

The following variables are only used when cluster_type = rosa.

rosa_token¤

Token to authenticate to the ROSA service. To obtain your API token login to the OpenShift cluster manager.

  • Required if cluster_type = rosa.
  • Environment Variable: ROSA_TOKEN
  • Default Value: None

rosa_cluster_admin_password¤

Password to set up for the cluster-admin user account on the OCP instance. You will need this to log onto the cluster after it is provisioned. If this is not set then a password is auto-generated.

  • Optional if cluster_type = rosa.
  • Environment Variable: ROSA_CLUSTER_ADMIN_PASSWORD
  • Default Value: None

rosa_compute_nodes¤

Number of compute nodes to deploy in the cluster.

  • Optional
  • Environment Variable: ROSA_COMPUTE_NODES
  • Default Value: 3

rosa_compute_machine_type¤

Worker nodes machine

  • Optional
  • Environment Variable: ROSA_COMPUTE_MACHINE_TYPE
  • Default Value: m5.4xlarge

rosa_config_dir¤

Config directory to hold the rosa-{{cluster_name}}-details.yaml file that contains the api endpoint and cluster-admin details

  • Optional
  • Environment Variable: ROSA_CONFIG_DIR
  • Default Value: None

Role Variables - FYRE¤

The following variables are only used when cluster_type = fyre.

fyre_username¤

Username to authenticate with Fyre API.

  • Required if cluster_type = fyre.
  • Environment Variable: FYRE_USERNAME
  • Default Value: None

fyre_apikey¤

API key to authenticate with Fyre API.

  • Required if cluster_type = fyre.
  • Environment Variable: FYRE_APIKEY
  • Default Value: None

fyre_quota_type¤

Type of quota to draw from when provisioning the cluster, valid options are quick_burn and product_group.

  • Required if cluster_type = fyre.
  • Environment Variable: FYRE_QUOTA_TYPE
  • Default Value: quick_burn

fyre_product_id¤

The Product ID that the cluster will be associated with for accounting purposes.

  • Required if cluster_type = fyre.
  • Environment Variable: FYRE_PRODUCT_ID
  • Default Value: None

fyre_site¤

Provide a site in Fyre where cluster will be provisioned

  • Optional
  • Environment Variable: FYRE_SITE
  • Default Value: svl

fyre_cluster_description¤

Provide a description for the cluster.

  • Optional
  • Environment Variable: FYRE_CLUSTER_DESCRIPTION
  • Default Value: None

ocp_fips_enabled¤

Set to true to provision a FIPS enabled cluster.

  • Optional
  • Environment Variable: OCP_FIPS_ENABLED
  • Default Value: false

fyre_cluster_size¤

The name of one of Fyre's pre-defined cluster sizes to use for the new cluster.

  • Required when cluster_type = fyre and fyre_quota_type = quick_burn.
  • Environment Variable: FYRE_CLUSTER_SIZE
  • Default Value: medium

fyre_worker_count¤

The number of worker nodes to provision in the cluster.

  • Required when cluster_type = fyre and fyre_quota_type = product_group.
  • Environment Variable: FYRE_WORKER_COUNT
  • Default Value: 2

fyre_worker_cpu¤

The amount of CPU to assign to each worker node (maximum value supported by FYRE 16).

  • Required when cluster_type = fyre and fyre_quota_type = product_group.
  • Environment Variable: FYRE_WORKER_CPU
  • Default Value: 8

fyre_worker_memory¤

The amount of memory to assign to each worker node (maximum value supported by FYRE 64).

  • Required when cluster_type = fyre and fyre_quota_type = product_group.
  • Environment Variable: FYRE_WORKER_MEMORY
  • Default Value: 32

fyre_worker_additional_disks¤

The size of additional disks in Gb added to each worker node, defined in a comma-seperated list, e.g. 400,400 will add two 400gb disks to each worker node. By default no additional disks will be attached.

  • Optional
  • Environment Variable: FYRE_WORKER_ADDITIONAL_DISKS
  • Default Value: None

fyre_nfs_image_registry_size¤

Defines the image registry storage size when configured to use NFS. The size allocated cannot be superior of storage available in the Fyre Infrastructure node.

  • Optional
  • Environment Variable: FYRE_NFS_IMAGE_REGISTRY_SIZE
  • Default: 100Gi

enable_ipv6¤

Enable IPv6. This is for Fyre at RTP site only.

  • Environment Variable: ENABLE_IPV6
  • Default: False

Role Variables - IPI¤

These variables are only used when cluster_type = ipi.

Note

IPI stands for Installer Provisioned Infrastructure. OpenShift offers two possible deployment methods: IPI and UPI (User Provisioned Infrastructure). The difference is the degree of automation and customization. IPI will not only deploy OpenShift but also all infrastructure components and configurations.

ipi_platform¤

Platform to create the cluster on. Technically, any platform supported by openshift-install should work here, but currently we have only specifically tested on aws and gcp , where aws is the default value.

  • Optional when cluster_type = ipi
  • Environment Variable: IPI_PLATFORM
  • Default Value: aws

ipi_region¤

Platform region where OCP cluster will be created.

  • Optional when cluster_type = ipi
  • Environment Variable: IPI_REGION
  • Default Value: us-east-1

ipi_base_domain¤

Specify the base domain of the cluster that will be provisioned.

  • Required when cluster_type = ipi
  • Environment Variable: IPI_BASE_DOMAIN
  • Default Value: None

ipi_pull_secret_file¤

Location of the file containing your Redhat OpenShift pull secret. This file can be obtained from the Red Hat Hybrid Cloud Console

  • Required when cluster_type = ipi
  • Environment Variable: IPI_PULL_SECRET_FILE
  • Default Value: None

ipi_dir¤

The working directory that is used to perform the installation, it will contain the openshift-install executable, its configuration files, & any generated logs.

  • Optional when cluster_type = ipi
  • Environment Variable: IPI_DIR
  • Default Value: ~/openshift-install

sshKey¤

Public SSH key value. It will be set in the OCP cluster nodes. Can be used to SSH into the OCP cluster nodes using a bastion.

  • Optional when cluster_type = ipi
  • Environment Variable: SSH_PUB_KEY

ipi_controlplane_type¤

Control plane node type.

  • Optional when cluster_type = ipi
  • Environment Variable: IPI_CONTROLPLANE_TYPE
  • Default Value: m5.4xlarge

ipi_controlplane_replicas¤

The number of master nodes to provision to form the control plane of your cluster.

  • Optional when cluster_type = ipi
  • Environment Variable: IPI_CONTROLPLANE_REPLICAS
  • Default Value: 3

ipi_compute_type¤

Compute node type.

  • Optional when cluster_type = ipi
  • Environment Variable: IPI_COMPUTE_TYPE
  • Default Value: m5.4xlarge

ipi_compute_replicas¤

The number of worker nodes to provsision in the cluster, providing your compute resource.

  • Optional when cluster_type = ipi
  • Environment Variable: IPI_COMPUTE_REPLICAS
  • Default Value: 3

ipi_rootvolume_size¤

The size of root volume in GiB.

  • Optional when cluster_type = ipi
  • Environment variable: IPI_ROOTVOLUME_SIZE

Role Variables - AWS¤

The following variables are only used when cluster_type = ipi and ipi_platform = aws.

aws_access_key_id¤

AWS access key associated with an IAM user or role. Make sure the access key has permissions to create instances.

  • Required when cluster_type = ipi and ipi_platform = aws
  • Environment Variable: AWS_ACCESS_KEY_ID
  • Default Value: None

aws_secret_access_key¤

AWS secret access key associated with an IAM user or role.

  • Required when cluster_type = aws-ipi and ipi_platform = aws
  • Environment Variable: AWS_SECRET_ACCESS_KEY
  • Default Value: None

Role Variables - GCP¤

The following variables are only used when cluster_type = ipi and ipi_platform = gcp.

gcp_service_account_file¤

GCP service account file path. Make sure the service account has permissions to create instances.

  • Required when cluster_type = ipi and ipi_platform = gcp
  • Environment Variable: GOOGLE_APPLICATION_CREDENTIALS
  • Default Value: None

ipi_gcp_projectid¤

GCP project id in which the cluster will be deployed.

  • Required when cluster_type = ipi and ipi_platform = gcp
  • Environment Variable: GOOGLE_PROJECTID
  • Default Value: None

Example Playbook¤

- hosts: localhost
  vars:
    cluster_type: roks
    cluster_name: mycluster
    ocp_version: 4.10

    ibmcloud_apikey: xxxxx
  roles:
    - ibm.mas_devops.ocp_provision

License¤

EPL-2.0